ptxas is a proprietary compiler from Nvidia that can compile PTX to machine code (SASS). It has a lot of diagnostics to catch errors in PTX, which can be used to verify PTX output from llc. Set -DPXTAS_EXECUTABLE=/path/to/ptxas CMake option to enable it. If this option is not set, then ptxas is substituted to true which effectively disables all ptxas RUN lines. LLVM_PTXAS_EXECUTABLE environment variable takes precedence over the CMake option, and allows to override ptxas executable that is used for LIT without complete re-configuration. Differential Revision: https://reviews.llvm.org/D121727
34 lines
957 B
LLVM
34 lines
957 B
LLVM
; RUN: llc < %s -march=nvptx -mcpu=sm_20 | FileCheck %s
|
|
; RUN: %if ptxas %{ llc < %s -march=nvptx -mcpu=sm_20 | %ptxas-verify %}
|
|
|
|
define void @foo(<2 x i32>* %a) {
|
|
; CHECK: .func foo
|
|
; CHECK: ld.v2.u32
|
|
; CHECK-NEXT: ld.v2.u32
|
|
; CHECK-NEXT: ld.v2.u32
|
|
; CHECK-NEXT: ld.v2.u32
|
|
; CHECK-NEXT: add.s32
|
|
; CHECK-NEXT: add.s32
|
|
; CHECK-NEXT: add.s32
|
|
; CHECK-NEXT: add.s32
|
|
; CHECK-NEXT: add.s32
|
|
; CHECK-NEXT: add.s32
|
|
%ptr0 = getelementptr <2 x i32>, <2 x i32>* %a, i32 0
|
|
%val0 = load <2 x i32>, <2 x i32>* %ptr0
|
|
%ptr1 = getelementptr <2 x i32>, <2 x i32>* %a, i32 1
|
|
%val1 = load <2 x i32>, <2 x i32>* %ptr1
|
|
%ptr2 = getelementptr <2 x i32>, <2 x i32>* %a, i32 2
|
|
%val2 = load <2 x i32>, <2 x i32>* %ptr2
|
|
%ptr3 = getelementptr <2 x i32>, <2 x i32>* %a, i32 3
|
|
%val3 = load <2 x i32>, <2 x i32>* %ptr3
|
|
|
|
%t0 = add <2 x i32> %val0, %val1
|
|
%t1 = add <2 x i32> %t0, %val2
|
|
%t2 = add <2 x i32> %t1, %val3
|
|
|
|
store <2 x i32> %t2, <2 x i32>* %a
|
|
|
|
ret void
|
|
}
|
|
|