caio/clang-p2996
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
Files
be1294de9dd93ae02002bd1aa4b4a8d62e562c21
clang-p2996/llvm/test/CodeGen/Mips/msa/f16-llvm-ir.ll
Bjorn Pettersson 4c7f820b2b Update @llvm.powi to handle different int sizes for the exponent 
This can be seen as a follow up to commit 0ee439b705,
that changed the second argument of __powidf2, __powisf2 and
__powitf2 in compiler-rt from si_int to int. That was to align with
how those runtimes are defined in libgcc.
One thing that seem to have been missing in that patch was to make
sure that the rest of LLVM also handle that the argument now depends
on the size of int (not using the si_int machine mode for 32-bit).
When using __builtin_powi for a target with 16-bit int clang crashed.
And when emitting libcalls to those rtlib functions, typically when
lowering @llvm.powi), the backend would always prepare the exponent
argument as an i32 which caused miscompiles when the rtlib was
compiled with 16-bit int.

The solution used here is to use an overloaded type for the second
argument in @llvm.powi. This way clang can use the "correct" type
when lowering __builtin_powi, and then later when emitting the libcall
it is assumed that the type used in @llvm.powi matches the rtlib
function.

One thing that needed some extra attention was that when vectorizing
calls several passes did not support that several arguments could
be overloaded in the intrinsics. This patch allows overload of a
scalar operand by adding hasVectorInstrinsicOverloadedScalarOpd, with
an entry for powi.

Differential Revision: https://reviews.llvm.org/D99439
2021-06-17 09:38:28 +02:00

3313 lines
119 KiB
LLVM
Raw Blame History

; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc -relocation-model=pic -mtriple=mipsel-- -mcpu=mips32r5 -mattr=+abs2008 \
; RUN: -mattr=+fp64,+msa -verify-machineinstrs -mips-jalr-reloc=false < %s | FileCheck %s \
; RUN: --check-prefixes=MIPS32,MIPS32-O32
; RUN: llc -relocation-model=pic -mtriple=mips64el-- -mcpu=mips64r5 -mattr=+abs2008 \
; RUN: -mattr=+fp64,+msa -verify-machineinstrs -target-abi n32 -mips-jalr-reloc=false < %s | FileCheck %s \
; RUN: --check-prefixes=MIPS64-N32,MIPS64R5-N32
; RUN: llc -relocation-model=pic -mtriple=mips64el-- -mcpu=mips64r5 -mattr=+abs2008 \
; RUN: -mattr=+fp64,+msa -verify-machineinstrs -target-abi n64 -mips-jalr-reloc=false < %s | FileCheck %s \
; RUN: --check-prefixes=MIPS64-N64,MIPS64R5-N64
; RUN: llc -relocation-model=pic -mtriple=mipsel-- -mcpu=mips32r6 \
; RUN: -mattr=+fp64,+msa -verify-machineinstrs -mips-jalr-reloc=false < %s | FileCheck %s \
; RUN: --check-prefixes=MIPS32,MIPSR6-O32
; RUN: llc -relocation-model=pic -mtriple=mips64el-- -mcpu=mips64r6 \
; RUN: -mattr=+fp64,+msa -verify-machineinstrs -target-abi n32 -mips-jalr-reloc=false < %s | FileCheck %s \
; RUN: --check-prefixes=MIPS64-N32,MIPSR6-N32
; RUN: llc -relocation-model=pic -mtriple=mips64el-- -mcpu=mips64r6 \
; RUN: -mattr=+fp64,+msa -verify-machineinstrs -target-abi n64 -mips-jalr-reloc=false < %s | FileCheck %s \
; RUN: --check-prefixes=MIPS64-N64,MIPSR6-N64
; Check the use of frame indexes in the msa pseudo f16 instructions.
@k = external global float
declare float @k2(half *)
define void @f3(i16 %b) {
; MIPS32-LABEL: f3:
; MIPS32: # %bb.0: # %entry
; MIPS32-NEXT: lui $2, %hi(_gp_disp)
; MIPS32-NEXT: addiu $2, $2, %lo(_gp_disp)
; MIPS32-NEXT: addiu $sp, $sp, -32
; MIPS32-NEXT: .cfi_def_cfa_offset 32
; MIPS32-NEXT: sw $ra, 28($sp) # 4-byte Folded Spill
; MIPS32-NEXT: sw $16, 24($sp) # 4-byte Folded Spill
; MIPS32-NEXT: .cfi_offset 31, -4
; MIPS32-NEXT: .cfi_offset 16, -8
; MIPS32-NEXT: addu $16, $2, $25
; MIPS32-NEXT: sh $4, 22($sp)
; MIPS32-NEXT: addiu $4, $sp, 22
; MIPS32-NEXT: lw $25, %call16(k2)($16)
; MIPS32-NEXT: jalr $25
; MIPS32-NEXT: move $gp, $16
; MIPS32-NEXT: lw $1, %got(k)($16)
; MIPS32-NEXT: swc1 $f0, 0($1)
; MIPS32-NEXT: lw $16, 24($sp) # 4-byte Folded Reload
; MIPS32-NEXT: lw $ra, 28($sp) # 4-byte Folded Reload
; MIPS32-NEXT: jr $ra
; MIPS32-NEXT: addiu $sp, $sp, 32
;
; MIPS64-N32-LABEL: f3:
; MIPS64-N32: # %bb.0: # %entry
; MIPS64-N32-NEXT: addiu $sp, $sp, -32
; MIPS64-N32-NEXT: .cfi_def_cfa_offset 32
; MIPS64-N32-NEXT: sd $ra, 24($sp) # 8-byte Folded Spill
; MIPS64-N32-NEXT: sd $gp, 16($sp) # 8-byte Folded Spill
; MIPS64-N32-NEXT: .cfi_offset 31, -8
; MIPS64-N32-NEXT: .cfi_offset 28, -16
; MIPS64-N32-NEXT: lui $1, %hi(%neg(%gp_rel(f3)))
; MIPS64-N32-NEXT: addu $1, $1, $25
; MIPS64-N32-NEXT: addiu $gp, $1, %lo(%neg(%gp_rel(f3)))
; MIPS64-N32-NEXT: sh $4, 14($sp)
; MIPS64-N32-NEXT: lw $25, %call16(k2)($gp)
; MIPS64-N32-NEXT: jalr $25
; MIPS64-N32-NEXT: addiu $4, $sp, 14
; MIPS64-N32-NEXT: lw $1, %got_disp(k)($gp)
; MIPS64-N32-NEXT: swc1 $f0, 0($1)
; MIPS64-N32-NEXT: ld $gp, 16($sp) # 8-byte Folded Reload
; MIPS64-N32-NEXT: ld $ra, 24($sp) # 8-byte Folded Reload
; MIPS64-N32-NEXT: jr $ra
; MIPS64-N32-NEXT: addiu $sp, $sp, 32
;
; MIPS64-N64-LABEL: f3:
; MIPS64-N64: # %bb.0: # %entry
; MIPS64-N64-NEXT: daddiu $sp, $sp, -32
; MIPS64-N64-NEXT: .cfi_def_cfa_offset 32
; MIPS64-N64-NEXT: sd $ra, 24($sp) # 8-byte Folded Spill
; MIPS64-N64-NEXT: sd $gp, 16($sp) # 8-byte Folded Spill
; MIPS64-N64-NEXT: .cfi_offset 31, -8
; MIPS64-N64-NEXT: .cfi_offset 28, -16
; MIPS64-N64-NEXT: lui $1, %hi(%neg(%gp_rel(f3)))
; MIPS64-N64-NEXT: daddu $1, $1, $25
; MIPS64-N64-NEXT: daddiu $gp, $1, %lo(%neg(%gp_rel(f3)))
; MIPS64-N64-NEXT: sh $4, 14($sp)
; MIPS64-N64-NEXT: ld $25, %call16(k2)($gp)
; MIPS64-N64-NEXT: jalr $25
; MIPS64-N64-NEXT: daddiu $4, $sp, 14
; MIPS64-N64-NEXT: ld $1, %got_disp(k)($gp)
; MIPS64-N64-NEXT: swc1 $f0, 0($1)
; MIPS64-N64-NEXT: ld $gp, 16($sp) # 8-byte Folded Reload
; MIPS64-N64-NEXT: ld $ra, 24($sp) # 8-byte Folded Reload
; MIPS64-N64-NEXT: jr $ra
; MIPS64-N64-NEXT: daddiu $sp, $sp, 32
entry:
%0 = alloca half
%1 = bitcast i16 %b to half
store half %1, half * %0
%2 = call float @k2(half * %0)
store float %2, float * @k
ret void
}
define void @f(i16 %b) {
; MIPS32-LABEL: f:
; MIPS32: # %bb.0:
; MIPS32-NEXT: lui $2, %hi(_gp_disp)
; MIPS32-NEXT: addiu $2, $2, %lo(_gp_disp)
; MIPS32-NEXT: addiu $sp, $sp, -8
; MIPS32-NEXT: .cfi_def_cfa_offset 8
; MIPS32-NEXT: addu $1, $2, $25
; MIPS32-NEXT: sh $4, 4($sp)
; MIPS32-NEXT: lh $2, 4($sp)
; MIPS32-NEXT: fill.h $w0, $2
; MIPS32-NEXT: fexupr.w $w0, $w0
; MIPS32-NEXT: copy_s.w $2, $w0[0]
; MIPS32-NEXT: mtc1 $2, $f0
; MIPS32-NEXT: lw $1, %got(k)($1)
; MIPS32-NEXT: swc1 $f0, 0($1)
; MIPS32-NEXT: jr $ra
; MIPS32-NEXT: addiu $sp, $sp, 8
;
; MIPS64-N32-LABEL: f:
; MIPS64-N32: # %bb.0:
; MIPS64-N32-NEXT: addiu $sp, $sp, -16
; MIPS64-N32-NEXT: .cfi_def_cfa_offset 16
; MIPS64-N32-NEXT: lui $1, %hi(%neg(%gp_rel(f)))
; MIPS64-N32-NEXT: addu $1, $1, $25
; MIPS64-N32-NEXT: addiu $1, $1, %lo(%neg(%gp_rel(f)))
; MIPS64-N32-NEXT: sh $4, 12($sp)
; MIPS64-N32-NEXT: lh $2, 12($sp)
; MIPS64-N32-NEXT: fill.h $w0, $2
; MIPS64-N32-NEXT: fexupr.w $w0, $w0
; MIPS64-N32-NEXT: copy_s.w $2, $w0[0]
; MIPS64-N32-NEXT: mtc1 $2, $f0
; MIPS64-N32-NEXT: lw $1, %got_disp(k)($1)
; MIPS64-N32-NEXT: swc1 $f0, 0($1)
; MIPS64-N32-NEXT: jr $ra
; MIPS64-N32-NEXT: addiu $sp, $sp, 16
;
; MIPS64-N64-LABEL: f:
; MIPS64-N64: # %bb.0:
; MIPS64-N64-NEXT: daddiu $sp, $sp, -16
; MIPS64-N64-NEXT: .cfi_def_cfa_offset 16
; MIPS64-N64-NEXT: lui $1, %hi(%neg(%gp_rel(f)))
; MIPS64-N64-NEXT: daddu $1, $1, $25
; MIPS64-N64-NEXT: daddiu $1, $1, %lo(%neg(%gp_rel(f)))
; MIPS64-N64-NEXT: sh $4, 12($sp)
; MIPS64-N64-NEXT: lh $2, 12($sp)
; MIPS64-N64-NEXT: fill.h $w0, $2
; MIPS64-N64-NEXT: fexupr.w $w0, $w0
; MIPS64-N64-NEXT: copy_s.w $2, $w0[0]
; MIPS64-N64-NEXT: mtc1 $2, $f0
; MIPS64-N64-NEXT: ld $1, %got_disp(k)($1)
; MIPS64-N64-NEXT: swc1 $f0, 0($1)
; MIPS64-N64-NEXT: jr $ra
; MIPS64-N64-NEXT: daddiu $sp, $sp, 16
%1 = bitcast i16 %b to half
%2 = fpext half %1 to float
store float %2, float * @k
ret void
}
@g = external global i16, align 2
@h = external global half, align 2
; Check that fext f16 to double has a fexupr.w, fexupr.d sequence.
; Check that ftrunc double to f16 has fexdo.w, fexdo.h sequence.
; Check that MIPS64R5+ uses 64-bit floating point <-> 64-bit GPR transfers.
; We don't need to check if pre-MIPSR5 expansions occur, the MSA ASE requires
; MIPSR5. Additionally, fp64 mode / FR=1 is required to use MSA.
define void @fadd_f64() {
; MIPS32-LABEL: fadd_f64:
; MIPS32: # %bb.0: # %entry
; MIPS32-NEXT: lui $2, %hi(_gp_disp)
; MIPS32-NEXT: addiu $2, $2, %lo(_gp_disp)
; MIPS32-NEXT: addu $1, $2, $25
; MIPS32-NEXT: lw $1, %got(h)($1)
; MIPS32-NEXT: lh $2, 0($1)
; MIPS32-NEXT: fill.h $w0, $2
; MIPS32-NEXT: fexupr.w $w0, $w0
; MIPS32-NEXT: fexupr.d $w0, $w0
; MIPS32-NEXT: copy_s.w $2, $w0[0]
; MIPS32-NEXT: mtc1 $2, $f1
; MIPS32-NEXT: copy_s.w $2, $w0[1]
; MIPS32-NEXT: mthc1 $2, $f1
; MIPS32-NEXT: add.d $f0, $f1, $f1
; MIPS32-NEXT: mfc1 $2, $f0
; MIPS32-NEXT: fill.w $w1, $2
; MIPS32-NEXT: mfhc1 $2, $f0
; MIPS32-NEXT: insert.w $w1[1], $2
; MIPS32-NEXT: insert.w $w1[3], $2
; MIPS32-NEXT: fexdo.w $w0, $w1, $w1
; MIPS32-NEXT: fexdo.h $w0, $w0, $w0
; MIPS32-NEXT: copy_u.h $2, $w0[0]
; MIPS32-NEXT: jr $ra
; MIPS32-NEXT: sh $2, 0($1)
;
; MIPS64-N32-LABEL: fadd_f64:
; MIPS64-N32: # %bb.0: # %entry
; MIPS64-N32-NEXT: lui $1, %hi(%neg(%gp_rel(fadd_f64)))
; MIPS64-N32-NEXT: addu $1, $1, $25
; MIPS64-N32-NEXT: addiu $1, $1, %lo(%neg(%gp_rel(fadd_f64)))
; MIPS64-N32-NEXT: lw $1, %got_disp(h)($1)
; MIPS64-N32-NEXT: lh $2, 0($1)
; MIPS64-N32-NEXT: fill.h $w0, $2
; MIPS64-N32-NEXT: fexupr.w $w0, $w0
; MIPS64-N32-NEXT: fexupr.d $w0, $w0
; MIPS64-N32-NEXT: copy_s.d $2, $w0[0]
; MIPS64-N32-NEXT: dmtc1 $2, $f0
; MIPS64-N32-NEXT: add.d $f0, $f0, $f0
; MIPS64-N32-NEXT: dmfc1 $2, $f0
; MIPS64-N32-NEXT: fill.d $w0, $2
; MIPS64-N32-NEXT: fexdo.w $w0, $w0, $w0
; MIPS64-N32-NEXT: fexdo.h $w0, $w0, $w0
; MIPS64-N32-NEXT: copy_u.h $2, $w0[0]
; MIPS64-N32-NEXT: jr $ra
; MIPS64-N32-NEXT: sh $2, 0($1)
;
; MIPS64-N64-LABEL: fadd_f64:
; MIPS64-N64: # %bb.0: # %entry
; MIPS64-N64-NEXT: lui $1, %hi(%neg(%gp_rel(fadd_f64)))
; MIPS64-N64-NEXT: daddu $1, $1, $25
; MIPS64-N64-NEXT: daddiu $1, $1, %lo(%neg(%gp_rel(fadd_f64)))
; MIPS64-N64-NEXT: ld $1, %got_disp(h)($1)
; MIPS64-N64-NEXT: lh $2, 0($1)
; MIPS64-N64-NEXT: fill.h $w0, $2
; MIPS64-N64-NEXT: fexupr.w $w0, $w0
; MIPS64-N64-NEXT: fexupr.d $w0, $w0
; MIPS64-N64-NEXT: copy_s.d $2, $w0[0]
; MIPS64-N64-NEXT: dmtc1 $2, $f0
; MIPS64-N64-NEXT: add.d $f0, $f0, $f0
; MIPS64-N64-NEXT: dmfc1 $2, $f0
; MIPS64-N64-NEXT: fill.d $w0, $2
; MIPS64-N64-NEXT: fexdo.w $w0, $w0, $w0
; MIPS64-N64-NEXT: fexdo.h $w0, $w0, $w0
; MIPS64-N64-NEXT: copy_u.h $2, $w0[0]
; MIPS64-N64-NEXT: jr $ra
; MIPS64-N64-NEXT: sh $2, 0($1)
entry:
%0 = load half, half * @h, align 2
%1 = fpext half %0 to double
%2 = load half, half * @h, align 2
%3 = fpext half %2 to double
%add = fadd double %1, %3
%4 = fptrunc double %add to half
store half %4, half * @h, align 2
ret void
}
; Entire fp16 (unsigned) range fits into (signed) i32.
define i32 @ffptoui() {
; MIPS32-LABEL: ffptoui:
; MIPS32: # %bb.0: # %entry
; MIPS32-NEXT: lui $2, %hi(_gp_disp)
; MIPS32-NEXT: addiu $2, $2, %lo(_gp_disp)
; MIPS32-NEXT: addu $1, $2, $25
; MIPS32-NEXT: lw $1, %got(h)($1)
; MIPS32-NEXT: lh $1, 0($1)
; MIPS32-NEXT: fill.h $w0, $1
; MIPS32-NEXT: fexupr.w $w0, $w0
; MIPS32-NEXT: fexupr.d $w0, $w0
; MIPS32-NEXT: copy_s.w $1, $w0[0]
; MIPS32-NEXT: mtc1 $1, $f1
; MIPS32-NEXT: copy_s.w $1, $w0[1]
; MIPS32-NEXT: mthc1 $1, $f1
; MIPS32-NEXT: trunc.w.d $f0, $f1
; MIPS32-NEXT: jr $ra
; MIPS32-NEXT: mfc1 $2, $f0
;
; MIPS64-N32-LABEL: ffptoui:
; MIPS64-N32: # %bb.0: # %entry
; MIPS64-N32-NEXT: lui $1, %hi(%neg(%gp_rel(ffptoui)))
; MIPS64-N32-NEXT: addu $1, $1, $25
; MIPS64-N32-NEXT: addiu $1, $1, %lo(%neg(%gp_rel(ffptoui)))
; MIPS64-N32-NEXT: lw $1, %got_disp(h)($1)
; MIPS64-N32-NEXT: lh $1, 0($1)
; MIPS64-N32-NEXT: fill.h $w0, $1
; MIPS64-N32-NEXT: fexupr.w $w0, $w0
; MIPS64-N32-NEXT: fexupr.d $w0, $w0
; MIPS64-N32-NEXT: copy_s.d $1, $w0[0]
; MIPS64-N32-NEXT: dmtc1 $1, $f0
; MIPS64-N32-NEXT: trunc.w.d $f0, $f0
; MIPS64-N32-NEXT: jr $ra
; MIPS64-N32-NEXT: mfc1 $2, $f0
;
; MIPS64-N64-LABEL: ffptoui:
; MIPS64-N64: # %bb.0: # %entry
; MIPS64-N64-NEXT: lui $1, %hi(%neg(%gp_rel(ffptoui)))
; MIPS64-N64-NEXT: daddu $1, $1, $25
; MIPS64-N64-NEXT: daddiu $1, $1, %lo(%neg(%gp_rel(ffptoui)))
; MIPS64-N64-NEXT: ld $1, %got_disp(h)($1)
; MIPS64-N64-NEXT: lh $1, 0($1)
; MIPS64-N64-NEXT: fill.h $w0, $1
; MIPS64-N64-NEXT: fexupr.w $w0, $w0
; MIPS64-N64-NEXT: fexupr.d $w0, $w0
; MIPS64-N64-NEXT: copy_s.d $1, $w0[0]
; MIPS64-N64-NEXT: dmtc1 $1, $f0
; MIPS64-N64-NEXT: trunc.w.d $f0, $f0
; MIPS64-N64-NEXT: jr $ra
; MIPS64-N64-NEXT: mfc1 $2, $f0
entry:
%0 = load half, half * @h, align 2
%1 = fptoui half %0 to i32
ret i32 %1
}
define i32 @ffptosi() {
; MIPS32-LABEL: ffptosi:
; MIPS32: # %bb.0: # %entry
; MIPS32-NEXT: lui $2, %hi(_gp_disp)
; MIPS32-NEXT: addiu $2, $2, %lo(_gp_disp)
; MIPS32-NEXT: addu $1, $2, $25
; MIPS32-NEXT: lw $1, %got(h)($1)
; MIPS32-NEXT: lh $1, 0($1)
; MIPS32-NEXT: fill.h $w0, $1
; MIPS32-NEXT: fexupr.w $w0, $w0
; MIPS32-NEXT: fexupr.d $w0, $w0
; MIPS32-NEXT: copy_s.w $1, $w0[0]
; MIPS32-NEXT: mtc1 $1, $f1
; MIPS32-NEXT: copy_s.w $1, $w0[1]
; MIPS32-NEXT: mthc1 $1, $f1
; MIPS32-NEXT: trunc.w.d $f0, $f1
; MIPS32-NEXT: jr $ra
; MIPS32-NEXT: mfc1 $2, $f0
;
; MIPS64-N32-LABEL: ffptosi:
; MIPS64-N32: # %bb.0: # %entry
; MIPS64-N32-NEXT: lui $1, %hi(%neg(%gp_rel(ffptosi)))
; MIPS64-N32-NEXT: addu $1, $1, $25
; MIPS64-N32-NEXT: addiu $1, $1, %lo(%neg(%gp_rel(ffptosi)))
; MIPS64-N32-NEXT: lw $1, %got_disp(h)($1)
; MIPS64-N32-NEXT: lh $1, 0($1)
; MIPS64-N32-NEXT: fill.h $w0, $1
; MIPS64-N32-NEXT: fexupr.w $w0, $w0
; MIPS64-N32-NEXT: fexupr.d $w0, $w0
; MIPS64-N32-NEXT: copy_s.d $1, $w0[0]
; MIPS64-N32-NEXT: dmtc1 $1, $f0
; MIPS64-N32-NEXT: trunc.w.d $f0, $f0
; MIPS64-N32-NEXT: jr $ra
; MIPS64-N32-NEXT: mfc1 $2, $f0
;
; MIPS64-N64-LABEL: ffptosi:
; MIPS64-N64: # %bb.0: # %entry
; MIPS64-N64-NEXT: lui $1, %hi(%neg(%gp_rel(ffptosi)))
; MIPS64-N64-NEXT: daddu $1, $1, $25
; MIPS64-N64-NEXT: daddiu $1, $1, %lo(%neg(%gp_rel(ffptosi)))
; MIPS64-N64-NEXT: ld $1, %got_disp(h)($1)
; MIPS64-N64-NEXT: lh $1, 0($1)
; MIPS64-N64-NEXT: fill.h $w0, $1
; MIPS64-N64-NEXT: fexupr.w $w0, $w0
; MIPS64-N64-NEXT: fexupr.d $w0, $w0
; MIPS64-N64-NEXT: copy_s.d $1, $w0[0]
; MIPS64-N64-NEXT: dmtc1 $1, $f0
; MIPS64-N64-NEXT: trunc.w.d $f0, $f0
; MIPS64-N64-NEXT: jr $ra
; MIPS64-N64-NEXT: mfc1 $2, $f0
entry:
%0 = load half, half * @h, align 2
%1 = fptosi half %0 to i32
ret i32 %1
}
define void @uitofp(i32 %a) {
; MIPS32-LABEL: uitofp:
; MIPS32: # %bb.0: # %entry
; MIPS32-NEXT: lui $2, %hi(_gp_disp)
; MIPS32-NEXT: addiu $2, $2, %lo(_gp_disp)
; MIPS32-NEXT: addiu $sp, $sp, -8
; MIPS32-NEXT: .cfi_def_cfa_offset 8
; MIPS32-NEXT: addu $1, $2, $25
; MIPS32-NEXT: lui $2, 17200
; MIPS32-NEXT: sw $2, 4($sp)
; MIPS32-NEXT: sw $4, 0($sp)
; MIPS32-NEXT: lw $2, %got($CPI5_0)($1)
; MIPS32-NEXT: ldc1 $f0, %lo($CPI5_0)($2)
; MIPS32-NEXT: ldc1 $f1, 0($sp)
; MIPS32-NEXT: sub.d $f0, $f1, $f0
; MIPS32-NEXT: mfc1 $2, $f0
; MIPS32-NEXT: fill.w $w1, $2
; MIPS32-NEXT: mfhc1 $2, $f0
; MIPS32-NEXT: insert.w $w1[1], $2
; MIPS32-NEXT: insert.w $w1[3], $2
; MIPS32-NEXT: fexdo.w $w0, $w1, $w1
; MIPS32-NEXT: fexdo.h $w0, $w0, $w0
; MIPS32-NEXT: lw $1, %got(h)($1)
; MIPS32-NEXT: copy_u.h $2, $w0[0]
; MIPS32-NEXT: sh $2, 0($1)
; MIPS32-NEXT: jr $ra
; MIPS32-NEXT: addiu $sp, $sp, 8
;
; MIPS64-N32-LABEL: uitofp:
; MIPS64-N32: # %bb.0: # %entry
; MIPS64-N32-NEXT: addiu $sp, $sp, -16
; MIPS64-N32-NEXT: .cfi_def_cfa_offset 16
; MIPS64-N32-NEXT: lui $1, %hi(%neg(%gp_rel(uitofp)))
; MIPS64-N32-NEXT: addu $1, $1, $25
; MIPS64-N32-NEXT: addiu $1, $1, %lo(%neg(%gp_rel(uitofp)))
; MIPS64-N32-NEXT: lui $2, 17200
; MIPS64-N32-NEXT: sw $2, 12($sp)
; MIPS64-N32-NEXT: sll $2, $4, 0
; MIPS64-N32-NEXT: sw $2, 8($sp)
; MIPS64-N32-NEXT: lw $2, %got_page(.LCPI5_0)($1)
; MIPS64-N32-NEXT: ldc1 $f0, %got_ofst(.LCPI5_0)($2)
; MIPS64-N32-NEXT: ldc1 $f1, 8($sp)
; MIPS64-N32-NEXT: sub.d $f0, $f1, $f0
; MIPS64-N32-NEXT: dmfc1 $2, $f0
; MIPS64-N32-NEXT: fill.d $w0, $2
; MIPS64-N32-NEXT: fexdo.w $w0, $w0, $w0
; MIPS64-N32-NEXT: fexdo.h $w0, $w0, $w0
; MIPS64-N32-NEXT: lw $1, %got_disp(h)($1)
; MIPS64-N32-NEXT: copy_u.h $2, $w0[0]
; MIPS64-N32-NEXT: sh $2, 0($1)
; MIPS64-N32-NEXT: jr $ra
; MIPS64-N32-NEXT: addiu $sp, $sp, 16
;
; MIPS64-N64-LABEL: uitofp:
; MIPS64-N64: # %bb.0: # %entry
; MIPS64-N64-NEXT: daddiu $sp, $sp, -16
; MIPS64-N64-NEXT: .cfi_def_cfa_offset 16
; MIPS64-N64-NEXT: lui $1, %hi(%neg(%gp_rel(uitofp)))
; MIPS64-N64-NEXT: daddu $1, $1, $25
; MIPS64-N64-NEXT: daddiu $1, $1, %lo(%neg(%gp_rel(uitofp)))
; MIPS64-N64-NEXT: lui $2, 17200
; MIPS64-N64-NEXT: sw $2, 12($sp)
; MIPS64-N64-NEXT: sll $2, $4, 0
; MIPS64-N64-NEXT: sw $2, 8($sp)
; MIPS64-N64-NEXT: ld $2, %got_page(.LCPI5_0)($1)
; MIPS64-N64-NEXT: ldc1 $f0, %got_ofst(.LCPI5_0)($2)
; MIPS64-N64-NEXT: ldc1 $f1, 8($sp)
; MIPS64-N64-NEXT: sub.d $f0, $f1, $f0
; MIPS64-N64-NEXT: dmfc1 $2, $f0
; MIPS64-N64-NEXT: fill.d $w0, $2
; MIPS64-N64-NEXT: fexdo.w $w0, $w0, $w0
; MIPS64-N64-NEXT: fexdo.h $w0, $w0, $w0
; MIPS64-N64-NEXT: ld $1, %got_disp(h)($1)
; MIPS64-N64-NEXT: copy_u.h $2, $w0[0]
; MIPS64-N64-NEXT: sh $2, 0($1)
; MIPS64-N64-NEXT: jr $ra
; MIPS64-N64-NEXT: daddiu $sp, $sp, 16
entry:
%0 = uitofp i32 %a to half
store half %0, half * @h, align 2
ret void
}
; Check that f16 is expanded to f32 and relevant transfer ops occur.
; We don't check f16 -> f64 expansion occurs, as we expand f16 to f32.
define void @fadd() {
; MIPS32-LABEL: fadd:
; MIPS32: # %bb.0: # %entry
; MIPS32-NEXT: lui $2, %hi(_gp_disp)
; MIPS32-NEXT: addiu $2, $2, %lo(_gp_disp)
; MIPS32-NEXT: addu $1, $2, $25
; MIPS32-NEXT: lw $1, %got(g)($1)
; MIPS32-NEXT: lh $2, 0($1)
; MIPS32-NEXT: fill.h $w0, $2
; MIPS32-NEXT: fexupr.w $w0, $w0
; MIPS32-NEXT: copy_s.w $2, $w0[0]
; MIPS32-NEXT: mtc1 $2, $f0
; MIPS32-NEXT: add.s $f0, $f0, $f0
; MIPS32-NEXT: mfc1 $2, $f0
; MIPS32-NEXT: fill.w $w0, $2
; MIPS32-NEXT: fexdo.h $w0, $w0, $w0
; MIPS32-NEXT: copy_u.h $2, $w0[0]
; MIPS32-NEXT: jr $ra
; MIPS32-NEXT: sh $2, 0($1)
;
; MIPS64-N32-LABEL: fadd:
; MIPS64-N32: # %bb.0: # %entry
; MIPS64-N32-NEXT: lui $1, %hi(%neg(%gp_rel(fadd)))
; MIPS64-N32-NEXT: addu $1, $1, $25
; MIPS64-N32-NEXT: addiu $1, $1, %lo(%neg(%gp_rel(fadd)))
; MIPS64-N32-NEXT: lw $1, %got_disp(g)($1)
; MIPS64-N32-NEXT: lh $2, 0($1)
; MIPS64-N32-NEXT: fill.h $w0, $2
; MIPS64-N32-NEXT: fexupr.w $w0, $w0
; MIPS64-N32-NEXT: copy_s.w $2, $w0[0]
; MIPS64-N32-NEXT: mtc1 $2, $f0
; MIPS64-N32-NEXT: add.s $f0, $f0, $f0
; MIPS64-N32-NEXT: mfc1 $2, $f0
; MIPS64-N32-NEXT: fill.w $w0, $2
; MIPS64-N32-NEXT: fexdo.h $w0, $w0, $w0
; MIPS64-N32-NEXT: copy_u.h $2, $w0[0]
; MIPS64-N32-NEXT: jr $ra
; MIPS64-N32-NEXT: sh $2, 0($1)
;
; MIPS64-N64-LABEL: fadd:
; MIPS64-N64: # %bb.0: # %entry
; MIPS64-N64-NEXT: lui $1, %hi(%neg(%gp_rel(fadd)))
; MIPS64-N64-NEXT: daddu $1, $1, $25
; MIPS64-N64-NEXT: daddiu $1, $1, %lo(%neg(%gp_rel(fadd)))
; MIPS64-N64-NEXT: ld $1, %got_disp(g)($1)
; MIPS64-N64-NEXT: lh $2, 0($1)
; MIPS64-N64-NEXT: fill.h $w0, $2
; MIPS64-N64-NEXT: fexupr.w $w0, $w0
; MIPS64-N64-NEXT: copy_s.w $2, $w0[0]
; MIPS64-N64-NEXT: mtc1 $2, $f0
; MIPS64-N64-NEXT: add.s $f0, $f0, $f0
; MIPS64-N64-NEXT: mfc1 $2, $f0
; MIPS64-N64-NEXT: fill.w $w0, $2
; MIPS64-N64-NEXT: fexdo.h $w0, $w0, $w0
; MIPS64-N64-NEXT: copy_u.h $2, $w0[0]
; MIPS64-N64-NEXT: jr $ra
; MIPS64-N64-NEXT: sh $2, 0($1)
entry:
%0 = load i16, i16* @g, align 2
%1 = call float @llvm.convert.from.fp16.f32(i16 %0)
%2 = load i16, i16* @g, align 2
%3 = call float @llvm.convert.from.fp16.f32(i16 %2)
%add = fadd float %1, %3
%4 = call i16 @llvm.convert.to.fp16.f32(float %add)
store i16 %4, i16* @g, align 2
ret void
}
; Function Attrs: nounwind readnone
declare float @llvm.convert.from.fp16.f32(i16)
; Function Attrs: nounwind readnone
declare i16 @llvm.convert.to.fp16.f32(float)
; Function Attrs: nounwind
define void @fsub() {
; MIPS32-LABEL: fsub:
; MIPS32: # %bb.0: # %entry
; MIPS32-NEXT: lui $2, %hi(_gp_disp)
; MIPS32-NEXT: addiu $2, $2, %lo(_gp_disp)
; MIPS32-NEXT: addu $1, $2, $25
; MIPS32-NEXT: lw $1, %got(g)($1)
; MIPS32-NEXT: lh $2, 0($1)
; MIPS32-NEXT: fill.h $w0, $2
; MIPS32-NEXT: fexupr.w $w0, $w0
; MIPS32-NEXT: copy_s.w $2, $w0[0]
; MIPS32-NEXT: mtc1 $2, $f0
; MIPS32-NEXT: sub.s $f0, $f0, $f0
; MIPS32-NEXT: mfc1 $2, $f0
; MIPS32-NEXT: fill.w $w0, $2
; MIPS32-NEXT: fexdo.h $w0, $w0, $w0
; MIPS32-NEXT: copy_u.h $2, $w0[0]
; MIPS32-NEXT: jr $ra
; MIPS32-NEXT: sh $2, 0($1)
;
; MIPS64-N32-LABEL: fsub:
; MIPS64-N32: # %bb.0: # %entry
; MIPS64-N32-NEXT: lui $1, %hi(%neg(%gp_rel(fsub)))
; MIPS64-N32-NEXT: addu $1, $1, $25
; MIPS64-N32-NEXT: addiu $1, $1, %lo(%neg(%gp_rel(fsub)))
; MIPS64-N32-NEXT: lw $1, %got_disp(g)($1)
; MIPS64-N32-NEXT: lh $2, 0($1)
; MIPS64-N32-NEXT: fill.h $w0, $2
; MIPS64-N32-NEXT: fexupr.w $w0, $w0
; MIPS64-N32-NEXT: copy_s.w $2, $w0[0]
; MIPS64-N32-NEXT: mtc1 $2, $f0
; MIPS64-N32-NEXT: sub.s $f0, $f0, $f0
; MIPS64-N32-NEXT: mfc1 $2, $f0
; MIPS64-N32-NEXT: fill.w $w0, $2
; MIPS64-N32-NEXT: fexdo.h $w0, $w0, $w0
; MIPS64-N32-NEXT: copy_u.h $2, $w0[0]
; MIPS64-N32-NEXT: jr $ra
; MIPS64-N32-NEXT: sh $2, 0($1)
;
; MIPS64-N64-LABEL: fsub:
; MIPS64-N64: # %bb.0: # %entry
; MIPS64-N64-NEXT: lui $1, %hi(%neg(%gp_rel(fsub)))
; MIPS64-N64-NEXT: daddu $1, $1, $25
; MIPS64-N64-NEXT: daddiu $1, $1, %lo(%neg(%gp_rel(fsub)))
; MIPS64-N64-NEXT: ld $1, %got_disp(g)($1)
; MIPS64-N64-NEXT: lh $2, 0($1)
; MIPS64-N64-NEXT: fill.h $w0, $2
; MIPS64-N64-NEXT: fexupr.w $w0, $w0
; MIPS64-N64-NEXT: copy_s.w $2, $w0[0]
; MIPS64-N64-NEXT: mtc1 $2, $f0
; MIPS64-N64-NEXT: sub.s $f0, $f0, $f0
; MIPS64-N64-NEXT: mfc1 $2, $f0
; MIPS64-N64-NEXT: fill.w $w0, $2
; MIPS64-N64-NEXT: fexdo.h $w0, $w0, $w0
; MIPS64-N64-NEXT: copy_u.h $2, $w0[0]
; MIPS64-N64-NEXT: jr $ra
; MIPS64-N64-NEXT: sh $2, 0($1)
entry:
%0 = load i16, i16* @g, align 2
%1 = call float @llvm.convert.from.fp16.f32(i16 %0)
%2 = load i16, i16* @g, align 2
%3 = call float @llvm.convert.from.fp16.f32(i16 %2)
%sub = fsub float %1, %3
%4 = call i16 @llvm.convert.to.fp16.f32(float %sub)
store i16 %4, i16* @g, align 2
ret void
}
define void @fmult() {
; MIPS32-LABEL: fmult:
; MIPS32: # %bb.0: # %entry
; MIPS32-NEXT: lui $2, %hi(_gp_disp)
; MIPS32-NEXT: addiu $2, $2, %lo(_gp_disp)
; MIPS32-NEXT: addu $1, $2, $25
; MIPS32-NEXT: lw $1, %got(g)($1)
; MIPS32-NEXT: lh $2, 0($1)
; MIPS32-NEXT: fill.h $w0, $2
; MIPS32-NEXT: fexupr.w $w0, $w0
; MIPS32-NEXT: copy_s.w $2, $w0[0]
; MIPS32-NEXT: mtc1 $2, $f0
; MIPS32-NEXT: mul.s $f0, $f0, $f0
; MIPS32-NEXT: mfc1 $2, $f0
; MIPS32-NEXT: fill.w $w0, $2
; MIPS32-NEXT: fexdo.h $w0, $w0, $w0
; MIPS32-NEXT: copy_u.h $2, $w0[0]
; MIPS32-NEXT: jr $ra
; MIPS32-NEXT: sh $2, 0($1)
;
; MIPS64-N32-LABEL: fmult:
; MIPS64-N32: # %bb.0: # %entry
; MIPS64-N32-NEXT: lui $1, %hi(%neg(%gp_rel(fmult)))
; MIPS64-N32-NEXT: addu $1, $1, $25
; MIPS64-N32-NEXT: addiu $1, $1, %lo(%neg(%gp_rel(fmult)))
; MIPS64-N32-NEXT: lw $1, %got_disp(g)($1)
; MIPS64-N32-NEXT: lh $2, 0($1)
; MIPS64-N32-NEXT: fill.h $w0, $2
; MIPS64-N32-NEXT: fexupr.w $w0, $w0
; MIPS64-N32-NEXT: copy_s.w $2, $w0[0]
; MIPS64-N32-NEXT: mtc1 $2, $f0
; MIPS64-N32-NEXT: mul.s $f0, $f0, $f0
; MIPS64-N32-NEXT: mfc1 $2, $f0
; MIPS64-N32-NEXT: fill.w $w0, $2
; MIPS64-N32-NEXT: fexdo.h $w0, $w0, $w0
; MIPS64-N32-NEXT: copy_u.h $2, $w0[0]
; MIPS64-N32-NEXT: jr $ra
; MIPS64-N32-NEXT: sh $2, 0($1)
;
; MIPS64-N64-LABEL: fmult:
; MIPS64-N64: # %bb.0: # %entry
; MIPS64-N64-NEXT: lui $1, %hi(%neg(%gp_rel(fmult)))
; MIPS64-N64-NEXT: daddu $1, $1, $25
; MIPS64-N64-NEXT: daddiu $1, $1, %lo(%neg(%gp_rel(fmult)))
; MIPS64-N64-NEXT: ld $1, %got_disp(g)($1)
; MIPS64-N64-NEXT: lh $2, 0($1)
; MIPS64-N64-NEXT: fill.h $w0, $2
; MIPS64-N64-NEXT: fexupr.w $w0, $w0
; MIPS64-N64-NEXT: copy_s.w $2, $w0[0]
; MIPS64-N64-NEXT: mtc1 $2, $f0
; MIPS64-N64-NEXT: mul.s $f0, $f0, $f0
; MIPS64-N64-NEXT: mfc1 $2, $f0
; MIPS64-N64-NEXT: fill.w $w0, $2
; MIPS64-N64-NEXT: fexdo.h $w0, $w0, $w0
; MIPS64-N64-NEXT: copy_u.h $2, $w0[0]
; MIPS64-N64-NEXT: jr $ra
; MIPS64-N64-NEXT: sh $2, 0($1)
entry:
%0 = load i16, i16* @g, align 2
%1 = call float @llvm.convert.from.fp16.f32(i16 %0)
%2 = load i16, i16* @g, align 2
%3 = call float @llvm.convert.from.fp16.f32(i16 %2)
%mul = fmul float %1, %3
%4 = call i16 @llvm.convert.to.fp16.f32(float %mul)
store i16 %4, i16* @g, align 2
ret void
}
define void @fdiv() {
; MIPS32-LABEL: fdiv:
; MIPS32: # %bb.0: # %entry
; MIPS32-NEXT: lui $2, %hi(_gp_disp)
; MIPS32-NEXT: addiu $2, $2, %lo(_gp_disp)
; MIPS32-NEXT: addu $1, $2, $25
; MIPS32-NEXT: lw $1, %got(g)($1)
; MIPS32-NEXT: lh $2, 0($1)
; MIPS32-NEXT: fill.h $w0, $2
; MIPS32-NEXT: fexupr.w $w0, $w0
; MIPS32-NEXT: copy_s.w $2, $w0[0]
; MIPS32-NEXT: mtc1 $2, $f0
; MIPS32-NEXT: div.s $f0, $f0, $f0
; MIPS32-NEXT: mfc1 $2, $f0
; MIPS32-NEXT: fill.w $w0, $2
; MIPS32-NEXT: fexdo.h $w0, $w0, $w0
; MIPS32-NEXT: copy_u.h $2, $w0[0]
; MIPS32-NEXT: jr $ra
; MIPS32-NEXT: sh $2, 0($1)
;
; MIPS64-N32-LABEL: fdiv:
; MIPS64-N32: # %bb.0: # %entry
; MIPS64-N32-NEXT: lui $1, %hi(%neg(%gp_rel(fdiv)))
; MIPS64-N32-NEXT: addu $1, $1, $25
; MIPS64-N32-NEXT: addiu $1, $1, %lo(%neg(%gp_rel(fdiv)))
; MIPS64-N32-NEXT: lw $1, %got_disp(g)($1)
; MIPS64-N32-NEXT: lh $2, 0($1)
; MIPS64-N32-NEXT: fill.h $w0, $2
; MIPS64-N32-NEXT: fexupr.w $w0, $w0
; MIPS64-N32-NEXT: copy_s.w $2, $w0[0]
; MIPS64-N32-NEXT: mtc1 $2, $f0
; MIPS64-N32-NEXT: div.s $f0, $f0, $f0
; MIPS64-N32-NEXT: mfc1 $2, $f0
; MIPS64-N32-NEXT: fill.w $w0, $2
; MIPS64-N32-NEXT: fexdo.h $w0, $w0, $w0
; MIPS64-N32-NEXT: copy_u.h $2, $w0[0]
; MIPS64-N32-NEXT: jr $ra
; MIPS64-N32-NEXT: sh $2, 0($1)
;
; MIPS64-N64-LABEL: fdiv:
; MIPS64-N64: # %bb.0: # %entry
; MIPS64-N64-NEXT: lui $1, %hi(%neg(%gp_rel(fdiv)))
; MIPS64-N64-NEXT: daddu $1, $1, $25
; MIPS64-N64-NEXT: daddiu $1, $1, %lo(%neg(%gp_rel(fdiv)))
; MIPS64-N64-NEXT: ld $1, %got_disp(g)($1)
; MIPS64-N64-NEXT: lh $2, 0($1)
; MIPS64-N64-NEXT: fill.h $w0, $2
; MIPS64-N64-NEXT: fexupr.w $w0, $w0
; MIPS64-N64-NEXT: copy_s.w $2, $w0[0]
; MIPS64-N64-NEXT: mtc1 $2, $f0
; MIPS64-N64-NEXT: div.s $f0, $f0, $f0
; MIPS64-N64-NEXT: mfc1 $2, $f0
; MIPS64-N64-NEXT: fill.w $w0, $2
; MIPS64-N64-NEXT: fexdo.h $w0, $w0, $w0
; MIPS64-N64-NEXT: copy_u.h $2, $w0[0]
; MIPS64-N64-NEXT: jr $ra
; MIPS64-N64-NEXT: sh $2, 0($1)
entry:
%0 = load i16, i16* @g, align 2
%1 = call float @llvm.convert.from.fp16.f32(i16 %0)
%2 = load i16, i16* @g, align 2
%3 = call float @llvm.convert.from.fp16.f32(i16 %2)
%div = fdiv float %1, %3
%4 = call i16 @llvm.convert.to.fp16.f32(float %div)
store i16 %4, i16* @g, align 2
ret void
}
define void @frem() {
; MIPS32-LABEL: frem:
; MIPS32: # %bb.0: # %entry
; MIPS32-NEXT: lui $2, %hi(_gp_disp)
; MIPS32-NEXT: addiu $2, $2, %lo(_gp_disp)
; MIPS32-NEXT: addiu $sp, $sp, -24
; MIPS32-NEXT: .cfi_def_cfa_offset 24
; MIPS32-NEXT: sw $ra, 20($sp) # 4-byte Folded Spill
; MIPS32-NEXT: sw $16, 16($sp) # 4-byte Folded Spill
; MIPS32-NEXT: .cfi_offset 31, -4
; MIPS32-NEXT: .cfi_offset 16, -8
; MIPS32-NEXT: addu $gp, $2, $25
; MIPS32-NEXT: lw $16, %got(g)($gp)
; MIPS32-NEXT: lh $1, 0($16)
; MIPS32-NEXT: fill.h $w0, $1
; MIPS32-NEXT: fexupr.w $w0, $w0
; MIPS32-NEXT: copy_s.w $1, $w0[0]
; MIPS32-NEXT: mtc1 $1, $f12
; MIPS32-NEXT: lw $25, %call16(fmodf)($gp)
; MIPS32-NEXT: jalr $25
; MIPS32-NEXT: mov.s $f14, $f12
; MIPS32-NEXT: mfc1 $1, $f0
; MIPS32-NEXT: fill.w $w0, $1
; MIPS32-NEXT: fexdo.h $w0, $w0, $w0
; MIPS32-NEXT: copy_u.h $1, $w0[0]
; MIPS32-NEXT: sh $1, 0($16)
; MIPS32-NEXT: lw $16, 16($sp) # 4-byte Folded Reload
; MIPS32-NEXT: lw $ra, 20($sp) # 4-byte Folded Reload
; MIPS32-NEXT: jr $ra
; MIPS32-NEXT: addiu $sp, $sp, 24
;
; MIPS64-N32-LABEL: frem:
; MIPS64-N32: # %bb.0: # %entry
; MIPS64-N32-NEXT: addiu $sp, $sp, -32
; MIPS64-N32-NEXT: .cfi_def_cfa_offset 32
; MIPS64-N32-NEXT: sd $ra, 24($sp) # 8-byte Folded Spill
; MIPS64-N32-NEXT: sd $gp, 16($sp) # 8-byte Folded Spill
; MIPS64-N32-NEXT: sd $16, 8($sp) # 8-byte Folded Spill
; MIPS64-N32-NEXT: .cfi_offset 31, -8
; MIPS64-N32-NEXT: .cfi_offset 28, -16
; MIPS64-N32-NEXT: .cfi_offset 16, -24
; MIPS64-N32-NEXT: lui $1, %hi(%neg(%gp_rel(frem)))
; MIPS64-N32-NEXT: addu $1, $1, $25
; MIPS64-N32-NEXT: addiu $gp, $1, %lo(%neg(%gp_rel(frem)))
; MIPS64-N32-NEXT: lw $16, %got_disp(g)($gp)
; MIPS64-N32-NEXT: lh $1, 0($16)
; MIPS64-N32-NEXT: fill.h $w0, $1
; MIPS64-N32-NEXT: fexupr.w $w0, $w0
; MIPS64-N32-NEXT: copy_s.w $1, $w0[0]
; MIPS64-N32-NEXT: mtc1 $1, $f12
; MIPS64-N32-NEXT: lw $25, %call16(fmodf)($gp)
; MIPS64-N32-NEXT: jalr $25
; MIPS64-N32-NEXT: mov.s $f13, $f12
; MIPS64-N32-NEXT: mfc1 $1, $f0
; MIPS64-N32-NEXT: fill.w $w0, $1
; MIPS64-N32-NEXT: fexdo.h $w0, $w0, $w0
; MIPS64-N32-NEXT: copy_u.h $1, $w0[0]
; MIPS64-N32-NEXT: sh $1, 0($16)
; MIPS64-N32-NEXT: ld $16, 8($sp) # 8-byte Folded Reload
; MIPS64-N32-NEXT: ld $gp, 16($sp) # 8-byte Folded Reload
; MIPS64-N32-NEXT: ld $ra, 24($sp) # 8-byte Folded Reload
; MIPS64-N32-NEXT: jr $ra
; MIPS64-N32-NEXT: addiu $sp, $sp, 32
;
; MIPS64-N64-LABEL: frem:
; MIPS64-N64: # %bb.0: # %entry
; MIPS64-N64-NEXT: daddiu $sp, $sp, -32
; MIPS64-N64-NEXT: .cfi_def_cfa_offset 32
; MIPS64-N64-NEXT: sd $ra, 24($sp) # 8-byte Folded Spill
; MIPS64-N64-NEXT: sd $gp, 16($sp) # 8-byte Folded Spill
; MIPS64-N64-NEXT: sd $16, 8($sp) # 8-byte Folded Spill
; MIPS64-N64-NEXT: .cfi_offset 31, -8
; MIPS64-N64-NEXT: .cfi_offset 28, -16
; MIPS64-N64-NEXT: .cfi_offset 16, -24
; MIPS64-N64-NEXT: lui $1, %hi(%neg(%gp_rel(frem)))
; MIPS64-N64-NEXT: daddu $1, $1, $25
; MIPS64-N64-NEXT: daddiu $gp, $1, %lo(%neg(%gp_rel(frem)))
; MIPS64-N64-NEXT: ld $16, %got_disp(g)($gp)
; MIPS64-N64-NEXT: lh $1, 0($16)
; MIPS64-N64-NEXT: fill.h $w0, $1
; MIPS64-N64-NEXT: fexupr.w $w0, $w0
; MIPS64-N64-NEXT: copy_s.w $1, $w0[0]
; MIPS64-N64-NEXT: mtc1 $1, $f12
; MIPS64-N64-NEXT: ld $25, %call16(fmodf)($gp)
; MIPS64-N64-NEXT: jalr $25
; MIPS64-N64-NEXT: mov.s $f13, $f12
; MIPS64-N64-NEXT: mfc1 $1, $f0
; MIPS64-N64-NEXT: fill.w $w0, $1
; MIPS64-N64-NEXT: fexdo.h $w0, $w0, $w0
; MIPS64-N64-NEXT: copy_u.h $1, $w0[0]
; MIPS64-N64-NEXT: sh $1, 0($16)
; MIPS64-N64-NEXT: ld $16, 8($sp) # 8-byte Folded Reload
; MIPS64-N64-NEXT: ld $gp, 16($sp) # 8-byte Folded Reload
; MIPS64-N64-NEXT: ld $ra, 24($sp) # 8-byte Folded Reload
; MIPS64-N64-NEXT: jr $ra
; MIPS64-N64-NEXT: daddiu $sp, $sp, 32
entry:
%0 = load i16, i16* @g, align 2
%1 = call float @llvm.convert.from.fp16.f32(i16 %0)
%2 = load i16, i16* @g, align 2
%3 = call float @llvm.convert.from.fp16.f32(i16 %2)
%rem = frem float %1, %3
%4 = call i16 @llvm.convert.to.fp16.f32(float %rem)
store i16 %4, i16* @g, align 2
ret void
}
@i1 = external global i16, align 1
define void @fcmp() {
; MIPS32-O32-LABEL: fcmp:
; MIPS32-O32: # %bb.0: # %entry
; MIPS32-O32-NEXT: lui $2, %hi(_gp_disp)
; MIPS32-O32-NEXT: addiu $2, $2, %lo(_gp_disp)
; MIPS32-O32-NEXT: addu $1, $2, $25
; MIPS32-O32-NEXT: lw $2, %got(g)($1)
; MIPS32-O32-NEXT: lh $2, 0($2)
; MIPS32-O32-NEXT: fill.h $w0, $2
; MIPS32-O32-NEXT: fexupr.w $w0, $w0
; MIPS32-O32-NEXT: copy_s.w $2, $w0[0]
; MIPS32-O32-NEXT: mtc1 $2, $f0
; MIPS32-O32-NEXT: addiu $2, $zero, 1
; MIPS32-O32-NEXT: c.un.s $f0, $f0
; MIPS32-O32-NEXT: movt $2, $zero, $fcc0
; MIPS32-O32-NEXT: lw $1, %got(i1)($1)
; MIPS32-O32-NEXT: jr $ra
; MIPS32-O32-NEXT: sh $2, 0($1)
;
; MIPS64R5-N32-LABEL: fcmp:
; MIPS64R5-N32: # %bb.0: # %entry
; MIPS64R5-N32-NEXT: lui $1, %hi(%neg(%gp_rel(fcmp)))
; MIPS64R5-N32-NEXT: addu $1, $1, $25
; MIPS64R5-N32-NEXT: addiu $1, $1, %lo(%neg(%gp_rel(fcmp)))
; MIPS64R5-N32-NEXT: lw $2, %got_disp(g)($1)
; MIPS64R5-N32-NEXT: lh $2, 0($2)
; MIPS64R5-N32-NEXT: fill.h $w0, $2
; MIPS64R5-N32-NEXT: fexupr.w $w0, $w0
; MIPS64R5-N32-NEXT: copy_s.w $2, $w0[0]
; MIPS64R5-N32-NEXT: mtc1 $2, $f0
; MIPS64R5-N32-NEXT: addiu $2, $zero, 1
; MIPS64R5-N32-NEXT: c.un.s $f0, $f0
; MIPS64R5-N32-NEXT: movt $2, $zero, $fcc0
; MIPS64R5-N32-NEXT: lw $1, %got_disp(i1)($1)
; MIPS64R5-N32-NEXT: jr $ra
; MIPS64R5-N32-NEXT: sh $2, 0($1)
;
; MIPS64R5-N64-LABEL: fcmp:
; MIPS64R5-N64: # %bb.0: # %entry
; MIPS64R5-N64-NEXT: lui $1, %hi(%neg(%gp_rel(fcmp)))
; MIPS64R5-N64-NEXT: daddu $1, $1, $25
; MIPS64R5-N64-NEXT: daddiu $1, $1, %lo(%neg(%gp_rel(fcmp)))
; MIPS64R5-N64-NEXT: ld $2, %got_disp(g)($1)
; MIPS64R5-N64-NEXT: lh $2, 0($2)
; MIPS64R5-N64-NEXT: fill.h $w0, $2
; MIPS64R5-N64-NEXT: fexupr.w $w0, $w0
; MIPS64R5-N64-NEXT: copy_s.w $2, $w0[0]
; MIPS64R5-N64-NEXT: mtc1 $2, $f0
; MIPS64R5-N64-NEXT: addiu $2, $zero, 1
; MIPS64R5-N64-NEXT: c.un.s $f0, $f0
; MIPS64R5-N64-NEXT: movt $2, $zero, $fcc0
; MIPS64R5-N64-NEXT: ld $1, %got_disp(i1)($1)
; MIPS64R5-N64-NEXT: jr $ra
; MIPS64R5-N64-NEXT: sh $2, 0($1)
;
; MIPSR6-O32-LABEL: fcmp:
; MIPSR6-O32: # %bb.0: # %entry
; MIPSR6-O32-NEXT: lui $2, %hi(_gp_disp)
; MIPSR6-O32-NEXT: addiu $2, $2, %lo(_gp_disp)
; MIPSR6-O32-NEXT: addu $1, $2, $25
; MIPSR6-O32-NEXT: lw $2, %got(g)($1)
; MIPSR6-O32-NEXT: lh $2, 0($2)
; MIPSR6-O32-NEXT: fill.h $w0, $2
; MIPSR6-O32-NEXT: fexupr.w $w0, $w0
; MIPSR6-O32-NEXT: copy_s.w $2, $w0[0]
; MIPSR6-O32-NEXT: mtc1 $2, $f0
; MIPSR6-O32-NEXT: cmp.un.s $f0, $f0, $f0
; MIPSR6-O32-NEXT: mfc1 $2, $f0
; MIPSR6-O32-NEXT: not $2, $2
; MIPSR6-O32-NEXT: andi $2, $2, 1
; MIPSR6-O32-NEXT: lw $1, %got(i1)($1)
; MIPSR6-O32-NEXT: jr $ra
; MIPSR6-O32-NEXT: sh $2, 0($1)
;
; MIPSR6-N32-LABEL: fcmp:
; MIPSR6-N32: # %bb.0: # %entry
; MIPSR6-N32-NEXT: lui $1, %hi(%neg(%gp_rel(fcmp)))
; MIPSR6-N32-NEXT: addu $1, $1, $25
; MIPSR6-N32-NEXT: addiu $1, $1, %lo(%neg(%gp_rel(fcmp)))
; MIPSR6-N32-NEXT: lw $2, %got_disp(g)($1)
; MIPSR6-N32-NEXT: lh $2, 0($2)
; MIPSR6-N32-NEXT: fill.h $w0, $2
; MIPSR6-N32-NEXT: fexupr.w $w0, $w0
; MIPSR6-N32-NEXT: copy_s.w $2, $w0[0]
; MIPSR6-N32-NEXT: mtc1 $2, $f0
; MIPSR6-N32-NEXT: cmp.un.s $f0, $f0, $f0
; MIPSR6-N32-NEXT: mfc1 $2, $f0
; MIPSR6-N32-NEXT: not $2, $2
; MIPSR6-N32-NEXT: andi $2, $2, 1
; MIPSR6-N32-NEXT: lw $1, %got_disp(i1)($1)
; MIPSR6-N32-NEXT: jr $ra
; MIPSR6-N32-NEXT: sh $2, 0($1)
;
; MIPSR6-N64-LABEL: fcmp:
; MIPSR6-N64: # %bb.0: # %entry
; MIPSR6-N64-NEXT: lui $1, %hi(%neg(%gp_rel(fcmp)))
; MIPSR6-N64-NEXT: daddu $1, $1, $25
; MIPSR6-N64-NEXT: daddiu $1, $1, %lo(%neg(%gp_rel(fcmp)))
; MIPSR6-N64-NEXT: ld $2, %got_disp(g)($1)
; MIPSR6-N64-NEXT: lh $2, 0($2)
; MIPSR6-N64-NEXT: fill.h $w0, $2
; MIPSR6-N64-NEXT: fexupr.w $w0, $w0
; MIPSR6-N64-NEXT: copy_s.w $2, $w0[0]
; MIPSR6-N64-NEXT: mtc1 $2, $f0
; MIPSR6-N64-NEXT: cmp.un.s $f0, $f0, $f0
; MIPSR6-N64-NEXT: mfc1 $2, $f0
; MIPSR6-N64-NEXT: not $2, $2
; MIPSR6-N64-NEXT: andi $2, $2, 1
; MIPSR6-N64-NEXT: ld $1, %got_disp(i1)($1)
; MIPSR6-N64-NEXT: jr $ra
; MIPSR6-N64-NEXT: sh $2, 0($1)
entry:
%0 = load i16, i16* @g, align 2
%1 = call float @llvm.convert.from.fp16.f32(i16 %0)
%2 = load i16, i16* @g, align 2
%3 = call float @llvm.convert.from.fp16.f32(i16 %2)
%fcmp = fcmp oeq float %1, %3
%4 = zext i1 %fcmp to i16
store i16 %4, i16* @i1, align 2
ret void
}
declare float @llvm.powi.f32.i32(float, i32)
define void @fpowi() {
; MIPS32-LABEL: fpowi:
; MIPS32: # %bb.0: # %entry
; MIPS32-NEXT: lui $2, %hi(_gp_disp)
; MIPS32-NEXT: addiu $2, $2, %lo(_gp_disp)
; MIPS32-NEXT: addu $1, $2, $25
; MIPS32-NEXT: lw $1, %got(g)($1)
; MIPS32-NEXT: lh $2, 0($1)
; MIPS32-NEXT: fill.h $w0, $2
; MIPS32-NEXT: fexupr.w $w0, $w0
; MIPS32-NEXT: copy_s.w $2, $w0[0]
; MIPS32-NEXT: mtc1 $2, $f0
; MIPS32-NEXT: mul.s $f0, $f0, $f0
; MIPS32-NEXT: mfc1 $2, $f0
; MIPS32-NEXT: fill.w $w0, $2
; MIPS32-NEXT: fexdo.h $w0, $w0, $w0
; MIPS32-NEXT: copy_u.h $2, $w0[0]
; MIPS32-NEXT: jr $ra
; MIPS32-NEXT: sh $2, 0($1)
;
; MIPS64-N32-LABEL: fpowi:
; MIPS64-N32: # %bb.0: # %entry
; MIPS64-N32-NEXT: lui $1, %hi(%neg(%gp_rel(fpowi)))
; MIPS64-N32-NEXT: addu $1, $1, $25
; MIPS64-N32-NEXT: addiu $1, $1, %lo(%neg(%gp_rel(fpowi)))
; MIPS64-N32-NEXT: lw $1, %got_disp(g)($1)
; MIPS64-N32-NEXT: lh $2, 0($1)
; MIPS64-N32-NEXT: fill.h $w0, $2
; MIPS64-N32-NEXT: fexupr.w $w0, $w0
; MIPS64-N32-NEXT: copy_s.w $2, $w0[0]
; MIPS64-N32-NEXT: mtc1 $2, $f0
; MIPS64-N32-NEXT: mul.s $f0, $f0, $f0
; MIPS64-N32-NEXT: mfc1 $2, $f0
; MIPS64-N32-NEXT: fill.w $w0, $2
; MIPS64-N32-NEXT: fexdo.h $w0, $w0, $w0
; MIPS64-N32-NEXT: copy_u.h $2, $w0[0]
; MIPS64-N32-NEXT: jr $ra
; MIPS64-N32-NEXT: sh $2, 0($1)
;
; MIPS64-N64-LABEL: fpowi:
; MIPS64-N64: # %bb.0: # %entry
; MIPS64-N64-NEXT: lui $1, %hi(%neg(%gp_rel(fpowi)))
; MIPS64-N64-NEXT: daddu $1, $1, $25
; MIPS64-N64-NEXT: daddiu $1, $1, %lo(%neg(%gp_rel(fpowi)))
; MIPS64-N64-NEXT: ld $1, %got_disp(g)($1)
; MIPS64-N64-NEXT: lh $2, 0($1)
; MIPS64-N64-NEXT: fill.h $w0, $2
; MIPS64-N64-NEXT: fexupr.w $w0, $w0
; MIPS64-N64-NEXT: copy_s.w $2, $w0[0]
; MIPS64-N64-NEXT: mtc1 $2, $f0
; MIPS64-N64-NEXT: mul.s $f0, $f0, $f0
; MIPS64-N64-NEXT: mfc1 $2, $f0
; MIPS64-N64-NEXT: fill.w $w0, $2
; MIPS64-N64-NEXT: fexdo.h $w0, $w0, $w0
; MIPS64-N64-NEXT: copy_u.h $2, $w0[0]
; MIPS64-N64-NEXT: jr $ra
; MIPS64-N64-NEXT: sh $2, 0($1)
entry:
%0 = load i16, i16* @g, align 2
%1 = call float @llvm.convert.from.fp16.f32(i16 %0)
%powi = call float @llvm.powi.f32.i32(float %1, i32 2)
%2 = call i16 @llvm.convert.to.fp16.f32(float %powi)
store i16 %2, i16* @g, align 2
ret void
}
define void @fpowi_var(i32 %var) {
; MIPS32-LABEL: fpowi_var:
; MIPS32: # %bb.0: # %entry
; MIPS32-NEXT: lui $2, %hi(_gp_disp)
; MIPS32-NEXT: addiu $2, $2, %lo(_gp_disp)
; MIPS32-NEXT: addiu $sp, $sp, -24
; MIPS32-NEXT: .cfi_def_cfa_offset 24
; MIPS32-NEXT: sw $ra, 20($sp) # 4-byte Folded Spill
; MIPS32-NEXT: sw $16, 16($sp) # 4-byte Folded Spill
; MIPS32-NEXT: .cfi_offset 31, -4
; MIPS32-NEXT: .cfi_offset 16, -8
; MIPS32-NEXT: addu $gp, $2, $25
; MIPS32-NEXT: lw $16, %got(g)($gp)
; MIPS32-NEXT: lh $1, 0($16)
; MIPS32-NEXT: fill.h $w0, $1
; MIPS32-NEXT: fexupr.w $w0, $w0
; MIPS32-NEXT: copy_s.w $1, $w0[0]
; MIPS32-NEXT: mtc1 $1, $f12
; MIPS32-NEXT: lw $25, %call16(__powisf2)($gp)
; MIPS32-NEXT: jalr $25
; MIPS32-NEXT: move $5, $4
; MIPS32-NEXT: mfc1 $1, $f0
; MIPS32-NEXT: fill.w $w0, $1
; MIPS32-NEXT: fexdo.h $w0, $w0, $w0
; MIPS32-NEXT: copy_u.h $1, $w0[0]
; MIPS32-NEXT: sh $1, 0($16)
; MIPS32-NEXT: lw $16, 16($sp) # 4-byte Folded Reload
; MIPS32-NEXT: lw $ra, 20($sp) # 4-byte Folded Reload
; MIPS32-NEXT: jr $ra
; MIPS32-NEXT: addiu $sp, $sp, 24
;
; MIPS64-N32-LABEL: fpowi_var:
; MIPS64-N32: # %bb.0: # %entry
; MIPS64-N32-NEXT: addiu $sp, $sp, -32
; MIPS64-N32-NEXT: .cfi_def_cfa_offset 32
; MIPS64-N32-NEXT: sd $ra, 24($sp) # 8-byte Folded Spill
; MIPS64-N32-NEXT: sd $gp, 16($sp) # 8-byte Folded Spill
; MIPS64-N32-NEXT: sd $16, 8($sp) # 8-byte Folded Spill
; MIPS64-N32-NEXT: .cfi_offset 31, -8
; MIPS64-N32-NEXT: .cfi_offset 28, -16
; MIPS64-N32-NEXT: .cfi_offset 16, -24
; MIPS64-N32-NEXT: lui $1, %hi(%neg(%gp_rel(fpowi_var)))
; MIPS64-N32-NEXT: addu $1, $1, $25
; MIPS64-N32-NEXT: addiu $gp, $1, %lo(%neg(%gp_rel(fpowi_var)))
; MIPS64-N32-NEXT: sll $5, $4, 0
; MIPS64-N32-NEXT: lw $16, %got_disp(g)($gp)
; MIPS64-N32-NEXT: lh $1, 0($16)
; MIPS64-N32-NEXT: fill.h $w0, $1
; MIPS64-N32-NEXT: fexupr.w $w0, $w0
; MIPS64-N32-NEXT: copy_s.w $1, $w0[0]
; MIPS64-N32-NEXT: lw $25, %call16(__powisf2)($gp)
; MIPS64-N32-NEXT: jalr $25
; MIPS64-N32-NEXT: mtc1 $1, $f12
; MIPS64-N32-NEXT: mfc1 $1, $f0
; MIPS64-N32-NEXT: fill.w $w0, $1
; MIPS64-N32-NEXT: fexdo.h $w0, $w0, $w0
; MIPS64-N32-NEXT: copy_u.h $1, $w0[0]
; MIPS64-N32-NEXT: sh $1, 0($16)
; MIPS64-N32-NEXT: ld $16, 8($sp) # 8-byte Folded Reload
; MIPS64-N32-NEXT: ld $gp, 16($sp) # 8-byte Folded Reload
; MIPS64-N32-NEXT: ld $ra, 24($sp) # 8-byte Folded Reload
; MIPS64-N32-NEXT: jr $ra
; MIPS64-N32-NEXT: addiu $sp, $sp, 32
;
; MIPS64-N64-LABEL: fpowi_var:
; MIPS64-N64: # %bb.0: # %entry
; MIPS64-N64-NEXT: daddiu $sp, $sp, -32
; MIPS64-N64-NEXT: .cfi_def_cfa_offset 32
; MIPS64-N64-NEXT: sd $ra, 24($sp) # 8-byte Folded Spill
; MIPS64-N64-NEXT: sd $gp, 16($sp) # 8-byte Folded Spill
; MIPS64-N64-NEXT: sd $16, 8($sp) # 8-byte Folded Spill
; MIPS64-N64-NEXT: .cfi_offset 31, -8
; MIPS64-N64-NEXT: .cfi_offset 28, -16
; MIPS64-N64-NEXT: .cfi_offset 16, -24
; MIPS64-N64-NEXT: lui $1, %hi(%neg(%gp_rel(fpowi_var)))
; MIPS64-N64-NEXT: daddu $1, $1, $25
; MIPS64-N64-NEXT: daddiu $gp, $1, %lo(%neg(%gp_rel(fpowi_var)))
; MIPS64-N64-NEXT: sll $5, $4, 0
; MIPS64-N64-NEXT: ld $16, %got_disp(g)($gp)
; MIPS64-N64-NEXT: lh $1, 0($16)
; MIPS64-N64-NEXT: fill.h $w0, $1
; MIPS64-N64-NEXT: fexupr.w $w0, $w0
; MIPS64-N64-NEXT: copy_s.w $1, $w0[0]
; MIPS64-N64-NEXT: ld $25, %call16(__powisf2)($gp)
; MIPS64-N64-NEXT: jalr $25
; MIPS64-N64-NEXT: mtc1 $1, $f12
; MIPS64-N64-NEXT: mfc1 $1, $f0
; MIPS64-N64-NEXT: fill.w $w0, $1
; MIPS64-N64-NEXT: fexdo.h $w0, $w0, $w0
; MIPS64-N64-NEXT: copy_u.h $1, $w0[0]
; MIPS64-N64-NEXT: sh $1, 0($16)
; MIPS64-N64-NEXT: ld $16, 8($sp) # 8-byte Folded Reload
; MIPS64-N64-NEXT: ld $gp, 16($sp) # 8-byte Folded Reload
; MIPS64-N64-NEXT: ld $ra, 24($sp) # 8-byte Folded Reload
; MIPS64-N64-NEXT: jr $ra
; MIPS64-N64-NEXT: daddiu $sp, $sp, 32
entry:
%0 = load i16, i16* @g, align 2
%1 = call float @llvm.convert.from.fp16.f32(i16 %0)
%powi = call float @llvm.powi.f32.i32(float %1, i32 %var)
%2 = call i16 @llvm.convert.to.fp16.f32(float %powi)
store i16 %2, i16* @g, align 2
ret void
}
declare float @llvm.pow.f32(float %Val, float %power)
define void @fpow(float %var) {
; MIPS32-LABEL: fpow:
; MIPS32: # %bb.0: # %entry
; MIPS32-NEXT: lui $2, %hi(_gp_disp)
; MIPS32-NEXT: addiu $2, $2, %lo(_gp_disp)
; MIPS32-NEXT: addiu $sp, $sp, -24
; MIPS32-NEXT: .cfi_def_cfa_offset 24
; MIPS32-NEXT: sw $ra, 20($sp) # 4-byte Folded Spill
; MIPS32-NEXT: sw $16, 16($sp) # 4-byte Folded Spill
; MIPS32-NEXT: .cfi_offset 31, -4
; MIPS32-NEXT: .cfi_offset 16, -8
; MIPS32-NEXT: addu $gp, $2, $25
; MIPS32-NEXT: mov.s $f14, $f12
; MIPS32-NEXT: lw $16, %got(g)($gp)
; MIPS32-NEXT: lh $1, 0($16)
; MIPS32-NEXT: fill.h $w0, $1
; MIPS32-NEXT: fexupr.w $w0, $w0
; MIPS32-NEXT: copy_s.w $1, $w0[0]
; MIPS32-NEXT: lw $25, %call16(powf)($gp)
; MIPS32-NEXT: jalr $25
; MIPS32-NEXT: mtc1 $1, $f12
; MIPS32-NEXT: mfc1 $1, $f0
; MIPS32-NEXT: fill.w $w0, $1
; MIPS32-NEXT: fexdo.h $w0, $w0, $w0
; MIPS32-NEXT: copy_u.h $1, $w0[0]
; MIPS32-NEXT: sh $1, 0($16)
; MIPS32-NEXT: lw $16, 16($sp) # 4-byte Folded Reload
; MIPS32-NEXT: lw $ra, 20($sp) # 4-byte Folded Reload
; MIPS32-NEXT: jr $ra
; MIPS32-NEXT: addiu $sp, $sp, 24
;
; MIPS64-N32-LABEL: fpow:
; MIPS64-N32: # %bb.0: # %entry
; MIPS64-N32-NEXT: addiu $sp, $sp, -32
; MIPS64-N32-NEXT: .cfi_def_cfa_offset 32
; MIPS64-N32-NEXT: sd $ra, 24($sp) # 8-byte Folded Spill
; MIPS64-N32-NEXT: sd $gp, 16($sp) # 8-byte Folded Spill
; MIPS64-N32-NEXT: sd $16, 8($sp) # 8-byte Folded Spill
; MIPS64-N32-NEXT: .cfi_offset 31, -8
; MIPS64-N32-NEXT: .cfi_offset 28, -16
; MIPS64-N32-NEXT: .cfi_offset 16, -24
; MIPS64-N32-NEXT: lui $1, %hi(%neg(%gp_rel(fpow)))
; MIPS64-N32-NEXT: addu $1, $1, $25
; MIPS64-N32-NEXT: addiu $gp, $1, %lo(%neg(%gp_rel(fpow)))
; MIPS64-N32-NEXT: mov.s $f13, $f12
; MIPS64-N32-NEXT: lw $16, %got_disp(g)($gp)
; MIPS64-N32-NEXT: lh $1, 0($16)
; MIPS64-N32-NEXT: fill.h $w0, $1
; MIPS64-N32-NEXT: fexupr.w $w0, $w0
; MIPS64-N32-NEXT: copy_s.w $1, $w0[0]
; MIPS64-N32-NEXT: lw $25, %call16(powf)($gp)
; MIPS64-N32-NEXT: jalr $25
; MIPS64-N32-NEXT: mtc1 $1, $f12
; MIPS64-N32-NEXT: mfc1 $1, $f0
; MIPS64-N32-NEXT: fill.w $w0, $1
; MIPS64-N32-NEXT: fexdo.h $w0, $w0, $w0
; MIPS64-N32-NEXT: copy_u.h $1, $w0[0]
; MIPS64-N32-NEXT: sh $1, 0($16)
; MIPS64-N32-NEXT: ld $16, 8($sp) # 8-byte Folded Reload
; MIPS64-N32-NEXT: ld $gp, 16($sp) # 8-byte Folded Reload
; MIPS64-N32-NEXT: ld $ra, 24($sp) # 8-byte Folded Reload
; MIPS64-N32-NEXT: jr $ra
; MIPS64-N32-NEXT: addiu $sp, $sp, 32
;
; MIPS64-N64-LABEL: fpow:
; MIPS64-N64: # %bb.0: # %entry
; MIPS64-N64-NEXT: daddiu $sp, $sp, -32
; MIPS64-N64-NEXT: .cfi_def_cfa_offset 32
; MIPS64-N64-NEXT: sd $ra, 24($sp) # 8-byte Folded Spill
; MIPS64-N64-NEXT: sd $gp, 16($sp) # 8-byte Folded Spill
; MIPS64-N64-NEXT: sd $16, 8($sp) # 8-byte Folded Spill
; MIPS64-N64-NEXT: .cfi_offset 31, -8
; MIPS64-N64-NEXT: .cfi_offset 28, -16
; MIPS64-N64-NEXT: .cfi_offset 16, -24
; MIPS64-N64-NEXT: lui $1, %hi(%neg(%gp_rel(fpow)))
; MIPS64-N64-NEXT: daddu $1, $1, $25
; MIPS64-N64-NEXT: daddiu $gp, $1, %lo(%neg(%gp_rel(fpow)))
; MIPS64-N64-NEXT: mov.s $f13, $f12
; MIPS64-N64-NEXT: ld $16, %got_disp(g)($gp)
; MIPS64-N64-NEXT: lh $1, 0($16)
; MIPS64-N64-NEXT: fill.h $w0, $1
; MIPS64-N64-NEXT: fexupr.w $w0, $w0
; MIPS64-N64-NEXT: copy_s.w $1, $w0[0]
; MIPS64-N64-NEXT: ld $25, %call16(powf)($gp)
; MIPS64-N64-NEXT: jalr $25
; MIPS64-N64-NEXT: mtc1 $1, $f12
; MIPS64-N64-NEXT: mfc1 $1, $f0
; MIPS64-N64-NEXT: fill.w $w0, $1
; MIPS64-N64-NEXT: fexdo.h $w0, $w0, $w0
; MIPS64-N64-NEXT: copy_u.h $1, $w0[0]
; MIPS64-N64-NEXT: sh $1, 0($16)
; MIPS64-N64-NEXT: ld $16, 8($sp) # 8-byte Folded Reload
; MIPS64-N64-NEXT: ld $gp, 16($sp) # 8-byte Folded Reload
; MIPS64-N64-NEXT: ld $ra, 24($sp) # 8-byte Folded Reload
; MIPS64-N64-NEXT: jr $ra
; MIPS64-N64-NEXT: daddiu $sp, $sp, 32
entry:
%0 = load i16, i16* @g, align 2
%1 = call float @llvm.convert.from.fp16.f32(i16 %0)
%powi = call float @llvm.pow.f32(float %1, float %var)
%2 = call i16 @llvm.convert.to.fp16.f32(float %powi)
store i16 %2, i16* @g, align 2
ret void
}
declare float @llvm.log2.f32(float %Val)
define void @flog2() {
; MIPS32-LABEL: flog2:
; MIPS32: # %bb.0: # %entry
; MIPS32-NEXT: lui $2, %hi(_gp_disp)
; MIPS32-NEXT: addiu $2, $2, %lo(_gp_disp)
; MIPS32-NEXT: addiu $sp, $sp, -24
; MIPS32-NEXT: .cfi_def_cfa_offset 24
; MIPS32-NEXT: sw $ra, 20($sp) # 4-byte Folded Spill
; MIPS32-NEXT: sw $16, 16($sp) # 4-byte Folded Spill
; MIPS32-NEXT: .cfi_offset 31, -4
; MIPS32-NEXT: .cfi_offset 16, -8
; MIPS32-NEXT: addu $gp, $2, $25
; MIPS32-NEXT: lw $16, %got(g)($gp)
; MIPS32-NEXT: lh $1, 0($16)
; MIPS32-NEXT: fill.h $w0, $1
; MIPS32-NEXT: fexupr.w $w0, $w0
; MIPS32-NEXT: copy_s.w $1, $w0[0]
; MIPS32-NEXT: lw $25, %call16(log2f)($gp)
; MIPS32-NEXT: jalr $25
; MIPS32-NEXT: mtc1 $1, $f12
; MIPS32-NEXT: mfc1 $1, $f0
; MIPS32-NEXT: fill.w $w0, $1
; MIPS32-NEXT: fexdo.h $w0, $w0, $w0
; MIPS32-NEXT: copy_u.h $1, $w0[0]
; MIPS32-NEXT: sh $1, 0($16)
; MIPS32-NEXT: lw $16, 16($sp) # 4-byte Folded Reload
; MIPS32-NEXT: lw $ra, 20($sp) # 4-byte Folded Reload
; MIPS32-NEXT: jr $ra
; MIPS32-NEXT: addiu $sp, $sp, 24
;
; MIPS64-N32-LABEL: flog2:
; MIPS64-N32: # %bb.0: # %entry
; MIPS64-N32-NEXT: addiu $sp, $sp, -32
; MIPS64-N32-NEXT: .cfi_def_cfa_offset 32
; MIPS64-N32-NEXT: sd $ra, 24($sp) # 8-byte Folded Spill
; MIPS64-N32-NEXT: sd $gp, 16($sp) # 8-byte Folded Spill
; MIPS64-N32-NEXT: sd $16, 8($sp) # 8-byte Folded Spill
; MIPS64-N32-NEXT: .cfi_offset 31, -8
; MIPS64-N32-NEXT: .cfi_offset 28, -16
; MIPS64-N32-NEXT: .cfi_offset 16, -24
; MIPS64-N32-NEXT: lui $1, %hi(%neg(%gp_rel(flog2)))
; MIPS64-N32-NEXT: addu $1, $1, $25
; MIPS64-N32-NEXT: addiu $gp, $1, %lo(%neg(%gp_rel(flog2)))
; MIPS64-N32-NEXT: lw $16, %got_disp(g)($gp)
; MIPS64-N32-NEXT: lh $1, 0($16)
; MIPS64-N32-NEXT: fill.h $w0, $1
; MIPS64-N32-NEXT: fexupr.w $w0, $w0
; MIPS64-N32-NEXT: copy_s.w $1, $w0[0]
; MIPS64-N32-NEXT: lw $25, %call16(log2f)($gp)
; MIPS64-N32-NEXT: jalr $25
; MIPS64-N32-NEXT: mtc1 $1, $f12
; MIPS64-N32-NEXT: mfc1 $1, $f0
; MIPS64-N32-NEXT: fill.w $w0, $1
; MIPS64-N32-NEXT: fexdo.h $w0, $w0, $w0
; MIPS64-N32-NEXT: copy_u.h $1, $w0[0]
; MIPS64-N32-NEXT: sh $1, 0($16)
; MIPS64-N32-NEXT: ld $16, 8($sp) # 8-byte Folded Reload
; MIPS64-N32-NEXT: ld $gp, 16($sp) # 8-byte Folded Reload
; MIPS64-N32-NEXT: ld $ra, 24($sp) # 8-byte Folded Reload
; MIPS64-N32-NEXT: jr $ra
; MIPS64-N32-NEXT: addiu $sp, $sp, 32
;
; MIPS64-N64-LABEL: flog2:
; MIPS64-N64: # %bb.0: # %entry
; MIPS64-N64-NEXT: daddiu $sp, $sp, -32
; MIPS64-N64-NEXT: .cfi_def_cfa_offset 32
; MIPS64-N64-NEXT: sd $ra, 24($sp) # 8-byte Folded Spill
; MIPS64-N64-NEXT: sd $gp, 16($sp) # 8-byte Folded Spill
; MIPS64-N64-NEXT: sd $16, 8($sp) # 8-byte Folded Spill
; MIPS64-N64-NEXT: .cfi_offset 31, -8
; MIPS64-N64-NEXT: .cfi_offset 28, -16
; MIPS64-N64-NEXT: .cfi_offset 16, -24
; MIPS64-N64-NEXT: lui $1, %hi(%neg(%gp_rel(flog2)))
; MIPS64-N64-NEXT: daddu $1, $1, $25
; MIPS64-N64-NEXT: daddiu $gp, $1, %lo(%neg(%gp_rel(flog2)))
; MIPS64-N64-NEXT: ld $16, %got_disp(g)($gp)
; MIPS64-N64-NEXT: lh $1, 0($16)
; MIPS64-N64-NEXT: fill.h $w0, $1
; MIPS64-N64-NEXT: fexupr.w $w0, $w0
; MIPS64-N64-NEXT: copy_s.w $1, $w0[0]
; MIPS64-N64-NEXT: ld $25, %call16(log2f)($gp)
; MIPS64-N64-NEXT: jalr $25
; MIPS64-N64-NEXT: mtc1 $1, $f12
; MIPS64-N64-NEXT: mfc1 $1, $f0
; MIPS64-N64-NEXT: fill.w $w0, $1
; MIPS64-N64-NEXT: fexdo.h $w0, $w0, $w0
; MIPS64-N64-NEXT: copy_u.h $1, $w0[0]
; MIPS64-N64-NEXT: sh $1, 0($16)
; MIPS64-N64-NEXT: ld $16, 8($sp) # 8-byte Folded Reload
; MIPS64-N64-NEXT: ld $gp, 16($sp) # 8-byte Folded Reload
; MIPS64-N64-NEXT: ld $ra, 24($sp) # 8-byte Folded Reload
; MIPS64-N64-NEXT: jr $ra
; MIPS64-N64-NEXT: daddiu $sp, $sp, 32
entry:
%0 = load i16, i16* @g, align 2
%1 = call float @llvm.convert.from.fp16.f32(i16 %0)
%log2 = call float @llvm.log2.f32(float %1)
%2 = call i16 @llvm.convert.to.fp16.f32(float %log2)
store i16 %2, i16* @g, align 2
ret void
}
declare float @llvm.log10.f32(float %Val)
define void @flog10() {
; MIPS32-LABEL: flog10:
; MIPS32: # %bb.0: # %entry
; MIPS32-NEXT: lui $2, %hi(_gp_disp)
; MIPS32-NEXT: addiu $2, $2, %lo(_gp_disp)
; MIPS32-NEXT: addiu $sp, $sp, -24
; MIPS32-NEXT: .cfi_def_cfa_offset 24
; MIPS32-NEXT: sw $ra, 20($sp) # 4-byte Folded Spill
; MIPS32-NEXT: sw $16, 16($sp) # 4-byte Folded Spill
; MIPS32-NEXT: .cfi_offset 31, -4
; MIPS32-NEXT: .cfi_offset 16, -8
; MIPS32-NEXT: addu $gp, $2, $25
; MIPS32-NEXT: lw $16, %got(g)($gp)
; MIPS32-NEXT: lh $1, 0($16)
; MIPS32-NEXT: fill.h $w0, $1
; MIPS32-NEXT: fexupr.w $w0, $w0
; MIPS32-NEXT: copy_s.w $1, $w0[0]
; MIPS32-NEXT: lw $25, %call16(log10f)($gp)
; MIPS32-NEXT: jalr $25
; MIPS32-NEXT: mtc1 $1, $f12
; MIPS32-NEXT: mfc1 $1, $f0
; MIPS32-NEXT: fill.w $w0, $1
; MIPS32-NEXT: fexdo.h $w0, $w0, $w0
; MIPS32-NEXT: copy_u.h $1, $w0[0]
; MIPS32-NEXT: sh $1, 0($16)
; MIPS32-NEXT: lw $16, 16($sp) # 4-byte Folded Reload
; MIPS32-NEXT: lw $ra, 20($sp) # 4-byte Folded Reload
; MIPS32-NEXT: jr $ra
; MIPS32-NEXT: addiu $sp, $sp, 24
;
; MIPS64-N32-LABEL: flog10:
; MIPS64-N32: # %bb.0: # %entry
; MIPS64-N32-NEXT: addiu $sp, $sp, -32
; MIPS64-N32-NEXT: .cfi_def_cfa_offset 32
; MIPS64-N32-NEXT: sd $ra, 24($sp) # 8-byte Folded Spill
; MIPS64-N32-NEXT: sd $gp, 16($sp) # 8-byte Folded Spill
; MIPS64-N32-NEXT: sd $16, 8($sp) # 8-byte Folded Spill
; MIPS64-N32-NEXT: .cfi_offset 31, -8
; MIPS64-N32-NEXT: .cfi_offset 28, -16
; MIPS64-N32-NEXT: .cfi_offset 16, -24
; MIPS64-N32-NEXT: lui $1, %hi(%neg(%gp_rel(flog10)))
; MIPS64-N32-NEXT: addu $1, $1, $25
; MIPS64-N32-NEXT: addiu $gp, $1, %lo(%neg(%gp_rel(flog10)))
; MIPS64-N32-NEXT: lw $16, %got_disp(g)($gp)
; MIPS64-N32-NEXT: lh $1, 0($16)
; MIPS64-N32-NEXT: fill.h $w0, $1
; MIPS64-N32-NEXT: fexupr.w $w0, $w0
; MIPS64-N32-NEXT: copy_s.w $1, $w0[0]
; MIPS64-N32-NEXT: lw $25, %call16(log10f)($gp)
; MIPS64-N32-NEXT: jalr $25
; MIPS64-N32-NEXT: mtc1 $1, $f12
; MIPS64-N32-NEXT: mfc1 $1, $f0
; MIPS64-N32-NEXT: fill.w $w0, $1
; MIPS64-N32-NEXT: fexdo.h $w0, $w0, $w0
; MIPS64-N32-NEXT: copy_u.h $1, $w0[0]
; MIPS64-N32-NEXT: sh $1, 0($16)
; MIPS64-N32-NEXT: ld $16, 8($sp) # 8-byte Folded Reload
; MIPS64-N32-NEXT: ld $gp, 16($sp) # 8-byte Folded Reload
; MIPS64-N32-NEXT: ld $ra, 24($sp) # 8-byte Folded Reload
; MIPS64-N32-NEXT: jr $ra
; MIPS64-N32-NEXT: addiu $sp, $sp, 32
;
; MIPS64-N64-LABEL: flog10:
; MIPS64-N64: # %bb.0: # %entry
; MIPS64-N64-NEXT: daddiu $sp, $sp, -32
; MIPS64-N64-NEXT: .cfi_def_cfa_offset 32
; MIPS64-N64-NEXT: sd $ra, 24($sp) # 8-byte Folded Spill
; MIPS64-N64-NEXT: sd $gp, 16($sp) # 8-byte Folded Spill
; MIPS64-N64-NEXT: sd $16, 8($sp) # 8-byte Folded Spill
; MIPS64-N64-NEXT: .cfi_offset 31, -8
; MIPS64-N64-NEXT: .cfi_offset 28, -16
; MIPS64-N64-NEXT: .cfi_offset 16, -24
; MIPS64-N64-NEXT: lui $1, %hi(%neg(%gp_rel(flog10)))
; MIPS64-N64-NEXT: daddu $1, $1, $25
; MIPS64-N64-NEXT: daddiu $gp, $1, %lo(%neg(%gp_rel(flog10)))
; MIPS64-N64-NEXT: ld $16, %got_disp(g)($gp)
; MIPS64-N64-NEXT: lh $1, 0($16)
; MIPS64-N64-NEXT: fill.h $w0, $1
; MIPS64-N64-NEXT: fexupr.w $w0, $w0
; MIPS64-N64-NEXT: copy_s.w $1, $w0[0]
; MIPS64-N64-NEXT: ld $25, %call16(log10f)($gp)
; MIPS64-N64-NEXT: jalr $25
; MIPS64-N64-NEXT: mtc1 $1, $f12
; MIPS64-N64-NEXT: mfc1 $1, $f0
; MIPS64-N64-NEXT: fill.w $w0, $1
; MIPS64-N64-NEXT: fexdo.h $w0, $w0, $w0
; MIPS64-N64-NEXT: copy_u.h $1, $w0[0]
; MIPS64-N64-NEXT: sh $1, 0($16)
; MIPS64-N64-NEXT: ld $16, 8($sp) # 8-byte Folded Reload
; MIPS64-N64-NEXT: ld $gp, 16($sp) # 8-byte Folded Reload
; MIPS64-N64-NEXT: ld $ra, 24($sp) # 8-byte Folded Reload
; MIPS64-N64-NEXT: jr $ra
; MIPS64-N64-NEXT: daddiu $sp, $sp, 32
entry:
%0 = load i16, i16* @g, align 2
%1 = call float @llvm.convert.from.fp16.f32(i16 %0)
%log10 = call float @llvm.log10.f32(float %1)
%2 = call i16 @llvm.convert.to.fp16.f32(float %log10)
store i16 %2, i16* @g, align 2
ret void
}
declare float @llvm.sqrt.f32(float %Val)
define void @fsqrt() {
; MIPS32-LABEL: fsqrt:
; MIPS32: # %bb.0: # %entry
; MIPS32-NEXT: lui $2, %hi(_gp_disp)
; MIPS32-NEXT: addiu $2, $2, %lo(_gp_disp)
; MIPS32-NEXT: addu $1, $2, $25
; MIPS32-NEXT: lw $1, %got(g)($1)
; MIPS32-NEXT: lh $2, 0($1)
; MIPS32-NEXT: fill.h $w0, $2
; MIPS32-NEXT: fexupr.w $w0, $w0
; MIPS32-NEXT: copy_s.w $2, $w0[0]
; MIPS32-NEXT: mtc1 $2, $f0
; MIPS32-NEXT: sqrt.s $f0, $f0
; MIPS32-NEXT: mfc1 $2, $f0
; MIPS32-NEXT: fill.w $w0, $2
; MIPS32-NEXT: fexdo.h $w0, $w0, $w0
; MIPS32-NEXT: copy_u.h $2, $w0[0]
; MIPS32-NEXT: jr $ra
; MIPS32-NEXT: sh $2, 0($1)
;
; MIPS64-N32-LABEL: fsqrt:
; MIPS64-N32: # %bb.0: # %entry
; MIPS64-N32-NEXT: lui $1, %hi(%neg(%gp_rel(fsqrt)))
; MIPS64-N32-NEXT: addu $1, $1, $25
; MIPS64-N32-NEXT: addiu $1, $1, %lo(%neg(%gp_rel(fsqrt)))
; MIPS64-N32-NEXT: lw $1, %got_disp(g)($1)
; MIPS64-N32-NEXT: lh $2, 0($1)
; MIPS64-N32-NEXT: fill.h $w0, $2
; MIPS64-N32-NEXT: fexupr.w $w0, $w0
; MIPS64-N32-NEXT: copy_s.w $2, $w0[0]
; MIPS64-N32-NEXT: mtc1 $2, $f0
; MIPS64-N32-NEXT: sqrt.s $f0, $f0
; MIPS64-N32-NEXT: mfc1 $2, $f0
; MIPS64-N32-NEXT: fill.w $w0, $2
; MIPS64-N32-NEXT: fexdo.h $w0, $w0, $w0
; MIPS64-N32-NEXT: copy_u.h $2, $w0[0]
; MIPS64-N32-NEXT: jr $ra
; MIPS64-N32-NEXT: sh $2, 0($1)
;
; MIPS64-N64-LABEL: fsqrt:
; MIPS64-N64: # %bb.0: # %entry
; MIPS64-N64-NEXT: lui $1, %hi(%neg(%gp_rel(fsqrt)))
; MIPS64-N64-NEXT: daddu $1, $1, $25
; MIPS64-N64-NEXT: daddiu $1, $1, %lo(%neg(%gp_rel(fsqrt)))
; MIPS64-N64-NEXT: ld $1, %got_disp(g)($1)
; MIPS64-N64-NEXT: lh $2, 0($1)
; MIPS64-N64-NEXT: fill.h $w0, $2
; MIPS64-N64-NEXT: fexupr.w $w0, $w0
; MIPS64-N64-NEXT: copy_s.w $2, $w0[0]
; MIPS64-N64-NEXT: mtc1 $2, $f0
; MIPS64-N64-NEXT: sqrt.s $f0, $f0
; MIPS64-N64-NEXT: mfc1 $2, $f0
; MIPS64-N64-NEXT: fill.w $w0, $2
; MIPS64-N64-NEXT: fexdo.h $w0, $w0, $w0
; MIPS64-N64-NEXT: copy_u.h $2, $w0[0]
; MIPS64-N64-NEXT: jr $ra
; MIPS64-N64-NEXT: sh $2, 0($1)
entry:
%0 = load i16, i16* @g, align 2
%1 = call float @llvm.convert.from.fp16.f32(i16 %0)
%sqrt = call float @llvm.sqrt.f32(float %1)
%2 = call i16 @llvm.convert.to.fp16.f32(float %sqrt)
store i16 %2, i16* @g, align 2
ret void
}
declare float @llvm.sin.f32(float %Val)
define void @fsin() {
; MIPS32-LABEL: fsin:
; MIPS32: # %bb.0: # %entry
; MIPS32-NEXT: lui $2, %hi(_gp_disp)
; MIPS32-NEXT: addiu $2, $2, %lo(_gp_disp)
; MIPS32-NEXT: addiu $sp, $sp, -24
; MIPS32-NEXT: .cfi_def_cfa_offset 24
; MIPS32-NEXT: sw $ra, 20($sp) # 4-byte Folded Spill
; MIPS32-NEXT: sw $16, 16($sp) # 4-byte Folded Spill
; MIPS32-NEXT: .cfi_offset 31, -4
; MIPS32-NEXT: .cfi_offset 16, -8
; MIPS32-NEXT: addu $gp, $2, $25
; MIPS32-NEXT: lw $16, %got(g)($gp)
; MIPS32-NEXT: lh $1, 0($16)
; MIPS32-NEXT: fill.h $w0, $1
; MIPS32-NEXT: fexupr.w $w0, $w0
; MIPS32-NEXT: copy_s.w $1, $w0[0]
; MIPS32-NEXT: lw $25, %call16(sinf)($gp)
; MIPS32-NEXT: jalr $25
; MIPS32-NEXT: mtc1 $1, $f12
; MIPS32-NEXT: mfc1 $1, $f0
; MIPS32-NEXT: fill.w $w0, $1
; MIPS32-NEXT: fexdo.h $w0, $w0, $w0
; MIPS32-NEXT: copy_u.h $1, $w0[0]
; MIPS32-NEXT: sh $1, 0($16)
; MIPS32-NEXT: lw $16, 16($sp) # 4-byte Folded Reload
; MIPS32-NEXT: lw $ra, 20($sp) # 4-byte Folded Reload
; MIPS32-NEXT: jr $ra
; MIPS32-NEXT: addiu $sp, $sp, 24
;
; MIPS64-N32-LABEL: fsin:
; MIPS64-N32: # %bb.0: # %entry
; MIPS64-N32-NEXT: addiu $sp, $sp, -32
; MIPS64-N32-NEXT: .cfi_def_cfa_offset 32
; MIPS64-N32-NEXT: sd $ra, 24($sp) # 8-byte Folded Spill
; MIPS64-N32-NEXT: sd $gp, 16($sp) # 8-byte Folded Spill
; MIPS64-N32-NEXT: sd $16, 8($sp) # 8-byte Folded Spill
; MIPS64-N32-NEXT: .cfi_offset 31, -8
; MIPS64-N32-NEXT: .cfi_offset 28, -16
; MIPS64-N32-NEXT: .cfi_offset 16, -24
; MIPS64-N32-NEXT: lui $1, %hi(%neg(%gp_rel(fsin)))
; MIPS64-N32-NEXT: addu $1, $1, $25
; MIPS64-N32-NEXT: addiu $gp, $1, %lo(%neg(%gp_rel(fsin)))
; MIPS64-N32-NEXT: lw $16, %got_disp(g)($gp)
; MIPS64-N32-NEXT: lh $1, 0($16)
; MIPS64-N32-NEXT: fill.h $w0, $1
; MIPS64-N32-NEXT: fexupr.w $w0, $w0
; MIPS64-N32-NEXT: copy_s.w $1, $w0[0]
; MIPS64-N32-NEXT: lw $25, %call16(sinf)($gp)
; MIPS64-N32-NEXT: jalr $25
; MIPS64-N32-NEXT: mtc1 $1, $f12
; MIPS64-N32-NEXT: mfc1 $1, $f0
; MIPS64-N32-NEXT: fill.w $w0, $1
; MIPS64-N32-NEXT: fexdo.h $w0, $w0, $w0
; MIPS64-N32-NEXT: copy_u.h $1, $w0[0]
; MIPS64-N32-NEXT: sh $1, 0($16)
; MIPS64-N32-NEXT: ld $16, 8($sp) # 8-byte Folded Reload
; MIPS64-N32-NEXT: ld $gp, 16($sp) # 8-byte Folded Reload
; MIPS64-N32-NEXT: ld $ra, 24($sp) # 8-byte Folded Reload
; MIPS64-N32-NEXT: jr $ra
; MIPS64-N32-NEXT: addiu $sp, $sp, 32
;
; MIPS64-N64-LABEL: fsin:
; MIPS64-N64: # %bb.0: # %entry
; MIPS64-N64-NEXT: daddiu $sp, $sp, -32
; MIPS64-N64-NEXT: .cfi_def_cfa_offset 32
; MIPS64-N64-NEXT: sd $ra, 24($sp) # 8-byte Folded Spill
; MIPS64-N64-NEXT: sd $gp, 16($sp) # 8-byte Folded Spill
; MIPS64-N64-NEXT: sd $16, 8($sp) # 8-byte Folded Spill
; MIPS64-N64-NEXT: .cfi_offset 31, -8
; MIPS64-N64-NEXT: .cfi_offset 28, -16
; MIPS64-N64-NEXT: .cfi_offset 16, -24
; MIPS64-N64-NEXT: lui $1, %hi(%neg(%gp_rel(fsin)))
; MIPS64-N64-NEXT: daddu $1, $1, $25
; MIPS64-N64-NEXT: daddiu $gp, $1, %lo(%neg(%gp_rel(fsin)))
; MIPS64-N64-NEXT: ld $16, %got_disp(g)($gp)
; MIPS64-N64-NEXT: lh $1, 0($16)
; MIPS64-N64-NEXT: fill.h $w0, $1
; MIPS64-N64-NEXT: fexupr.w $w0, $w0
; MIPS64-N64-NEXT: copy_s.w $1, $w0[0]
; MIPS64-N64-NEXT: ld $25, %call16(sinf)($gp)
; MIPS64-N64-NEXT: jalr $25
; MIPS64-N64-NEXT: mtc1 $1, $f12
; MIPS64-N64-NEXT: mfc1 $1, $f0
; MIPS64-N64-NEXT: fill.w $w0, $1
; MIPS64-N64-NEXT: fexdo.h $w0, $w0, $w0
; MIPS64-N64-NEXT: copy_u.h $1, $w0[0]
; MIPS64-N64-NEXT: sh $1, 0($16)
; MIPS64-N64-NEXT: ld $16, 8($sp) # 8-byte Folded Reload
; MIPS64-N64-NEXT: ld $gp, 16($sp) # 8-byte Folded Reload
; MIPS64-N64-NEXT: ld $ra, 24($sp) # 8-byte Folded Reload
; MIPS64-N64-NEXT: jr $ra
; MIPS64-N64-NEXT: daddiu $sp, $sp, 32
entry:
%0 = load i16, i16* @g, align 2
%1 = call float @llvm.convert.from.fp16.f32(i16 %0)
%sin = call float @llvm.sin.f32(float %1)
%2 = call i16 @llvm.convert.to.fp16.f32(float %sin)
store i16 %2, i16* @g, align 2
ret void
}
declare float @llvm.cos.f32(float %Val)
define void @fcos() {
; MIPS32-LABEL: fcos:
; MIPS32: # %bb.0: # %entry
; MIPS32-NEXT: lui $2, %hi(_gp_disp)
; MIPS32-NEXT: addiu $2, $2, %lo(_gp_disp)
; MIPS32-NEXT: addiu $sp, $sp, -24
; MIPS32-NEXT: .cfi_def_cfa_offset 24
; MIPS32-NEXT: sw $ra, 20($sp) # 4-byte Folded Spill
; MIPS32-NEXT: sw $16, 16($sp) # 4-byte Folded Spill
; MIPS32-NEXT: .cfi_offset 31, -4
; MIPS32-NEXT: .cfi_offset 16, -8
; MIPS32-NEXT: addu $gp, $2, $25
; MIPS32-NEXT: lw $16, %got(g)($gp)
; MIPS32-NEXT: lh $1, 0($16)
; MIPS32-NEXT: fill.h $w0, $1
; MIPS32-NEXT: fexupr.w $w0, $w0
; MIPS32-NEXT: copy_s.w $1, $w0[0]
; MIPS32-NEXT: lw $25, %call16(cosf)($gp)
; MIPS32-NEXT: jalr $25
; MIPS32-NEXT: mtc1 $1, $f12
; MIPS32-NEXT: mfc1 $1, $f0
; MIPS32-NEXT: fill.w $w0, $1
; MIPS32-NEXT: fexdo.h $w0, $w0, $w0
; MIPS32-NEXT: copy_u.h $1, $w0[0]
; MIPS32-NEXT: sh $1, 0($16)
; MIPS32-NEXT: lw $16, 16($sp) # 4-byte Folded Reload
; MIPS32-NEXT: lw $ra, 20($sp) # 4-byte Folded Reload
; MIPS32-NEXT: jr $ra
; MIPS32-NEXT: addiu $sp, $sp, 24
;
; MIPS64-N32-LABEL: fcos:
; MIPS64-N32: # %bb.0: # %entry
; MIPS64-N32-NEXT: addiu $sp, $sp, -32
; MIPS64-N32-NEXT: .cfi_def_cfa_offset 32
; MIPS64-N32-NEXT: sd $ra, 24($sp) # 8-byte Folded Spill
; MIPS64-N32-NEXT: sd $gp, 16($sp) # 8-byte Folded Spill
; MIPS64-N32-NEXT: sd $16, 8($sp) # 8-byte Folded Spill
; MIPS64-N32-NEXT: .cfi_offset 31, -8
; MIPS64-N32-NEXT: .cfi_offset 28, -16
; MIPS64-N32-NEXT: .cfi_offset 16, -24
; MIPS64-N32-NEXT: lui $1, %hi(%neg(%gp_rel(fcos)))
; MIPS64-N32-NEXT: addu $1, $1, $25
; MIPS64-N32-NEXT: addiu $gp, $1, %lo(%neg(%gp_rel(fcos)))
; MIPS64-N32-NEXT: lw $16, %got_disp(g)($gp)
; MIPS64-N32-NEXT: lh $1, 0($16)
; MIPS64-N32-NEXT: fill.h $w0, $1
; MIPS64-N32-NEXT: fexupr.w $w0, $w0
; MIPS64-N32-NEXT: copy_s.w $1, $w0[0]
; MIPS64-N32-NEXT: lw $25, %call16(cosf)($gp)
; MIPS64-N32-NEXT: jalr $25
; MIPS64-N32-NEXT: mtc1 $1, $f12
; MIPS64-N32-NEXT: mfc1 $1, $f0
; MIPS64-N32-NEXT: fill.w $w0, $1
; MIPS64-N32-NEXT: fexdo.h $w0, $w0, $w0
; MIPS64-N32-NEXT: copy_u.h $1, $w0[0]
; MIPS64-N32-NEXT: sh $1, 0($16)
; MIPS64-N32-NEXT: ld $16, 8($sp) # 8-byte Folded Reload
; MIPS64-N32-NEXT: ld $gp, 16($sp) # 8-byte Folded Reload
; MIPS64-N32-NEXT: ld $ra, 24($sp) # 8-byte Folded Reload
; MIPS64-N32-NEXT: jr $ra
; MIPS64-N32-NEXT: addiu $sp, $sp, 32
;
; MIPS64-N64-LABEL: fcos:
; MIPS64-N64: # %bb.0: # %entry
; MIPS64-N64-NEXT: daddiu $sp, $sp, -32
; MIPS64-N64-NEXT: .cfi_def_cfa_offset 32
; MIPS64-N64-NEXT: sd $ra, 24($sp) # 8-byte Folded Spill
; MIPS64-N64-NEXT: sd $gp, 16($sp) # 8-byte Folded Spill
; MIPS64-N64-NEXT: sd $16, 8($sp) # 8-byte Folded Spill
; MIPS64-N64-NEXT: .cfi_offset 31, -8
; MIPS64-N64-NEXT: .cfi_offset 28, -16
; MIPS64-N64-NEXT: .cfi_offset 16, -24
; MIPS64-N64-NEXT: lui $1, %hi(%neg(%gp_rel(fcos)))
; MIPS64-N64-NEXT: daddu $1, $1, $25
; MIPS64-N64-NEXT: daddiu $gp, $1, %lo(%neg(%gp_rel(fcos)))
; MIPS64-N64-NEXT: ld $16, %got_disp(g)($gp)
; MIPS64-N64-NEXT: lh $1, 0($16)
; MIPS64-N64-NEXT: fill.h $w0, $1
; MIPS64-N64-NEXT: fexupr.w $w0, $w0
; MIPS64-N64-NEXT: copy_s.w $1, $w0[0]
; MIPS64-N64-NEXT: ld $25, %call16(cosf)($gp)
; MIPS64-N64-NEXT: jalr $25
; MIPS64-N64-NEXT: mtc1 $1, $f12
; MIPS64-N64-NEXT: mfc1 $1, $f0
; MIPS64-N64-NEXT: fill.w $w0, $1
; MIPS64-N64-NEXT: fexdo.h $w0, $w0, $w0
; MIPS64-N64-NEXT: copy_u.h $1, $w0[0]
; MIPS64-N64-NEXT: sh $1, 0($16)
; MIPS64-N64-NEXT: ld $16, 8($sp) # 8-byte Folded Reload
; MIPS64-N64-NEXT: ld $gp, 16($sp) # 8-byte Folded Reload
; MIPS64-N64-NEXT: ld $ra, 24($sp) # 8-byte Folded Reload
; MIPS64-N64-NEXT: jr $ra
; MIPS64-N64-NEXT: daddiu $sp, $sp, 32
entry:
%0 = load i16, i16* @g, align 2
%1 = call float @llvm.convert.from.fp16.f32(i16 %0)
%cos = call float @llvm.cos.f32(float %1)
%2 = call i16 @llvm.convert.to.fp16.f32(float %cos)
store i16 %2, i16* @g, align 2
ret void
}
declare float @llvm.exp.f32(float %Val)
define void @fexp() {
; MIPS32-LABEL: fexp:
; MIPS32: # %bb.0: # %entry
; MIPS32-NEXT: lui $2, %hi(_gp_disp)
; MIPS32-NEXT: addiu $2, $2, %lo(_gp_disp)
; MIPS32-NEXT: addiu $sp, $sp, -24
; MIPS32-NEXT: .cfi_def_cfa_offset 24
; MIPS32-NEXT: sw $ra, 20($sp) # 4-byte Folded Spill
; MIPS32-NEXT: sw $16, 16($sp) # 4-byte Folded Spill
; MIPS32-NEXT: .cfi_offset 31, -4
; MIPS32-NEXT: .cfi_offset 16, -8
; MIPS32-NEXT: addu $gp, $2, $25
; MIPS32-NEXT: lw $16, %got(g)($gp)
; MIPS32-NEXT: lh $1, 0($16)
; MIPS32-NEXT: fill.h $w0, $1
; MIPS32-NEXT: fexupr.w $w0, $w0
; MIPS32-NEXT: copy_s.w $1, $w0[0]
; MIPS32-NEXT: lw $25, %call16(expf)($gp)
; MIPS32-NEXT: jalr $25
; MIPS32-NEXT: mtc1 $1, $f12
; MIPS32-NEXT: mfc1 $1, $f0
; MIPS32-NEXT: fill.w $w0, $1
; MIPS32-NEXT: fexdo.h $w0, $w0, $w0
; MIPS32-NEXT: copy_u.h $1, $w0[0]
; MIPS32-NEXT: sh $1, 0($16)
; MIPS32-NEXT: lw $16, 16($sp) # 4-byte Folded Reload
; MIPS32-NEXT: lw $ra, 20($sp) # 4-byte Folded Reload
; MIPS32-NEXT: jr $ra
; MIPS32-NEXT: addiu $sp, $sp, 24
;
; MIPS64-N32-LABEL: fexp:
; MIPS64-N32: # %bb.0: # %entry
; MIPS64-N32-NEXT: addiu $sp, $sp, -32
; MIPS64-N32-NEXT: .cfi_def_cfa_offset 32
; MIPS64-N32-NEXT: sd $ra, 24($sp) # 8-byte Folded Spill
; MIPS64-N32-NEXT: sd $gp, 16($sp) # 8-byte Folded Spill
; MIPS64-N32-NEXT: sd $16, 8($sp) # 8-byte Folded Spill
; MIPS64-N32-NEXT: .cfi_offset 31, -8
; MIPS64-N32-NEXT: .cfi_offset 28, -16
; MIPS64-N32-NEXT: .cfi_offset 16, -24
; MIPS64-N32-NEXT: lui $1, %hi(%neg(%gp_rel(fexp)))
; MIPS64-N32-NEXT: addu $1, $1, $25
; MIPS64-N32-NEXT: addiu $gp, $1, %lo(%neg(%gp_rel(fexp)))
; MIPS64-N32-NEXT: lw $16, %got_disp(g)($gp)
; MIPS64-N32-NEXT: lh $1, 0($16)
; MIPS64-N32-NEXT: fill.h $w0, $1
; MIPS64-N32-NEXT: fexupr.w $w0, $w0
; MIPS64-N32-NEXT: copy_s.w $1, $w0[0]
; MIPS64-N32-NEXT: lw $25, %call16(expf)($gp)
; MIPS64-N32-NEXT: jalr $25
; MIPS64-N32-NEXT: mtc1 $1, $f12
; MIPS64-N32-NEXT: mfc1 $1, $f0
; MIPS64-N32-NEXT: fill.w $w0, $1
; MIPS64-N32-NEXT: fexdo.h $w0, $w0, $w0
; MIPS64-N32-NEXT: copy_u.h $1, $w0[0]
; MIPS64-N32-NEXT: sh $1, 0($16)
; MIPS64-N32-NEXT: ld $16, 8($sp) # 8-byte Folded Reload
; MIPS64-N32-NEXT: ld $gp, 16($sp) # 8-byte Folded Reload
; MIPS64-N32-NEXT: ld $ra, 24($sp) # 8-byte Folded Reload
; MIPS64-N32-NEXT: jr $ra
; MIPS64-N32-NEXT: addiu $sp, $sp, 32
;
; MIPS64-N64-LABEL: fexp:
; MIPS64-N64: # %bb.0: # %entry
; MIPS64-N64-NEXT: daddiu $sp, $sp, -32
; MIPS64-N64-NEXT: .cfi_def_cfa_offset 32
; MIPS64-N64-NEXT: sd $ra, 24($sp) # 8-byte Folded Spill
; MIPS64-N64-NEXT: sd $gp, 16($sp) # 8-byte Folded Spill
; MIPS64-N64-NEXT: sd $16, 8($sp) # 8-byte Folded Spill
; MIPS64-N64-NEXT: .cfi_offset 31, -8
; MIPS64-N64-NEXT: .cfi_offset 28, -16
; MIPS64-N64-NEXT: .cfi_offset 16, -24
; MIPS64-N64-NEXT: lui $1, %hi(%neg(%gp_rel(fexp)))
; MIPS64-N64-NEXT: daddu $1, $1, $25
; MIPS64-N64-NEXT: daddiu $gp, $1, %lo(%neg(%gp_rel(fexp)))
; MIPS64-N64-NEXT: ld $16, %got_disp(g)($gp)
; MIPS64-N64-NEXT: lh $1, 0($16)
; MIPS64-N64-NEXT: fill.h $w0, $1
; MIPS64-N64-NEXT: fexupr.w $w0, $w0
; MIPS64-N64-NEXT: copy_s.w $1, $w0[0]
; MIPS64-N64-NEXT: ld $25, %call16(expf)($gp)
; MIPS64-N64-NEXT: jalr $25
; MIPS64-N64-NEXT: mtc1 $1, $f12
; MIPS64-N64-NEXT: mfc1 $1, $f0
; MIPS64-N64-NEXT: fill.w $w0, $1
; MIPS64-N64-NEXT: fexdo.h $w0, $w0, $w0
; MIPS64-N64-NEXT: copy_u.h $1, $w0[0]
; MIPS64-N64-NEXT: sh $1, 0($16)
; MIPS64-N64-NEXT: ld $16, 8($sp) # 8-byte Folded Reload
; MIPS64-N64-NEXT: ld $gp, 16($sp) # 8-byte Folded Reload
; MIPS64-N64-NEXT: ld $ra, 24($sp) # 8-byte Folded Reload
; MIPS64-N64-NEXT: jr $ra
; MIPS64-N64-NEXT: daddiu $sp, $sp, 32
entry:
%0 = load i16, i16* @g, align 2
%1 = call float @llvm.convert.from.fp16.f32(i16 %0)
%exp = call float @llvm.exp.f32(float %1)
%2 = call i16 @llvm.convert.to.fp16.f32(float %exp)
store i16 %2, i16* @g, align 2
ret void
}
declare float @llvm.exp2.f32(float %Val)
define void @fexp2() {
; MIPS32-LABEL: fexp2:
; MIPS32: # %bb.0: # %entry
; MIPS32-NEXT: lui $2, %hi(_gp_disp)
; MIPS32-NEXT: addiu $2, $2, %lo(_gp_disp)
; MIPS32-NEXT: addiu $sp, $sp, -24
; MIPS32-NEXT: .cfi_def_cfa_offset 24
; MIPS32-NEXT: sw $ra, 20($sp) # 4-byte Folded Spill
; MIPS32-NEXT: sw $16, 16($sp) # 4-byte Folded Spill
; MIPS32-NEXT: .cfi_offset 31, -4
; MIPS32-NEXT: .cfi_offset 16, -8
; MIPS32-NEXT: addu $gp, $2, $25
; MIPS32-NEXT: lw $16, %got(g)($gp)
; MIPS32-NEXT: lh $1, 0($16)
; MIPS32-NEXT: fill.h $w0, $1
; MIPS32-NEXT: fexupr.w $w0, $w0
; MIPS32-NEXT: copy_s.w $1, $w0[0]
; MIPS32-NEXT: lw $25, %call16(exp2f)($gp)
; MIPS32-NEXT: jalr $25
; MIPS32-NEXT: mtc1 $1, $f12
; MIPS32-NEXT: mfc1 $1, $f0
; MIPS32-NEXT: fill.w $w0, $1
; MIPS32-NEXT: fexdo.h $w0, $w0, $w0
; MIPS32-NEXT: copy_u.h $1, $w0[0]
; MIPS32-NEXT: sh $1, 0($16)
; MIPS32-NEXT: lw $16, 16($sp) # 4-byte Folded Reload
; MIPS32-NEXT: lw $ra, 20($sp) # 4-byte Folded Reload
; MIPS32-NEXT: jr $ra
; MIPS32-NEXT: addiu $sp, $sp, 24
;
; MIPS64-N32-LABEL: fexp2:
; MIPS64-N32: # %bb.0: # %entry
; MIPS64-N32-NEXT: addiu $sp, $sp, -32
; MIPS64-N32-NEXT: .cfi_def_cfa_offset 32
; MIPS64-N32-NEXT: sd $ra, 24($sp) # 8-byte Folded Spill
; MIPS64-N32-NEXT: sd $gp, 16($sp) # 8-byte Folded Spill
; MIPS64-N32-NEXT: sd $16, 8($sp) # 8-byte Folded Spill
; MIPS64-N32-NEXT: .cfi_offset 31, -8
; MIPS64-N32-NEXT: .cfi_offset 28, -16
; MIPS64-N32-NEXT: .cfi_offset 16, -24
; MIPS64-N32-NEXT: lui $1, %hi(%neg(%gp_rel(fexp2)))
; MIPS64-N32-NEXT: addu $1, $1, $25
; MIPS64-N32-NEXT: addiu $gp, $1, %lo(%neg(%gp_rel(fexp2)))
; MIPS64-N32-NEXT: lw $16, %got_disp(g)($gp)
; MIPS64-N32-NEXT: lh $1, 0($16)
; MIPS64-N32-NEXT: fill.h $w0, $1
; MIPS64-N32-NEXT: fexupr.w $w0, $w0
; MIPS64-N32-NEXT: copy_s.w $1, $w0[0]
; MIPS64-N32-NEXT: lw $25, %call16(exp2f)($gp)
; MIPS64-N32-NEXT: jalr $25
; MIPS64-N32-NEXT: mtc1 $1, $f12
; MIPS64-N32-NEXT: mfc1 $1, $f0
; MIPS64-N32-NEXT: fill.w $w0, $1
; MIPS64-N32-NEXT: fexdo.h $w0, $w0, $w0
; MIPS64-N32-NEXT: copy_u.h $1, $w0[0]
; MIPS64-N32-NEXT: sh $1, 0($16)
; MIPS64-N32-NEXT: ld $16, 8($sp) # 8-byte Folded Reload
; MIPS64-N32-NEXT: ld $gp, 16($sp) # 8-byte Folded Reload
; MIPS64-N32-NEXT: ld $ra, 24($sp) # 8-byte Folded Reload
; MIPS64-N32-NEXT: jr $ra
; MIPS64-N32-NEXT: addiu $sp, $sp, 32
;
; MIPS64-N64-LABEL: fexp2:
; MIPS64-N64: # %bb.0: # %entry
; MIPS64-N64-NEXT: daddiu $sp, $sp, -32
; MIPS64-N64-NEXT: .cfi_def_cfa_offset 32
; MIPS64-N64-NEXT: sd $ra, 24($sp) # 8-byte Folded Spill
; MIPS64-N64-NEXT: sd $gp, 16($sp) # 8-byte Folded Spill
; MIPS64-N64-NEXT: sd $16, 8($sp) # 8-byte Folded Spill
; MIPS64-N64-NEXT: .cfi_offset 31, -8
; MIPS64-N64-NEXT: .cfi_offset 28, -16
; MIPS64-N64-NEXT: .cfi_offset 16, -24
; MIPS64-N64-NEXT: lui $1, %hi(%neg(%gp_rel(fexp2)))
; MIPS64-N64-NEXT: daddu $1, $1, $25
; MIPS64-N64-NEXT: daddiu $gp, $1, %lo(%neg(%gp_rel(fexp2)))
; MIPS64-N64-NEXT: ld $16, %got_disp(g)($gp)
; MIPS64-N64-NEXT: lh $1, 0($16)
; MIPS64-N64-NEXT: fill.h $w0, $1
; MIPS64-N64-NEXT: fexupr.w $w0, $w0
; MIPS64-N64-NEXT: copy_s.w $1, $w0[0]
; MIPS64-N64-NEXT: ld $25, %call16(exp2f)($gp)
; MIPS64-N64-NEXT: jalr $25
; MIPS64-N64-NEXT: mtc1 $1, $f12
; MIPS64-N64-NEXT: mfc1 $1, $f0
; MIPS64-N64-NEXT: fill.w $w0, $1
; MIPS64-N64-NEXT: fexdo.h $w0, $w0, $w0
; MIPS64-N64-NEXT: copy_u.h $1, $w0[0]
; MIPS64-N64-NEXT: sh $1, 0($16)
; MIPS64-N64-NEXT: ld $16, 8($sp) # 8-byte Folded Reload
; MIPS64-N64-NEXT: ld $gp, 16($sp) # 8-byte Folded Reload
; MIPS64-N64-NEXT: ld $ra, 24($sp) # 8-byte Folded Reload
; MIPS64-N64-NEXT: jr $ra
; MIPS64-N64-NEXT: daddiu $sp, $sp, 32
entry:
%0 = load i16, i16* @g, align 2
%1 = call float @llvm.convert.from.fp16.f32(i16 %0)
%exp2 = call float @llvm.exp2.f32(float %1)
%2 = call i16 @llvm.convert.to.fp16.f32(float %exp2)
store i16 %2, i16* @g, align 2
ret void
}
declare float @llvm.fma.f32(float, float, float)
define void @ffma(float %b, float %c) {
; MIPS32-LABEL: ffma:
; MIPS32: # %bb.0: # %entry
; MIPS32-NEXT: lui $2, %hi(_gp_disp)
; MIPS32-NEXT: addiu $2, $2, %lo(_gp_disp)
; MIPS32-NEXT: addiu $sp, $sp, -24
; MIPS32-NEXT: .cfi_def_cfa_offset 24
; MIPS32-NEXT: sw $ra, 20($sp) # 4-byte Folded Spill
; MIPS32-NEXT: sw $16, 16($sp) # 4-byte Folded Spill
; MIPS32-NEXT: .cfi_offset 31, -4
; MIPS32-NEXT: .cfi_offset 16, -8
; MIPS32-NEXT: addu $gp, $2, $25
; MIPS32-NEXT: mov.s $f0, $f12
; MIPS32-NEXT: mfc1 $6, $f14
; MIPS32-NEXT: lw $16, %got(g)($gp)
; MIPS32-NEXT: lh $1, 0($16)
; MIPS32-NEXT: fill.h $w1, $1
; MIPS32-NEXT: fexupr.w $w1, $w1
; MIPS32-NEXT: copy_s.w $1, $w1[0]
; MIPS32-NEXT: mtc1 $1, $f12
; MIPS32-NEXT: lw $25, %call16(fmaf)($gp)
; MIPS32-NEXT: jalr $25
; MIPS32-NEXT: mov.s $f14, $f0
; MIPS32-NEXT: mfc1 $1, $f0
; MIPS32-NEXT: fill.w $w0, $1
; MIPS32-NEXT: fexdo.h $w0, $w0, $w0
; MIPS32-NEXT: copy_u.h $1, $w0[0]
; MIPS32-NEXT: sh $1, 0($16)
; MIPS32-NEXT: lw $16, 16($sp) # 4-byte Folded Reload
; MIPS32-NEXT: lw $ra, 20($sp) # 4-byte Folded Reload
; MIPS32-NEXT: jr $ra
; MIPS32-NEXT: addiu $sp, $sp, 24
;
; MIPS64-N32-LABEL: ffma:
; MIPS64-N32: # %bb.0: # %entry
; MIPS64-N32-NEXT: addiu $sp, $sp, -32
; MIPS64-N32-NEXT: .cfi_def_cfa_offset 32
; MIPS64-N32-NEXT: sd $ra, 24($sp) # 8-byte Folded Spill
; MIPS64-N32-NEXT: sd $gp, 16($sp) # 8-byte Folded Spill
; MIPS64-N32-NEXT: sd $16, 8($sp) # 8-byte Folded Spill
; MIPS64-N32-NEXT: .cfi_offset 31, -8
; MIPS64-N32-NEXT: .cfi_offset 28, -16
; MIPS64-N32-NEXT: .cfi_offset 16, -24
; MIPS64-N32-NEXT: lui $1, %hi(%neg(%gp_rel(ffma)))
; MIPS64-N32-NEXT: addu $1, $1, $25
; MIPS64-N32-NEXT: addiu $gp, $1, %lo(%neg(%gp_rel(ffma)))
; MIPS64-N32-NEXT: mov.s $f14, $f13
; MIPS64-N32-NEXT: mov.s $f13, $f12
; MIPS64-N32-NEXT: lw $16, %got_disp(g)($gp)
; MIPS64-N32-NEXT: lh $1, 0($16)
; MIPS64-N32-NEXT: fill.h $w0, $1
; MIPS64-N32-NEXT: fexupr.w $w0, $w0
; MIPS64-N32-NEXT: copy_s.w $1, $w0[0]
; MIPS64-N32-NEXT: lw $25, %call16(fmaf)($gp)
; MIPS64-N32-NEXT: jalr $25
; MIPS64-N32-NEXT: mtc1 $1, $f12
; MIPS64-N32-NEXT: mfc1 $1, $f0
; MIPS64-N32-NEXT: fill.w $w0, $1
; MIPS64-N32-NEXT: fexdo.h $w0, $w0, $w0
; MIPS64-N32-NEXT: copy_u.h $1, $w0[0]
; MIPS64-N32-NEXT: sh $1, 0($16)
; MIPS64-N32-NEXT: ld $16, 8($sp) # 8-byte Folded Reload
; MIPS64-N32-NEXT: ld $gp, 16($sp) # 8-byte Folded Reload
; MIPS64-N32-NEXT: ld $ra, 24($sp) # 8-byte Folded Reload
; MIPS64-N32-NEXT: jr $ra
; MIPS64-N32-NEXT: addiu $sp, $sp, 32
;
; MIPS64-N64-LABEL: ffma:
; MIPS64-N64: # %bb.0: # %entry
; MIPS64-N64-NEXT: daddiu $sp, $sp, -32
; MIPS64-N64-NEXT: .cfi_def_cfa_offset 32
; MIPS64-N64-NEXT: sd $ra, 24($sp) # 8-byte Folded Spill
; MIPS64-N64-NEXT: sd $gp, 16($sp) # 8-byte Folded Spill
; MIPS64-N64-NEXT: sd $16, 8($sp) # 8-byte Folded Spill
; MIPS64-N64-NEXT: .cfi_offset 31, -8
; MIPS64-N64-NEXT: .cfi_offset 28, -16
; MIPS64-N64-NEXT: .cfi_offset 16, -24
; MIPS64-N64-NEXT: lui $1, %hi(%neg(%gp_rel(ffma)))
; MIPS64-N64-NEXT: daddu $1, $1, $25
; MIPS64-N64-NEXT: daddiu $gp, $1, %lo(%neg(%gp_rel(ffma)))
; MIPS64-N64-NEXT: mov.s $f14, $f13
; MIPS64-N64-NEXT: mov.s $f13, $f12
; MIPS64-N64-NEXT: ld $16, %got_disp(g)($gp)
; MIPS64-N64-NEXT: lh $1, 0($16)
; MIPS64-N64-NEXT: fill.h $w0, $1
; MIPS64-N64-NEXT: fexupr.w $w0, $w0
; MIPS64-N64-NEXT: copy_s.w $1, $w0[0]
; MIPS64-N64-NEXT: ld $25, %call16(fmaf)($gp)
; MIPS64-N64-NEXT: jalr $25
; MIPS64-N64-NEXT: mtc1 $1, $f12
; MIPS64-N64-NEXT: mfc1 $1, $f0
; MIPS64-N64-NEXT: fill.w $w0, $1
; MIPS64-N64-NEXT: fexdo.h $w0, $w0, $w0
; MIPS64-N64-NEXT: copy_u.h $1, $w0[0]
; MIPS64-N64-NEXT: sh $1, 0($16)
; MIPS64-N64-NEXT: ld $16, 8($sp) # 8-byte Folded Reload
; MIPS64-N64-NEXT: ld $gp, 16($sp) # 8-byte Folded Reload
; MIPS64-N64-NEXT: ld $ra, 24($sp) # 8-byte Folded Reload
; MIPS64-N64-NEXT: jr $ra
; MIPS64-N64-NEXT: daddiu $sp, $sp, 32
entry:
%0 = load i16, i16* @g, align 2
%1 = call float @llvm.convert.from.fp16.f32(i16 %0)
%fma = call float @llvm.fma.f32(float %1, float %b, float %c)
%2 = call i16 @llvm.convert.to.fp16.f32(float %fma)
store i16 %2, i16* @g, align 2
ret void
}
; FIXME: For MIPSR6, this should produced the maddf.s instruction. MIPSR5 cannot
; fuse the operation such that the intermediate result is not rounded.
declare float @llvm.fmuladd.f32(float, float, float)
define void @ffmuladd(float %b, float %c) {
; MIPS32-O32-LABEL: ffmuladd:
; MIPS32-O32: # %bb.0: # %entry
; MIPS32-O32-NEXT: lui $2, %hi(_gp_disp)
; MIPS32-O32-NEXT: addiu $2, $2, %lo(_gp_disp)
; MIPS32-O32-NEXT: addu $1, $2, $25
; MIPS32-O32-NEXT: lw $1, %got(g)($1)
; MIPS32-O32-NEXT: lh $2, 0($1)
; MIPS32-O32-NEXT: fill.h $w0, $2
; MIPS32-O32-NEXT: fexupr.w $w0, $w0
; MIPS32-O32-NEXT: copy_s.w $2, $w0[0]
; MIPS32-O32-NEXT: mtc1 $2, $f0
; MIPS32-O32-NEXT: madd.s $f0, $f14, $f0, $f12
; MIPS32-O32-NEXT: mfc1 $2, $f0
; MIPS32-O32-NEXT: fill.w $w0, $2
; MIPS32-O32-NEXT: fexdo.h $w0, $w0, $w0
; MIPS32-O32-NEXT: copy_u.h $2, $w0[0]
; MIPS32-O32-NEXT: jr $ra
; MIPS32-O32-NEXT: sh $2, 0($1)
;
; MIPS64R5-N32-LABEL: ffmuladd:
; MIPS64R5-N32: # %bb.0: # %entry
; MIPS64R5-N32-NEXT: lui $1, %hi(%neg(%gp_rel(ffmuladd)))
; MIPS64R5-N32-NEXT: addu $1, $1, $25
; MIPS64R5-N32-NEXT: addiu $1, $1, %lo(%neg(%gp_rel(ffmuladd)))
; MIPS64R5-N32-NEXT: lw $1, %got_disp(g)($1)
; MIPS64R5-N32-NEXT: lh $2, 0($1)
; MIPS64R5-N32-NEXT: fill.h $w0, $2
; MIPS64R5-N32-NEXT: fexupr.w $w0, $w0
; MIPS64R5-N32-NEXT: copy_s.w $2, $w0[0]
; MIPS64R5-N32-NEXT: mtc1 $2, $f0
; MIPS64R5-N32-NEXT: madd.s $f0, $f13, $f0, $f12
; MIPS64R5-N32-NEXT: mfc1 $2, $f0
; MIPS64R5-N32-NEXT: fill.w $w0, $2
; MIPS64R5-N32-NEXT: fexdo.h $w0, $w0, $w0
; MIPS64R5-N32-NEXT: copy_u.h $2, $w0[0]
; MIPS64R5-N32-NEXT: jr $ra
; MIPS64R5-N32-NEXT: sh $2, 0($1)
;
; MIPS64R5-N64-LABEL: ffmuladd:
; MIPS64R5-N64: # %bb.0: # %entry
; MIPS64R5-N64-NEXT: lui $1, %hi(%neg(%gp_rel(ffmuladd)))
; MIPS64R5-N64-NEXT: daddu $1, $1, $25
; MIPS64R5-N64-NEXT: daddiu $1, $1, %lo(%neg(%gp_rel(ffmuladd)))
; MIPS64R5-N64-NEXT: ld $1, %got_disp(g)($1)
; MIPS64R5-N64-NEXT: lh $2, 0($1)
; MIPS64R5-N64-NEXT: fill.h $w0, $2
; MIPS64R5-N64-NEXT: fexupr.w $w0, $w0
; MIPS64R5-N64-NEXT: copy_s.w $2, $w0[0]
; MIPS64R5-N64-NEXT: mtc1 $2, $f0
; MIPS64R5-N64-NEXT: madd.s $f0, $f13, $f0, $f12
; MIPS64R5-N64-NEXT: mfc1 $2, $f0
; MIPS64R5-N64-NEXT: fill.w $w0, $2
; MIPS64R5-N64-NEXT: fexdo.h $w0, $w0, $w0
; MIPS64R5-N64-NEXT: copy_u.h $2, $w0[0]
; MIPS64R5-N64-NEXT: jr $ra
; MIPS64R5-N64-NEXT: sh $2, 0($1)
;
; MIPSR6-O32-LABEL: ffmuladd:
; MIPSR6-O32: # %bb.0: # %entry
; MIPSR6-O32-NEXT: lui $2, %hi(_gp_disp)
; MIPSR6-O32-NEXT: addiu $2, $2, %lo(_gp_disp)
; MIPSR6-O32-NEXT: addu $1, $2, $25
; MIPSR6-O32-NEXT: lw $1, %got(g)($1)
; MIPSR6-O32-NEXT: lh $2, 0($1)
; MIPSR6-O32-NEXT: fill.h $w0, $2
; MIPSR6-O32-NEXT: fexupr.w $w0, $w0
; MIPSR6-O32-NEXT: copy_s.w $2, $w0[0]
; MIPSR6-O32-NEXT: mtc1 $2, $f0
; MIPSR6-O32-NEXT: mul.s $f0, $f0, $f12
; MIPSR6-O32-NEXT: add.s $f0, $f0, $f14
; MIPSR6-O32-NEXT: mfc1 $2, $f0
; MIPSR6-O32-NEXT: fill.w $w0, $2
; MIPSR6-O32-NEXT: fexdo.h $w0, $w0, $w0
; MIPSR6-O32-NEXT: copy_u.h $2, $w0[0]
; MIPSR6-O32-NEXT: jr $ra
; MIPSR6-O32-NEXT: sh $2, 0($1)
;
; MIPSR6-N32-LABEL: ffmuladd:
; MIPSR6-N32: # %bb.0: # %entry
; MIPSR6-N32-NEXT: lui $1, %hi(%neg(%gp_rel(ffmuladd)))
; MIPSR6-N32-NEXT: addu $1, $1, $25
; MIPSR6-N32-NEXT: addiu $1, $1, %lo(%neg(%gp_rel(ffmuladd)))
; MIPSR6-N32-NEXT: lw $1, %got_disp(g)($1)
; MIPSR6-N32-NEXT: lh $2, 0($1)
; MIPSR6-N32-NEXT: fill.h $w0, $2
; MIPSR6-N32-NEXT: fexupr.w $w0, $w0
; MIPSR6-N32-NEXT: copy_s.w $2, $w0[0]
; MIPSR6-N32-NEXT: mtc1 $2, $f0
; MIPSR6-N32-NEXT: mul.s $f0, $f0, $f12
; MIPSR6-N32-NEXT: add.s $f0, $f0, $f13
; MIPSR6-N32-NEXT: mfc1 $2, $f0
; MIPSR6-N32-NEXT: fill.w $w0, $2
; MIPSR6-N32-NEXT: fexdo.h $w0, $w0, $w0
; MIPSR6-N32-NEXT: copy_u.h $2, $w0[0]
; MIPSR6-N32-NEXT: jr $ra
; MIPSR6-N32-NEXT: sh $2, 0($1)
;
; MIPSR6-N64-LABEL: ffmuladd:
; MIPSR6-N64: # %bb.0: # %entry
; MIPSR6-N64-NEXT: lui $1, %hi(%neg(%gp_rel(ffmuladd)))
; MIPSR6-N64-NEXT: daddu $1, $1, $25
; MIPSR6-N64-NEXT: daddiu $1, $1, %lo(%neg(%gp_rel(ffmuladd)))
; MIPSR6-N64-NEXT: ld $1, %got_disp(g)($1)
; MIPSR6-N64-NEXT: lh $2, 0($1)
; MIPSR6-N64-NEXT: fill.h $w0, $2
; MIPSR6-N64-NEXT: fexupr.w $w0, $w0
; MIPSR6-N64-NEXT: copy_s.w $2, $w0[0]
; MIPSR6-N64-NEXT: mtc1 $2, $f0
; MIPSR6-N64-NEXT: mul.s $f0, $f0, $f12
; MIPSR6-N64-NEXT: add.s $f0, $f0, $f13
; MIPSR6-N64-NEXT: mfc1 $2, $f0
; MIPSR6-N64-NEXT: fill.w $w0, $2
; MIPSR6-N64-NEXT: fexdo.h $w0, $w0, $w0
; MIPSR6-N64-NEXT: copy_u.h $2, $w0[0]
; MIPSR6-N64-NEXT: jr $ra
; MIPSR6-N64-NEXT: sh $2, 0($1)
entry:
%0 = load i16, i16* @g, align 2
%1 = call float @llvm.convert.from.fp16.f32(i16 %0)
; MIPS32-N32: madd.s $f[[F1:[0-9]]], $f13, $f[[F0]], $f12
; MIPS32-N64: madd.s $f[[F1:[0-9]]], $f13, $f[[F0]], $f12
%fmuladd = call float @llvm.fmuladd.f32(float %1, float %b, float %c)
%2 = call i16 @llvm.convert.to.fp16.f32(float %fmuladd)
store i16 %2, i16* @g, align 2
ret void
}
declare float @llvm.fabs.f32(float %Val)
define void @ffabs() {
; MIPS32-LABEL: ffabs:
; MIPS32: # %bb.0: # %entry
; MIPS32-NEXT: lui $2, %hi(_gp_disp)
; MIPS32-NEXT: addiu $2, $2, %lo(_gp_disp)
; MIPS32-NEXT: addu $1, $2, $25
; MIPS32-NEXT: lw $1, %got(g)($1)
; MIPS32-NEXT: lh $2, 0($1)
; MIPS32-NEXT: fill.h $w0, $2
; MIPS32-NEXT: fexupr.w $w0, $w0
; MIPS32-NEXT: copy_s.w $2, $w0[0]
; MIPS32-NEXT: mtc1 $2, $f0
; MIPS32-NEXT: abs.s $f0, $f0
; MIPS32-NEXT: mfc1 $2, $f0
; MIPS32-NEXT: fill.w $w0, $2
; MIPS32-NEXT: fexdo.h $w0, $w0, $w0
; MIPS32-NEXT: copy_u.h $2, $w0[0]
; MIPS32-NEXT: jr $ra
; MIPS32-NEXT: sh $2, 0($1)
;
; MIPS64-N32-LABEL: ffabs:
; MIPS64-N32: # %bb.0: # %entry
; MIPS64-N32-NEXT: lui $1, %hi(%neg(%gp_rel(ffabs)))
; MIPS64-N32-NEXT: addu $1, $1, $25
; MIPS64-N32-NEXT: addiu $1, $1, %lo(%neg(%gp_rel(ffabs)))
; MIPS64-N32-NEXT: lw $1, %got_disp(g)($1)
; MIPS64-N32-NEXT: lh $2, 0($1)
; MIPS64-N32-NEXT: fill.h $w0, $2
; MIPS64-N32-NEXT: fexupr.w $w0, $w0
; MIPS64-N32-NEXT: copy_s.w $2, $w0[0]
; MIPS64-N32-NEXT: mtc1 $2, $f0
; MIPS64-N32-NEXT: abs.s $f0, $f0
; MIPS64-N32-NEXT: mfc1 $2, $f0
; MIPS64-N32-NEXT: fill.w $w0, $2
; MIPS64-N32-NEXT: fexdo.h $w0, $w0, $w0
; MIPS64-N32-NEXT: copy_u.h $2, $w0[0]
; MIPS64-N32-NEXT: jr $ra
; MIPS64-N32-NEXT: sh $2, 0($1)
;
; MIPS64-N64-LABEL: ffabs:
; MIPS64-N64: # %bb.0: # %entry
; MIPS64-N64-NEXT: lui $1, %hi(%neg(%gp_rel(ffabs)))
; MIPS64-N64-NEXT: daddu $1, $1, $25
; MIPS64-N64-NEXT: daddiu $1, $1, %lo(%neg(%gp_rel(ffabs)))
; MIPS64-N64-NEXT: ld $1, %got_disp(g)($1)
; MIPS64-N64-NEXT: lh $2, 0($1)
; MIPS64-N64-NEXT: fill.h $w0, $2
; MIPS64-N64-NEXT: fexupr.w $w0, $w0
; MIPS64-N64-NEXT: copy_s.w $2, $w0[0]
; MIPS64-N64-NEXT: mtc1 $2, $f0
; MIPS64-N64-NEXT: abs.s $f0, $f0
; MIPS64-N64-NEXT: mfc1 $2, $f0
; MIPS64-N64-NEXT: fill.w $w0, $2
; MIPS64-N64-NEXT: fexdo.h $w0, $w0, $w0
; MIPS64-N64-NEXT: copy_u.h $2, $w0[0]
; MIPS64-N64-NEXT: jr $ra
; MIPS64-N64-NEXT: sh $2, 0($1)
entry:
%0 = load i16, i16* @g, align 2
%1 = call float @llvm.convert.from.fp16.f32(i16 %0)
%fabs = call float @llvm.fabs.f32(float %1)
%2 = call i16 @llvm.convert.to.fp16.f32(float %fabs)
store i16 %2, i16* @g, align 2
ret void
}
declare float @llvm.minnum.f32(float %Val, float %b)
define void @fminnum(float %b) {
; MIPS32-LABEL: fminnum:
; MIPS32: # %bb.0: # %entry
; MIPS32-NEXT: lui $2, %hi(_gp_disp)
; MIPS32-NEXT: addiu $2, $2, %lo(_gp_disp)
; MIPS32-NEXT: addiu $sp, $sp, -24
; MIPS32-NEXT: .cfi_def_cfa_offset 24
; MIPS32-NEXT: sw $ra, 20($sp) # 4-byte Folded Spill
; MIPS32-NEXT: sw $16, 16($sp) # 4-byte Folded Spill
; MIPS32-NEXT: .cfi_offset 31, -4
; MIPS32-NEXT: .cfi_offset 16, -8
; MIPS32-NEXT: addu $gp, $2, $25
; MIPS32-NEXT: mov.s $f14, $f12
; MIPS32-NEXT: lw $16, %got(g)($gp)
; MIPS32-NEXT: lh $1, 0($16)
; MIPS32-NEXT: fill.h $w0, $1
; MIPS32-NEXT: fexupr.w $w0, $w0
; MIPS32-NEXT: copy_s.w $1, $w0[0]
; MIPS32-NEXT: lw $25, %call16(fminf)($gp)
; MIPS32-NEXT: jalr $25
; MIPS32-NEXT: mtc1 $1, $f12
; MIPS32-NEXT: mfc1 $1, $f0
; MIPS32-NEXT: fill.w $w0, $1
; MIPS32-NEXT: fexdo.h $w0, $w0, $w0
; MIPS32-NEXT: copy_u.h $1, $w0[0]
; MIPS32-NEXT: sh $1, 0($16)
; MIPS32-NEXT: lw $16, 16($sp) # 4-byte Folded Reload
; MIPS32-NEXT: lw $ra, 20($sp) # 4-byte Folded Reload
; MIPS32-NEXT: jr $ra
; MIPS32-NEXT: addiu $sp, $sp, 24
;
; MIPS64-N32-LABEL: fminnum:
; MIPS64-N32: # %bb.0: # %entry
; MIPS64-N32-NEXT: addiu $sp, $sp, -32
; MIPS64-N32-NEXT: .cfi_def_cfa_offset 32
; MIPS64-N32-NEXT: sd $ra, 24($sp) # 8-byte Folded Spill
; MIPS64-N32-NEXT: sd $gp, 16($sp) # 8-byte Folded Spill
; MIPS64-N32-NEXT: sd $16, 8($sp) # 8-byte Folded Spill
; MIPS64-N32-NEXT: .cfi_offset 31, -8
; MIPS64-N32-NEXT: .cfi_offset 28, -16
; MIPS64-N32-NEXT: .cfi_offset 16, -24
; MIPS64-N32-NEXT: lui $1, %hi(%neg(%gp_rel(fminnum)))
; MIPS64-N32-NEXT: addu $1, $1, $25
; MIPS64-N32-NEXT: addiu $gp, $1, %lo(%neg(%gp_rel(fminnum)))
; MIPS64-N32-NEXT: mov.s $f13, $f12
; MIPS64-N32-NEXT: lw $16, %got_disp(g)($gp)
; MIPS64-N32-NEXT: lh $1, 0($16)
; MIPS64-N32-NEXT: fill.h $w0, $1
; MIPS64-N32-NEXT: fexupr.w $w0, $w0
; MIPS64-N32-NEXT: copy_s.w $1, $w0[0]
; MIPS64-N32-NEXT: lw $25, %call16(fminf)($gp)
; MIPS64-N32-NEXT: jalr $25
; MIPS64-N32-NEXT: mtc1 $1, $f12
; MIPS64-N32-NEXT: mfc1 $1, $f0
; MIPS64-N32-NEXT: fill.w $w0, $1
; MIPS64-N32-NEXT: fexdo.h $w0, $w0, $w0
; MIPS64-N32-NEXT: copy_u.h $1, $w0[0]
; MIPS64-N32-NEXT: sh $1, 0($16)
; MIPS64-N32-NEXT: ld $16, 8($sp) # 8-byte Folded Reload
; MIPS64-N32-NEXT: ld $gp, 16($sp) # 8-byte Folded Reload
; MIPS64-N32-NEXT: ld $ra, 24($sp) # 8-byte Folded Reload
; MIPS64-N32-NEXT: jr $ra
; MIPS64-N32-NEXT: addiu $sp, $sp, 32
;
; MIPS64-N64-LABEL: fminnum:
; MIPS64-N64: # %bb.0: # %entry
; MIPS64-N64-NEXT: daddiu $sp, $sp, -32
; MIPS64-N64-NEXT: .cfi_def_cfa_offset 32
; MIPS64-N64-NEXT: sd $ra, 24($sp) # 8-byte Folded Spill
; MIPS64-N64-NEXT: sd $gp, 16($sp) # 8-byte Folded Spill
; MIPS64-N64-NEXT: sd $16, 8($sp) # 8-byte Folded Spill
; MIPS64-N64-NEXT: .cfi_offset 31, -8
; MIPS64-N64-NEXT: .cfi_offset 28, -16
; MIPS64-N64-NEXT: .cfi_offset 16, -24
; MIPS64-N64-NEXT: lui $1, %hi(%neg(%gp_rel(fminnum)))
; MIPS64-N64-NEXT: daddu $1, $1, $25
; MIPS64-N64-NEXT: daddiu $gp, $1, %lo(%neg(%gp_rel(fminnum)))
; MIPS64-N64-NEXT: mov.s $f13, $f12
; MIPS64-N64-NEXT: ld $16, %got_disp(g)($gp)
; MIPS64-N64-NEXT: lh $1, 0($16)
; MIPS64-N64-NEXT: fill.h $w0, $1
; MIPS64-N64-NEXT: fexupr.w $w0, $w0
; MIPS64-N64-NEXT: copy_s.w $1, $w0[0]
; MIPS64-N64-NEXT: ld $25, %call16(fminf)($gp)
; MIPS64-N64-NEXT: jalr $25
; MIPS64-N64-NEXT: mtc1 $1, $f12
; MIPS64-N64-NEXT: mfc1 $1, $f0
; MIPS64-N64-NEXT: fill.w $w0, $1
; MIPS64-N64-NEXT: fexdo.h $w0, $w0, $w0
; MIPS64-N64-NEXT: copy_u.h $1, $w0[0]
; MIPS64-N64-NEXT: sh $1, 0($16)
; MIPS64-N64-NEXT: ld $16, 8($sp) # 8-byte Folded Reload
; MIPS64-N64-NEXT: ld $gp, 16($sp) # 8-byte Folded Reload
; MIPS64-N64-NEXT: ld $ra, 24($sp) # 8-byte Folded Reload
; MIPS64-N64-NEXT: jr $ra
; MIPS64-N64-NEXT: daddiu $sp, $sp, 32
entry:
%0 = load i16, i16* @g, align 2
%1 = call float @llvm.convert.from.fp16.f32(i16 %0)
%minnum = call float @llvm.minnum.f32(float %1, float %b)
%2 = call i16 @llvm.convert.to.fp16.f32(float %minnum)
store i16 %2, i16* @g, align 2
ret void
}
declare float @llvm.maxnum.f32(float %Val, float %b)
define void @fmaxnum(float %b) {
; MIPS32-LABEL: fmaxnum:
; MIPS32: # %bb.0: # %entry
; MIPS32-NEXT: lui $2, %hi(_gp_disp)
; MIPS32-NEXT: addiu $2, $2, %lo(_gp_disp)
; MIPS32-NEXT: addiu $sp, $sp, -24
; MIPS32-NEXT: .cfi_def_cfa_offset 24
; MIPS32-NEXT: sw $ra, 20($sp) # 4-byte Folded Spill
; MIPS32-NEXT: sw $16, 16($sp) # 4-byte Folded Spill
; MIPS32-NEXT: .cfi_offset 31, -4
; MIPS32-NEXT: .cfi_offset 16, -8
; MIPS32-NEXT: addu $gp, $2, $25
; MIPS32-NEXT: mov.s $f14, $f12
; MIPS32-NEXT: lw $16, %got(g)($gp)
; MIPS32-NEXT: lh $1, 0($16)
; MIPS32-NEXT: fill.h $w0, $1
; MIPS32-NEXT: fexupr.w $w0, $w0
; MIPS32-NEXT: copy_s.w $1, $w0[0]
; MIPS32-NEXT: lw $25, %call16(fmaxf)($gp)
; MIPS32-NEXT: jalr $25
; MIPS32-NEXT: mtc1 $1, $f12
; MIPS32-NEXT: mfc1 $1, $f0
; MIPS32-NEXT: fill.w $w0, $1
; MIPS32-NEXT: fexdo.h $w0, $w0, $w0
; MIPS32-NEXT: copy_u.h $1, $w0[0]
; MIPS32-NEXT: sh $1, 0($16)
; MIPS32-NEXT: lw $16, 16($sp) # 4-byte Folded Reload
; MIPS32-NEXT: lw $ra, 20($sp) # 4-byte Folded Reload
; MIPS32-NEXT: jr $ra
; MIPS32-NEXT: addiu $sp, $sp, 24
;
; MIPS64-N32-LABEL: fmaxnum:
; MIPS64-N32: # %bb.0: # %entry
; MIPS64-N32-NEXT: addiu $sp, $sp, -32
; MIPS64-N32-NEXT: .cfi_def_cfa_offset 32
; MIPS64-N32-NEXT: sd $ra, 24($sp) # 8-byte Folded Spill
; MIPS64-N32-NEXT: sd $gp, 16($sp) # 8-byte Folded Spill
; MIPS64-N32-NEXT: sd $16, 8($sp) # 8-byte Folded Spill
; MIPS64-N32-NEXT: .cfi_offset 31, -8
; MIPS64-N32-NEXT: .cfi_offset 28, -16
; MIPS64-N32-NEXT: .cfi_offset 16, -24
; MIPS64-N32-NEXT: lui $1, %hi(%neg(%gp_rel(fmaxnum)))
; MIPS64-N32-NEXT: addu $1, $1, $25
; MIPS64-N32-NEXT: addiu $gp, $1, %lo(%neg(%gp_rel(fmaxnum)))
; MIPS64-N32-NEXT: mov.s $f13, $f12
; MIPS64-N32-NEXT: lw $16, %got_disp(g)($gp)
; MIPS64-N32-NEXT: lh $1, 0($16)
; MIPS64-N32-NEXT: fill.h $w0, $1
; MIPS64-N32-NEXT: fexupr.w $w0, $w0
; MIPS64-N32-NEXT: copy_s.w $1, $w0[0]
; MIPS64-N32-NEXT: lw $25, %call16(fmaxf)($gp)
; MIPS64-N32-NEXT: jalr $25
; MIPS64-N32-NEXT: mtc1 $1, $f12
; MIPS64-N32-NEXT: mfc1 $1, $f0
; MIPS64-N32-NEXT: fill.w $w0, $1
; MIPS64-N32-NEXT: fexdo.h $w0, $w0, $w0
; MIPS64-N32-NEXT: copy_u.h $1, $w0[0]
; MIPS64-N32-NEXT: sh $1, 0($16)
; MIPS64-N32-NEXT: ld $16, 8($sp) # 8-byte Folded Reload
; MIPS64-N32-NEXT: ld $gp, 16($sp) # 8-byte Folded Reload
; MIPS64-N32-NEXT: ld $ra, 24($sp) # 8-byte Folded Reload
; MIPS64-N32-NEXT: jr $ra
; MIPS64-N32-NEXT: addiu $sp, $sp, 32
;
; MIPS64-N64-LABEL: fmaxnum:
; MIPS64-N64: # %bb.0: # %entry
; MIPS64-N64-NEXT: daddiu $sp, $sp, -32
; MIPS64-N64-NEXT: .cfi_def_cfa_offset 32
; MIPS64-N64-NEXT: sd $ra, 24($sp) # 8-byte Folded Spill
; MIPS64-N64-NEXT: sd $gp, 16($sp) # 8-byte Folded Spill
; MIPS64-N64-NEXT: sd $16, 8($sp) # 8-byte Folded Spill
; MIPS64-N64-NEXT: .cfi_offset 31, -8
; MIPS64-N64-NEXT: .cfi_offset 28, -16
; MIPS64-N64-NEXT: .cfi_offset 16, -24
; MIPS64-N64-NEXT: lui $1, %hi(%neg(%gp_rel(fmaxnum)))
; MIPS64-N64-NEXT: daddu $1, $1, $25
; MIPS64-N64-NEXT: daddiu $gp, $1, %lo(%neg(%gp_rel(fmaxnum)))
; MIPS64-N64-NEXT: mov.s $f13, $f12
; MIPS64-N64-NEXT: ld $16, %got_disp(g)($gp)
; MIPS64-N64-NEXT: lh $1, 0($16)
; MIPS64-N64-NEXT: fill.h $w0, $1
; MIPS64-N64-NEXT: fexupr.w $w0, $w0
; MIPS64-N64-NEXT: copy_s.w $1, $w0[0]
; MIPS64-N64-NEXT: ld $25, %call16(fmaxf)($gp)
; MIPS64-N64-NEXT: jalr $25
; MIPS64-N64-NEXT: mtc1 $1, $f12
; MIPS64-N64-NEXT: mfc1 $1, $f0
; MIPS64-N64-NEXT: fill.w $w0, $1
; MIPS64-N64-NEXT: fexdo.h $w0, $w0, $w0
; MIPS64-N64-NEXT: copy_u.h $1, $w0[0]
; MIPS64-N64-NEXT: sh $1, 0($16)
; MIPS64-N64-NEXT: ld $16, 8($sp) # 8-byte Folded Reload
; MIPS64-N64-NEXT: ld $gp, 16($sp) # 8-byte Folded Reload
; MIPS64-N64-NEXT: ld $ra, 24($sp) # 8-byte Folded Reload
; MIPS64-N64-NEXT: jr $ra
; MIPS64-N64-NEXT: daddiu $sp, $sp, 32
entry:
%0 = load i16, i16* @g, align 2
%1 = call float @llvm.convert.from.fp16.f32(i16 %0)
%maxnum = call float @llvm.maxnum.f32(float %1, float %b)
%2 = call i16 @llvm.convert.to.fp16.f32(float %maxnum)
store i16 %2, i16* @g, align 2
ret void
}
; This expansion of fcopysign could be done without converting f16 to float.
declare float @llvm.copysign.f32(float %Val, float %b)
define void @fcopysign(float %b) {
; MIPS32-LABEL: fcopysign:
; MIPS32: # %bb.0: # %entry
; MIPS32-NEXT: lui $2, %hi(_gp_disp)
; MIPS32-NEXT: addiu $2, $2, %lo(_gp_disp)
; MIPS32-NEXT: addu $1, $2, $25
; MIPS32-NEXT: lw $1, %got(g)($1)
; MIPS32-NEXT: lh $2, 0($1)
; MIPS32-NEXT: fill.h $w0, $2
; MIPS32-NEXT: fexupr.w $w0, $w0
; MIPS32-NEXT: copy_s.w $2, $w0[0]
; MIPS32-NEXT: mfc1 $3, $f12
; MIPS32-NEXT: ext $3, $3, 31, 1
; MIPS32-NEXT: ins $2, $3, 31, 1
; MIPS32-NEXT: fill.w $w0, $2
; MIPS32-NEXT: fexdo.h $w0, $w0, $w0
; MIPS32-NEXT: copy_u.h $2, $w0[0]
; MIPS32-NEXT: jr $ra
; MIPS32-NEXT: sh $2, 0($1)
;
; MIPS64-N32-LABEL: fcopysign:
; MIPS64-N32: # %bb.0: # %entry
; MIPS64-N32-NEXT: lui $1, %hi(%neg(%gp_rel(fcopysign)))
; MIPS64-N32-NEXT: addu $1, $1, $25
; MIPS64-N32-NEXT: addiu $1, $1, %lo(%neg(%gp_rel(fcopysign)))
; MIPS64-N32-NEXT: lw $1, %got_disp(g)($1)
; MIPS64-N32-NEXT: lh $2, 0($1)
; MIPS64-N32-NEXT: fill.h $w0, $2
; MIPS64-N32-NEXT: fexupr.w $w0, $w0
; MIPS64-N32-NEXT: copy_s.w $2, $w0[0]
; MIPS64-N32-NEXT: mfc1 $3, $f12
; MIPS64-N32-NEXT: ext $3, $3, 31, 1
; MIPS64-N32-NEXT: ins $2, $3, 31, 1
; MIPS64-N32-NEXT: fill.w $w0, $2
; MIPS64-N32-NEXT: fexdo.h $w0, $w0, $w0
; MIPS64-N32-NEXT: copy_u.h $2, $w0[0]
; MIPS64-N32-NEXT: jr $ra
; MIPS64-N32-NEXT: sh $2, 0($1)
;
; MIPS64-N64-LABEL: fcopysign:
; MIPS64-N64: # %bb.0: # %entry
; MIPS64-N64-NEXT: lui $1, %hi(%neg(%gp_rel(fcopysign)))
; MIPS64-N64-NEXT: daddu $1, $1, $25
; MIPS64-N64-NEXT: daddiu $1, $1, %lo(%neg(%gp_rel(fcopysign)))
; MIPS64-N64-NEXT: ld $1, %got_disp(g)($1)
; MIPS64-N64-NEXT: lh $2, 0($1)
; MIPS64-N64-NEXT: fill.h $w0, $2
; MIPS64-N64-NEXT: fexupr.w $w0, $w0
; MIPS64-N64-NEXT: copy_s.w $2, $w0[0]
; MIPS64-N64-NEXT: mfc1 $3, $f12
; MIPS64-N64-NEXT: ext $3, $3, 31, 1
; MIPS64-N64-NEXT: ins $2, $3, 31, 1
; MIPS64-N64-NEXT: fill.w $w0, $2
; MIPS64-N64-NEXT: fexdo.h $w0, $w0, $w0
; MIPS64-N64-NEXT: copy_u.h $2, $w0[0]
; MIPS64-N64-NEXT: jr $ra
; MIPS64-N64-NEXT: sh $2, 0($1)
entry:
%0 = load i16, i16* @g, align 2
%1 = call float @llvm.convert.from.fp16.f32(i16 %0)
%copysign = call float @llvm.copysign.f32(float %1, float %b)
%2 = call i16 @llvm.convert.to.fp16.f32(float %copysign)
store i16 %2, i16* @g, align 2
ret void
}
declare float @llvm.floor.f32(float %Val)
define void @ffloor() {
; MIPS32-LABEL: ffloor:
; MIPS32: # %bb.0: # %entry
; MIPS32-NEXT: lui $2, %hi(_gp_disp)
; MIPS32-NEXT: addiu $2, $2, %lo(_gp_disp)
; MIPS32-NEXT: addiu $sp, $sp, -24
; MIPS32-NEXT: .cfi_def_cfa_offset 24
; MIPS32-NEXT: sw $ra, 20($sp) # 4-byte Folded Spill
; MIPS32-NEXT: sw $16, 16($sp) # 4-byte Folded Spill
; MIPS32-NEXT: .cfi_offset 31, -4
; MIPS32-NEXT: .cfi_offset 16, -8
; MIPS32-NEXT: addu $gp, $2, $25
; MIPS32-NEXT: lw $16, %got(g)($gp)
; MIPS32-NEXT: lh $1, 0($16)
; MIPS32-NEXT: fill.h $w0, $1
; MIPS32-NEXT: fexupr.w $w0, $w0
; MIPS32-NEXT: copy_s.w $1, $w0[0]
; MIPS32-NEXT: lw $25, %call16(floorf)($gp)
; MIPS32-NEXT: jalr $25
; MIPS32-NEXT: mtc1 $1, $f12
; MIPS32-NEXT: mfc1 $1, $f0
; MIPS32-NEXT: fill.w $w0, $1
; MIPS32-NEXT: fexdo.h $w0, $w0, $w0
; MIPS32-NEXT: copy_u.h $1, $w0[0]
; MIPS32-NEXT: sh $1, 0($16)
; MIPS32-NEXT: lw $16, 16($sp) # 4-byte Folded Reload
; MIPS32-NEXT: lw $ra, 20($sp) # 4-byte Folded Reload
; MIPS32-NEXT: jr $ra
; MIPS32-NEXT: addiu $sp, $sp, 24
;
; MIPS64-N32-LABEL: ffloor:
; MIPS64-N32: # %bb.0: # %entry
; MIPS64-N32-NEXT: addiu $sp, $sp, -32
; MIPS64-N32-NEXT: .cfi_def_cfa_offset 32
; MIPS64-N32-NEXT: sd $ra, 24($sp) # 8-byte Folded Spill
; MIPS64-N32-NEXT: sd $gp, 16($sp) # 8-byte Folded Spill
; MIPS64-N32-NEXT: sd $16, 8($sp) # 8-byte Folded Spill
; MIPS64-N32-NEXT: .cfi_offset 31, -8
; MIPS64-N32-NEXT: .cfi_offset 28, -16
; MIPS64-N32-NEXT: .cfi_offset 16, -24
; MIPS64-N32-NEXT: lui $1, %hi(%neg(%gp_rel(ffloor)))
; MIPS64-N32-NEXT: addu $1, $1, $25
; MIPS64-N32-NEXT: addiu $gp, $1, %lo(%neg(%gp_rel(ffloor)))
; MIPS64-N32-NEXT: lw $16, %got_disp(g)($gp)
; MIPS64-N32-NEXT: lh $1, 0($16)
; MIPS64-N32-NEXT: fill.h $w0, $1
; MIPS64-N32-NEXT: fexupr.w $w0, $w0
; MIPS64-N32-NEXT: copy_s.w $1, $w0[0]
; MIPS64-N32-NEXT: lw $25, %call16(floorf)($gp)
; MIPS64-N32-NEXT: jalr $25
; MIPS64-N32-NEXT: mtc1 $1, $f12
; MIPS64-N32-NEXT: mfc1 $1, $f0
; MIPS64-N32-NEXT: fill.w $w0, $1
; MIPS64-N32-NEXT: fexdo.h $w0, $w0, $w0
; MIPS64-N32-NEXT: copy_u.h $1, $w0[0]
; MIPS64-N32-NEXT: sh $1, 0($16)
; MIPS64-N32-NEXT: ld $16, 8($sp) # 8-byte Folded Reload
; MIPS64-N32-NEXT: ld $gp, 16($sp) # 8-byte Folded Reload
; MIPS64-N32-NEXT: ld $ra, 24($sp) # 8-byte Folded Reload
; MIPS64-N32-NEXT: jr $ra
; MIPS64-N32-NEXT: addiu $sp, $sp, 32
;
; MIPS64-N64-LABEL: ffloor:
; MIPS64-N64: # %bb.0: # %entry
; MIPS64-N64-NEXT: daddiu $sp, $sp, -32
; MIPS64-N64-NEXT: .cfi_def_cfa_offset 32
; MIPS64-N64-NEXT: sd $ra, 24($sp) # 8-byte Folded Spill
; MIPS64-N64-NEXT: sd $gp, 16($sp) # 8-byte Folded Spill
; MIPS64-N64-NEXT: sd $16, 8($sp) # 8-byte Folded Spill
; MIPS64-N64-NEXT: .cfi_offset 31, -8
; MIPS64-N64-NEXT: .cfi_offset 28, -16
; MIPS64-N64-NEXT: .cfi_offset 16, -24
; MIPS64-N64-NEXT: lui $1, %hi(%neg(%gp_rel(ffloor)))
; MIPS64-N64-NEXT: daddu $1, $1, $25
; MIPS64-N64-NEXT: daddiu $gp, $1, %lo(%neg(%gp_rel(ffloor)))
; MIPS64-N64-NEXT: ld $16, %got_disp(g)($gp)
; MIPS64-N64-NEXT: lh $1, 0($16)
; MIPS64-N64-NEXT: fill.h $w0, $1
; MIPS64-N64-NEXT: fexupr.w $w0, $w0
; MIPS64-N64-NEXT: copy_s.w $1, $w0[0]
; MIPS64-N64-NEXT: ld $25, %call16(floorf)($gp)
; MIPS64-N64-NEXT: jalr $25
; MIPS64-N64-NEXT: mtc1 $1, $f12
; MIPS64-N64-NEXT: mfc1 $1, $f0
; MIPS64-N64-NEXT: fill.w $w0, $1
; MIPS64-N64-NEXT: fexdo.h $w0, $w0, $w0
; MIPS64-N64-NEXT: copy_u.h $1, $w0[0]
; MIPS64-N64-NEXT: sh $1, 0($16)
; MIPS64-N64-NEXT: ld $16, 8($sp) # 8-byte Folded Reload
; MIPS64-N64-NEXT: ld $gp, 16($sp) # 8-byte Folded Reload
; MIPS64-N64-NEXT: ld $ra, 24($sp) # 8-byte Folded Reload
; MIPS64-N64-NEXT: jr $ra
; MIPS64-N64-NEXT: daddiu $sp, $sp, 32
entry:
%0 = load i16, i16* @g, align 2
%1 = call float @llvm.convert.from.fp16.f32(i16 %0)
%floor = call float @llvm.floor.f32(float %1)
%2 = call i16 @llvm.convert.to.fp16.f32(float %floor)
store i16 %2, i16* @g, align 2
ret void
}
declare float @llvm.ceil.f32(float %Val)
define void @fceil() {
; MIPS32-LABEL: fceil:
; MIPS32: # %bb.0: # %entry
; MIPS32-NEXT: lui $2, %hi(_gp_disp)
; MIPS32-NEXT: addiu $2, $2, %lo(_gp_disp)
; MIPS32-NEXT: addiu $sp, $sp, -24
; MIPS32-NEXT: .cfi_def_cfa_offset 24
; MIPS32-NEXT: sw $ra, 20($sp) # 4-byte Folded Spill
; MIPS32-NEXT: sw $16, 16($sp) # 4-byte Folded Spill
; MIPS32-NEXT: .cfi_offset 31, -4
; MIPS32-NEXT: .cfi_offset 16, -8
; MIPS32-NEXT: addu $gp, $2, $25
; MIPS32-NEXT: lw $16, %got(g)($gp)
; MIPS32-NEXT: lh $1, 0($16)
; MIPS32-NEXT: fill.h $w0, $1
; MIPS32-NEXT: fexupr.w $w0, $w0
; MIPS32-NEXT: copy_s.w $1, $w0[0]
; MIPS32-NEXT: lw $25, %call16(ceilf)($gp)
; MIPS32-NEXT: jalr $25
; MIPS32-NEXT: mtc1 $1, $f12
; MIPS32-NEXT: mfc1 $1, $f0
; MIPS32-NEXT: fill.w $w0, $1
; MIPS32-NEXT: fexdo.h $w0, $w0, $w0
; MIPS32-NEXT: copy_u.h $1, $w0[0]
; MIPS32-NEXT: sh $1, 0($16)
; MIPS32-NEXT: lw $16, 16($sp) # 4-byte Folded Reload
; MIPS32-NEXT: lw $ra, 20($sp) # 4-byte Folded Reload
; MIPS32-NEXT: jr $ra
; MIPS32-NEXT: addiu $sp, $sp, 24
;
; MIPS64-N32-LABEL: fceil:
; MIPS64-N32: # %bb.0: # %entry
; MIPS64-N32-NEXT: addiu $sp, $sp, -32
; MIPS64-N32-NEXT: .cfi_def_cfa_offset 32
; MIPS64-N32-NEXT: sd $ra, 24($sp) # 8-byte Folded Spill
; MIPS64-N32-NEXT: sd $gp, 16($sp) # 8-byte Folded Spill
; MIPS64-N32-NEXT: sd $16, 8($sp) # 8-byte Folded Spill
; MIPS64-N32-NEXT: .cfi_offset 31, -8
; MIPS64-N32-NEXT: .cfi_offset 28, -16
; MIPS64-N32-NEXT: .cfi_offset 16, -24
; MIPS64-N32-NEXT: lui $1, %hi(%neg(%gp_rel(fceil)))
; MIPS64-N32-NEXT: addu $1, $1, $25
; MIPS64-N32-NEXT: addiu $gp, $1, %lo(%neg(%gp_rel(fceil)))
; MIPS64-N32-NEXT: lw $16, %got_disp(g)($gp)
; MIPS64-N32-NEXT: lh $1, 0($16)
; MIPS64-N32-NEXT: fill.h $w0, $1
; MIPS64-N32-NEXT: fexupr.w $w0, $w0
; MIPS64-N32-NEXT: copy_s.w $1, $w0[0]
; MIPS64-N32-NEXT: lw $25, %call16(ceilf)($gp)
; MIPS64-N32-NEXT: jalr $25
; MIPS64-N32-NEXT: mtc1 $1, $f12
; MIPS64-N32-NEXT: mfc1 $1, $f0
; MIPS64-N32-NEXT: fill.w $w0, $1
; MIPS64-N32-NEXT: fexdo.h $w0, $w0, $w0
; MIPS64-N32-NEXT: copy_u.h $1, $w0[0]
; MIPS64-N32-NEXT: sh $1, 0($16)
; MIPS64-N32-NEXT: ld $16, 8($sp) # 8-byte Folded Reload
; MIPS64-N32-NEXT: ld $gp, 16($sp) # 8-byte Folded Reload
; MIPS64-N32-NEXT: ld $ra, 24($sp) # 8-byte Folded Reload
; MIPS64-N32-NEXT: jr $ra
; MIPS64-N32-NEXT: addiu $sp, $sp, 32
;
; MIPS64-N64-LABEL: fceil:
; MIPS64-N64: # %bb.0: # %entry
; MIPS64-N64-NEXT: daddiu $sp, $sp, -32
; MIPS64-N64-NEXT: .cfi_def_cfa_offset 32
; MIPS64-N64-NEXT: sd $ra, 24($sp) # 8-byte Folded Spill
; MIPS64-N64-NEXT: sd $gp, 16($sp) # 8-byte Folded Spill
; MIPS64-N64-NEXT: sd $16, 8($sp) # 8-byte Folded Spill
; MIPS64-N64-NEXT: .cfi_offset 31, -8
; MIPS64-N64-NEXT: .cfi_offset 28, -16
; MIPS64-N64-NEXT: .cfi_offset 16, -24
; MIPS64-N64-NEXT: lui $1, %hi(%neg(%gp_rel(fceil)))
; MIPS64-N64-NEXT: daddu $1, $1, $25
; MIPS64-N64-NEXT: daddiu $gp, $1, %lo(%neg(%gp_rel(fceil)))
; MIPS64-N64-NEXT: ld $16, %got_disp(g)($gp)
; MIPS64-N64-NEXT: lh $1, 0($16)
; MIPS64-N64-NEXT: fill.h $w0, $1
; MIPS64-N64-NEXT: fexupr.w $w0, $w0
; MIPS64-N64-NEXT: copy_s.w $1, $w0[0]
; MIPS64-N64-NEXT: ld $25, %call16(ceilf)($gp)
; MIPS64-N64-NEXT: jalr $25
; MIPS64-N64-NEXT: mtc1 $1, $f12
; MIPS64-N64-NEXT: mfc1 $1, $f0
; MIPS64-N64-NEXT: fill.w $w0, $1
; MIPS64-N64-NEXT: fexdo.h $w0, $w0, $w0
; MIPS64-N64-NEXT: copy_u.h $1, $w0[0]
; MIPS64-N64-NEXT: sh $1, 0($16)
; MIPS64-N64-NEXT: ld $16, 8($sp) # 8-byte Folded Reload
; MIPS64-N64-NEXT: ld $gp, 16($sp) # 8-byte Folded Reload
; MIPS64-N64-NEXT: ld $ra, 24($sp) # 8-byte Folded Reload
; MIPS64-N64-NEXT: jr $ra
; MIPS64-N64-NEXT: daddiu $sp, $sp, 32
entry:
%0 = load i16, i16* @g, align 2
%1 = call float @llvm.convert.from.fp16.f32(i16 %0)
%ceil = call float @llvm.ceil.f32(float %1)
%2 = call i16 @llvm.convert.to.fp16.f32(float %ceil)
store i16 %2, i16* @g, align 2
ret void
}
declare float @llvm.trunc.f32(float %Val)
define void @ftrunc() {
; MIPS32-LABEL: ftrunc:
; MIPS32: # %bb.0: # %entry
; MIPS32-NEXT: lui $2, %hi(_gp_disp)
; MIPS32-NEXT: addiu $2, $2, %lo(_gp_disp)
; MIPS32-NEXT: addiu $sp, $sp, -24
; MIPS32-NEXT: .cfi_def_cfa_offset 24
; MIPS32-NEXT: sw $ra, 20($sp) # 4-byte Folded Spill
; MIPS32-NEXT: sw $16, 16($sp) # 4-byte Folded Spill
; MIPS32-NEXT: .cfi_offset 31, -4
; MIPS32-NEXT: .cfi_offset 16, -8
; MIPS32-NEXT: addu $gp, $2, $25
; MIPS32-NEXT: lw $16, %got(g)($gp)
; MIPS32-NEXT: lh $1, 0($16)
; MIPS32-NEXT: fill.h $w0, $1
; MIPS32-NEXT: fexupr.w $w0, $w0
; MIPS32-NEXT: copy_s.w $1, $w0[0]
; MIPS32-NEXT: lw $25, %call16(truncf)($gp)
; MIPS32-NEXT: jalr $25
; MIPS32-NEXT: mtc1 $1, $f12
; MIPS32-NEXT: mfc1 $1, $f0
; MIPS32-NEXT: fill.w $w0, $1
; MIPS32-NEXT: fexdo.h $w0, $w0, $w0
; MIPS32-NEXT: copy_u.h $1, $w0[0]
; MIPS32-NEXT: sh $1, 0($16)
; MIPS32-NEXT: lw $16, 16($sp) # 4-byte Folded Reload
; MIPS32-NEXT: lw $ra, 20($sp) # 4-byte Folded Reload
; MIPS32-NEXT: jr $ra
; MIPS32-NEXT: addiu $sp, $sp, 24
;
; MIPS64-N32-LABEL: ftrunc:
; MIPS64-N32: # %bb.0: # %entry
; MIPS64-N32-NEXT: addiu $sp, $sp, -32
; MIPS64-N32-NEXT: .cfi_def_cfa_offset 32
; MIPS64-N32-NEXT: sd $ra, 24($sp) # 8-byte Folded Spill
; MIPS64-N32-NEXT: sd $gp, 16($sp) # 8-byte Folded Spill
; MIPS64-N32-NEXT: sd $16, 8($sp) # 8-byte Folded Spill
; MIPS64-N32-NEXT: .cfi_offset 31, -8
; MIPS64-N32-NEXT: .cfi_offset 28, -16
; MIPS64-N32-NEXT: .cfi_offset 16, -24
; MIPS64-N32-NEXT: lui $1, %hi(%neg(%gp_rel(ftrunc)))
; MIPS64-N32-NEXT: addu $1, $1, $25
; MIPS64-N32-NEXT: addiu $gp, $1, %lo(%neg(%gp_rel(ftrunc)))
; MIPS64-N32-NEXT: lw $16, %got_disp(g)($gp)
; MIPS64-N32-NEXT: lh $1, 0($16)
; MIPS64-N32-NEXT: fill.h $w0, $1
; MIPS64-N32-NEXT: fexupr.w $w0, $w0
; MIPS64-N32-NEXT: copy_s.w $1, $w0[0]
; MIPS64-N32-NEXT: lw $25, %call16(truncf)($gp)
; MIPS64-N32-NEXT: jalr $25
; MIPS64-N32-NEXT: mtc1 $1, $f12
; MIPS64-N32-NEXT: mfc1 $1, $f0
; MIPS64-N32-NEXT: fill.w $w0, $1
; MIPS64-N32-NEXT: fexdo.h $w0, $w0, $w0
; MIPS64-N32-NEXT: copy_u.h $1, $w0[0]
; MIPS64-N32-NEXT: sh $1, 0($16)
; MIPS64-N32-NEXT: ld $16, 8($sp) # 8-byte Folded Reload
; MIPS64-N32-NEXT: ld $gp, 16($sp) # 8-byte Folded Reload
; MIPS64-N32-NEXT: ld $ra, 24($sp) # 8-byte Folded Reload
; MIPS64-N32-NEXT: jr $ra
; MIPS64-N32-NEXT: addiu $sp, $sp, 32
;
; MIPS64-N64-LABEL: ftrunc:
; MIPS64-N64: # %bb.0: # %entry
; MIPS64-N64-NEXT: daddiu $sp, $sp, -32
; MIPS64-N64-NEXT: .cfi_def_cfa_offset 32
; MIPS64-N64-NEXT: sd $ra, 24($sp) # 8-byte Folded Spill
; MIPS64-N64-NEXT: sd $gp, 16($sp) # 8-byte Folded Spill
; MIPS64-N64-NEXT: sd $16, 8($sp) # 8-byte Folded Spill
; MIPS64-N64-NEXT: .cfi_offset 31, -8
; MIPS64-N64-NEXT: .cfi_offset 28, -16
; MIPS64-N64-NEXT: .cfi_offset 16, -24
; MIPS64-N64-NEXT: lui $1, %hi(%neg(%gp_rel(ftrunc)))
; MIPS64-N64-NEXT: daddu $1, $1, $25
; MIPS64-N64-NEXT: daddiu $gp, $1, %lo(%neg(%gp_rel(ftrunc)))
; MIPS64-N64-NEXT: ld $16, %got_disp(g)($gp)
; MIPS64-N64-NEXT: lh $1, 0($16)
; MIPS64-N64-NEXT: fill.h $w0, $1
; MIPS64-N64-NEXT: fexupr.w $w0, $w0
; MIPS64-N64-NEXT: copy_s.w $1, $w0[0]
; MIPS64-N64-NEXT: ld $25, %call16(truncf)($gp)
; MIPS64-N64-NEXT: jalr $25
; MIPS64-N64-NEXT: mtc1 $1, $f12
; MIPS64-N64-NEXT: mfc1 $1, $f0
; MIPS64-N64-NEXT: fill.w $w0, $1
; MIPS64-N64-NEXT: fexdo.h $w0, $w0, $w0
; MIPS64-N64-NEXT: copy_u.h $1, $w0[0]
; MIPS64-N64-NEXT: sh $1, 0($16)
; MIPS64-N64-NEXT: ld $16, 8($sp) # 8-byte Folded Reload
; MIPS64-N64-NEXT: ld $gp, 16($sp) # 8-byte Folded Reload
; MIPS64-N64-NEXT: ld $ra, 24($sp) # 8-byte Folded Reload
; MIPS64-N64-NEXT: jr $ra
; MIPS64-N64-NEXT: daddiu $sp, $sp, 32
entry:
%0 = load i16, i16* @g, align 2
%1 = call float @llvm.convert.from.fp16.f32(i16 %0)
%trunc = call float @llvm.trunc.f32(float %1)
%2 = call i16 @llvm.convert.to.fp16.f32(float %trunc)
store i16 %2, i16* @g, align 2
ret void
}
declare float @llvm.rint.f32(float %Val)
define void @frint() {
; MIPS32-LABEL: frint:
; MIPS32: # %bb.0: # %entry
; MIPS32-NEXT: lui $2, %hi(_gp_disp)
; MIPS32-NEXT: addiu $2, $2, %lo(_gp_disp)
; MIPS32-NEXT: addiu $sp, $sp, -24
; MIPS32-NEXT: .cfi_def_cfa_offset 24
; MIPS32-NEXT: sw $ra, 20($sp) # 4-byte Folded Spill
; MIPS32-NEXT: sw $16, 16($sp) # 4-byte Folded Spill
; MIPS32-NEXT: .cfi_offset 31, -4
; MIPS32-NEXT: .cfi_offset 16, -8
; MIPS32-NEXT: addu $gp, $2, $25
; MIPS32-NEXT: lw $16, %got(g)($gp)
; MIPS32-NEXT: lh $1, 0($16)
; MIPS32-NEXT: fill.h $w0, $1
; MIPS32-NEXT: fexupr.w $w0, $w0
; MIPS32-NEXT: copy_s.w $1, $w0[0]
; MIPS32-NEXT: lw $25, %call16(rintf)($gp)
; MIPS32-NEXT: jalr $25
; MIPS32-NEXT: mtc1 $1, $f12
; MIPS32-NEXT: mfc1 $1, $f0
; MIPS32-NEXT: fill.w $w0, $1
; MIPS32-NEXT: fexdo.h $w0, $w0, $w0
; MIPS32-NEXT: copy_u.h $1, $w0[0]
; MIPS32-NEXT: sh $1, 0($16)
; MIPS32-NEXT: lw $16, 16($sp) # 4-byte Folded Reload
; MIPS32-NEXT: lw $ra, 20($sp) # 4-byte Folded Reload
; MIPS32-NEXT: jr $ra
; MIPS32-NEXT: addiu $sp, $sp, 24
;
; MIPS64-N32-LABEL: frint:
; MIPS64-N32: # %bb.0: # %entry
; MIPS64-N32-NEXT: addiu $sp, $sp, -32
; MIPS64-N32-NEXT: .cfi_def_cfa_offset 32
; MIPS64-N32-NEXT: sd $ra, 24($sp) # 8-byte Folded Spill
; MIPS64-N32-NEXT: sd $gp, 16($sp) # 8-byte Folded Spill
; MIPS64-N32-NEXT: sd $16, 8($sp) # 8-byte Folded Spill
; MIPS64-N32-NEXT: .cfi_offset 31, -8
; MIPS64-N32-NEXT: .cfi_offset 28, -16
; MIPS64-N32-NEXT: .cfi_offset 16, -24
; MIPS64-N32-NEXT: lui $1, %hi(%neg(%gp_rel(frint)))
; MIPS64-N32-NEXT: addu $1, $1, $25
; MIPS64-N32-NEXT: addiu $gp, $1, %lo(%neg(%gp_rel(frint)))
; MIPS64-N32-NEXT: lw $16, %got_disp(g)($gp)
; MIPS64-N32-NEXT: lh $1, 0($16)
; MIPS64-N32-NEXT: fill.h $w0, $1
; MIPS64-N32-NEXT: fexupr.w $w0, $w0
; MIPS64-N32-NEXT: copy_s.w $1, $w0[0]
; MIPS64-N32-NEXT: lw $25, %call16(rintf)($gp)
; MIPS64-N32-NEXT: jalr $25
; MIPS64-N32-NEXT: mtc1 $1, $f12
; MIPS64-N32-NEXT: mfc1 $1, $f0
; MIPS64-N32-NEXT: fill.w $w0, $1
; MIPS64-N32-NEXT: fexdo.h $w0, $w0, $w0
; MIPS64-N32-NEXT: copy_u.h $1, $w0[0]
; MIPS64-N32-NEXT: sh $1, 0($16)
; MIPS64-N32-NEXT: ld $16, 8($sp) # 8-byte Folded Reload
; MIPS64-N32-NEXT: ld $gp, 16($sp) # 8-byte Folded Reload
; MIPS64-N32-NEXT: ld $ra, 24($sp) # 8-byte Folded Reload
; MIPS64-N32-NEXT: jr $ra
; MIPS64-N32-NEXT: addiu $sp, $sp, 32
;
; MIPS64-N64-LABEL: frint:
; MIPS64-N64: # %bb.0: # %entry
; MIPS64-N64-NEXT: daddiu $sp, $sp, -32
; MIPS64-N64-NEXT: .cfi_def_cfa_offset 32
; MIPS64-N64-NEXT: sd $ra, 24($sp) # 8-byte Folded Spill
; MIPS64-N64-NEXT: sd $gp, 16($sp) # 8-byte Folded Spill
; MIPS64-N64-NEXT: sd $16, 8($sp) # 8-byte Folded Spill
; MIPS64-N64-NEXT: .cfi_offset 31, -8
; MIPS64-N64-NEXT: .cfi_offset 28, -16
; MIPS64-N64-NEXT: .cfi_offset 16, -24
; MIPS64-N64-NEXT: lui $1, %hi(%neg(%gp_rel(frint)))
; MIPS64-N64-NEXT: daddu $1, $1, $25
; MIPS64-N64-NEXT: daddiu $gp, $1, %lo(%neg(%gp_rel(frint)))
; MIPS64-N64-NEXT: ld $16, %got_disp(g)($gp)
; MIPS64-N64-NEXT: lh $1, 0($16)
; MIPS64-N64-NEXT: fill.h $w0, $1
; MIPS64-N64-NEXT: fexupr.w $w0, $w0
; MIPS64-N64-NEXT: copy_s.w $1, $w0[0]
; MIPS64-N64-NEXT: ld $25, %call16(rintf)($gp)
; MIPS64-N64-NEXT: jalr $25
; MIPS64-N64-NEXT: mtc1 $1, $f12
; MIPS64-N64-NEXT: mfc1 $1, $f0
; MIPS64-N64-NEXT: fill.w $w0, $1
; MIPS64-N64-NEXT: fexdo.h $w0, $w0, $w0
; MIPS64-N64-NEXT: copy_u.h $1, $w0[0]
; MIPS64-N64-NEXT: sh $1, 0($16)
; MIPS64-N64-NEXT: ld $16, 8($sp) # 8-byte Folded Reload
; MIPS64-N64-NEXT: ld $gp, 16($sp) # 8-byte Folded Reload
; MIPS64-N64-NEXT: ld $ra, 24($sp) # 8-byte Folded Reload
; MIPS64-N64-NEXT: jr $ra
; MIPS64-N64-NEXT: daddiu $sp, $sp, 32
entry:
%0 = load i16, i16* @g, align 2
%1 = call float @llvm.convert.from.fp16.f32(i16 %0)
%rint = call float @llvm.rint.f32(float %1)
%2 = call i16 @llvm.convert.to.fp16.f32(float %rint)
store i16 %2, i16* @g, align 2
ret void
}
declare float @llvm.nearbyint.f32(float %Val)
define void @fnearbyint() {
; MIPS32-LABEL: fnearbyint:
; MIPS32: # %bb.0: # %entry
; MIPS32-NEXT: lui $2, %hi(_gp_disp)
; MIPS32-NEXT: addiu $2, $2, %lo(_gp_disp)
; MIPS32-NEXT: addiu $sp, $sp, -24
; MIPS32-NEXT: .cfi_def_cfa_offset 24
; MIPS32-NEXT: sw $ra, 20($sp) # 4-byte Folded Spill
; MIPS32-NEXT: sw $16, 16($sp) # 4-byte Folded Spill
; MIPS32-NEXT: .cfi_offset 31, -4
; MIPS32-NEXT: .cfi_offset 16, -8
; MIPS32-NEXT: addu $gp, $2, $25
; MIPS32-NEXT: lw $16, %got(g)($gp)
; MIPS32-NEXT: lh $1, 0($16)
; MIPS32-NEXT: fill.h $w0, $1
; MIPS32-NEXT: fexupr.w $w0, $w0
; MIPS32-NEXT: copy_s.w $1, $w0[0]
; MIPS32-NEXT: lw $25, %call16(nearbyintf)($gp)
; MIPS32-NEXT: jalr $25
; MIPS32-NEXT: mtc1 $1, $f12
; MIPS32-NEXT: mfc1 $1, $f0
; MIPS32-NEXT: fill.w $w0, $1
; MIPS32-NEXT: fexdo.h $w0, $w0, $w0
; MIPS32-NEXT: copy_u.h $1, $w0[0]
; MIPS32-NEXT: sh $1, 0($16)
; MIPS32-NEXT: lw $16, 16($sp) # 4-byte Folded Reload
; MIPS32-NEXT: lw $ra, 20($sp) # 4-byte Folded Reload
; MIPS32-NEXT: jr $ra
; MIPS32-NEXT: addiu $sp, $sp, 24
;
; MIPS64-N32-LABEL: fnearbyint:
; MIPS64-N32: # %bb.0: # %entry
; MIPS64-N32-NEXT: addiu $sp, $sp, -32
; MIPS64-N32-NEXT: .cfi_def_cfa_offset 32
; MIPS64-N32-NEXT: sd $ra, 24($sp) # 8-byte Folded Spill
; MIPS64-N32-NEXT: sd $gp, 16($sp) # 8-byte Folded Spill
; MIPS64-N32-NEXT: sd $16, 8($sp) # 8-byte Folded Spill
; MIPS64-N32-NEXT: .cfi_offset 31, -8
; MIPS64-N32-NEXT: .cfi_offset 28, -16
; MIPS64-N32-NEXT: .cfi_offset 16, -24
; MIPS64-N32-NEXT: lui $1, %hi(%neg(%gp_rel(fnearbyint)))
; MIPS64-N32-NEXT: addu $1, $1, $25
; MIPS64-N32-NEXT: addiu $gp, $1, %lo(%neg(%gp_rel(fnearbyint)))
; MIPS64-N32-NEXT: lw $16, %got_disp(g)($gp)
; MIPS64-N32-NEXT: lh $1, 0($16)
; MIPS64-N32-NEXT: fill.h $w0, $1
; MIPS64-N32-NEXT: fexupr.w $w0, $w0
; MIPS64-N32-NEXT: copy_s.w $1, $w0[0]
; MIPS64-N32-NEXT: lw $25, %call16(nearbyintf)($gp)
; MIPS64-N32-NEXT: jalr $25
; MIPS64-N32-NEXT: mtc1 $1, $f12
; MIPS64-N32-NEXT: mfc1 $1, $f0
; MIPS64-N32-NEXT: fill.w $w0, $1
; MIPS64-N32-NEXT: fexdo.h $w0, $w0, $w0
; MIPS64-N32-NEXT: copy_u.h $1, $w0[0]
; MIPS64-N32-NEXT: sh $1, 0($16)
; MIPS64-N32-NEXT: ld $16, 8($sp) # 8-byte Folded Reload
; MIPS64-N32-NEXT: ld $gp, 16($sp) # 8-byte Folded Reload
; MIPS64-N32-NEXT: ld $ra, 24($sp) # 8-byte Folded Reload
; MIPS64-N32-NEXT: jr $ra
; MIPS64-N32-NEXT: addiu $sp, $sp, 32
;
; MIPS64-N64-LABEL: fnearbyint:
; MIPS64-N64: # %bb.0: # %entry
; MIPS64-N64-NEXT: daddiu $sp, $sp, -32
; MIPS64-N64-NEXT: .cfi_def_cfa_offset 32
; MIPS64-N64-NEXT: sd $ra, 24($sp) # 8-byte Folded Spill
; MIPS64-N64-NEXT: sd $gp, 16($sp) # 8-byte Folded Spill
; MIPS64-N64-NEXT: sd $16, 8($sp) # 8-byte Folded Spill
; MIPS64-N64-NEXT: .cfi_offset 31, -8
; MIPS64-N64-NEXT: .cfi_offset 28, -16
; MIPS64-N64-NEXT: .cfi_offset 16, -24
; MIPS64-N64-NEXT: lui $1, %hi(%neg(%gp_rel(fnearbyint)))
; MIPS64-N64-NEXT: daddu $1, $1, $25
; MIPS64-N64-NEXT: daddiu $gp, $1, %lo(%neg(%gp_rel(fnearbyint)))
; MIPS64-N64-NEXT: ld $16, %got_disp(g)($gp)
; MIPS64-N64-NEXT: lh $1, 0($16)
; MIPS64-N64-NEXT: fill.h $w0, $1
; MIPS64-N64-NEXT: fexupr.w $w0, $w0
; MIPS64-N64-NEXT: copy_s.w $1, $w0[0]
; MIPS64-N64-NEXT: ld $25, %call16(nearbyintf)($gp)
; MIPS64-N64-NEXT: jalr $25
; MIPS64-N64-NEXT: mtc1 $1, $f12
; MIPS64-N64-NEXT: mfc1 $1, $f0
; MIPS64-N64-NEXT: fill.w $w0, $1
; MIPS64-N64-NEXT: fexdo.h $w0, $w0, $w0
; MIPS64-N64-NEXT: copy_u.h $1, $w0[0]
; MIPS64-N64-NEXT: sh $1, 0($16)
; MIPS64-N64-NEXT: ld $16, 8($sp) # 8-byte Folded Reload
; MIPS64-N64-NEXT: ld $gp, 16($sp) # 8-byte Folded Reload
; MIPS64-N64-NEXT: ld $ra, 24($sp) # 8-byte Folded Reload
; MIPS64-N64-NEXT: jr $ra
; MIPS64-N64-NEXT: daddiu $sp, $sp, 32
entry:
%0 = load i16, i16* @g, align 2
%1 = call float @llvm.convert.from.fp16.f32(i16 %0)
%nearbyint = call float @llvm.nearbyint.f32(float %1)
%2 = call i16 @llvm.convert.to.fp16.f32(float %nearbyint)
store i16 %2, i16* @g, align 2
ret void
}
declare float @llvm.round.f32(float %Val)
define void @fround() {
; MIPS32-LABEL: fround:
; MIPS32: # %bb.0: # %entry
; MIPS32-NEXT: lui $2, %hi(_gp_disp)
; MIPS32-NEXT: addiu $2, $2, %lo(_gp_disp)
; MIPS32-NEXT: addiu $sp, $sp, -24
; MIPS32-NEXT: .cfi_def_cfa_offset 24
; MIPS32-NEXT: sw $ra, 20($sp) # 4-byte Folded Spill
; MIPS32-NEXT: sw $16, 16($sp) # 4-byte Folded Spill
; MIPS32-NEXT: .cfi_offset 31, -4
; MIPS32-NEXT: .cfi_offset 16, -8
; MIPS32-NEXT: addu $gp, $2, $25
; MIPS32-NEXT: lw $16, %got(g)($gp)
; MIPS32-NEXT: lh $1, 0($16)
; MIPS32-NEXT: fill.h $w0, $1
; MIPS32-NEXT: fexupr.w $w0, $w0
; MIPS32-NEXT: copy_s.w $1, $w0[0]
; MIPS32-NEXT: lw $25, %call16(roundf)($gp)
; MIPS32-NEXT: jalr $25
; MIPS32-NEXT: mtc1 $1, $f12
; MIPS32-NEXT: mfc1 $1, $f0
; MIPS32-NEXT: fill.w $w0, $1
; MIPS32-NEXT: fexdo.h $w0, $w0, $w0
; MIPS32-NEXT: copy_u.h $1, $w0[0]
; MIPS32-NEXT: sh $1, 0($16)
; MIPS32-NEXT: lw $16, 16($sp) # 4-byte Folded Reload
; MIPS32-NEXT: lw $ra, 20($sp) # 4-byte Folded Reload
; MIPS32-NEXT: jr $ra
; MIPS32-NEXT: addiu $sp, $sp, 24
;
; MIPS64-N32-LABEL: fround:
; MIPS64-N32: # %bb.0: # %entry
; MIPS64-N32-NEXT: addiu $sp, $sp, -32
; MIPS64-N32-NEXT: .cfi_def_cfa_offset 32
; MIPS64-N32-NEXT: sd $ra, 24($sp) # 8-byte Folded Spill
; MIPS64-N32-NEXT: sd $gp, 16($sp) # 8-byte Folded Spill
; MIPS64-N32-NEXT: sd $16, 8($sp) # 8-byte Folded Spill
; MIPS64-N32-NEXT: .cfi_offset 31, -8
; MIPS64-N32-NEXT: .cfi_offset 28, -16
; MIPS64-N32-NEXT: .cfi_offset 16, -24
; MIPS64-N32-NEXT: lui $1, %hi(%neg(%gp_rel(fround)))
; MIPS64-N32-NEXT: addu $1, $1, $25
; MIPS64-N32-NEXT: addiu $gp, $1, %lo(%neg(%gp_rel(fround)))
; MIPS64-N32-NEXT: lw $16, %got_disp(g)($gp)
; MIPS64-N32-NEXT: lh $1, 0($16)
; MIPS64-N32-NEXT: fill.h $w0, $1
; MIPS64-N32-NEXT: fexupr.w $w0, $w0
; MIPS64-N32-NEXT: copy_s.w $1, $w0[0]
; MIPS64-N32-NEXT: lw $25, %call16(roundf)($gp)
; MIPS64-N32-NEXT: jalr $25
; MIPS64-N32-NEXT: mtc1 $1, $f12
; MIPS64-N32-NEXT: mfc1 $1, $f0
; MIPS64-N32-NEXT: fill.w $w0, $1
; MIPS64-N32-NEXT: fexdo.h $w0, $w0, $w0
; MIPS64-N32-NEXT: copy_u.h $1, $w0[0]
; MIPS64-N32-NEXT: sh $1, 0($16)
; MIPS64-N32-NEXT: ld $16, 8($sp) # 8-byte Folded Reload
; MIPS64-N32-NEXT: ld $gp, 16($sp) # 8-byte Folded Reload
; MIPS64-N32-NEXT: ld $ra, 24($sp) # 8-byte Folded Reload
; MIPS64-N32-NEXT: jr $ra
; MIPS64-N32-NEXT: addiu $sp, $sp, 32
;
; MIPS64-N64-LABEL: fround:
; MIPS64-N64: # %bb.0: # %entry
; MIPS64-N64-NEXT: daddiu $sp, $sp, -32
; MIPS64-N64-NEXT: .cfi_def_cfa_offset 32
; MIPS64-N64-NEXT: sd $ra, 24($sp) # 8-byte Folded Spill
; MIPS64-N64-NEXT: sd $gp, 16($sp) # 8-byte Folded Spill
; MIPS64-N64-NEXT: sd $16, 8($sp) # 8-byte Folded Spill
; MIPS64-N64-NEXT: .cfi_offset 31, -8
; MIPS64-N64-NEXT: .cfi_offset 28, -16
; MIPS64-N64-NEXT: .cfi_offset 16, -24
; MIPS64-N64-NEXT: lui $1, %hi(%neg(%gp_rel(fround)))
; MIPS64-N64-NEXT: daddu $1, $1, $25
; MIPS64-N64-NEXT: daddiu $gp, $1, %lo(%neg(%gp_rel(fround)))
; MIPS64-N64-NEXT: ld $16, %got_disp(g)($gp)
; MIPS64-N64-NEXT: lh $1, 0($16)
; MIPS64-N64-NEXT: fill.h $w0, $1
; MIPS64-N64-NEXT: fexupr.w $w0, $w0
; MIPS64-N64-NEXT: copy_s.w $1, $w0[0]
; MIPS64-N64-NEXT: ld $25, %call16(roundf)($gp)
; MIPS64-N64-NEXT: jalr $25
; MIPS64-N64-NEXT: mtc1 $1, $f12
; MIPS64-N64-NEXT: mfc1 $1, $f0
; MIPS64-N64-NEXT: fill.w $w0, $1
; MIPS64-N64-NEXT: fexdo.h $w0, $w0, $w0
; MIPS64-N64-NEXT: copy_u.h $1, $w0[0]
; MIPS64-N64-NEXT: sh $1, 0($16)
; MIPS64-N64-NEXT: ld $16, 8($sp) # 8-byte Folded Reload
; MIPS64-N64-NEXT: ld $gp, 16($sp) # 8-byte Folded Reload
; MIPS64-N64-NEXT: ld $ra, 24($sp) # 8-byte Folded Reload
; MIPS64-N64-NEXT: jr $ra
; MIPS64-N64-NEXT: daddiu $sp, $sp, 32
entry:
%0 = load i16, i16* @g, align 2
%1 = call float @llvm.convert.from.fp16.f32(i16 %0)
%round = call float @llvm.round.f32(float %1)
%2 = call i16 @llvm.convert.to.fp16.f32(float %round)
store i16 %2, i16* @g, align 2
ret void
}
Reference in New Issue View Git Blame Copy Permalink