130080fab1
Don't count register uses when determining the maximum number of registers used by a function. Count only the defs. This is really an underestimate of the true register usage, but in practice that's not a problem because if a function uses a register, then it has either defined it earlier, or some other function that executed before has defined it. In particular, the register counts are used: 1. When launching an entry function - in which case we're safe because the register counts of the entry function will include the register counts of all callees. 2. At function boundaries in dynamic VGPR mode. In this case it's safe because whenever we set the new VGPR allocation we take into account the outgoing_vgpr_count set by the middle-end. The main advantage of doing this is that the artificial VGPR arguments used only for preserving the inactive lanes when using the llvm.amdgcn.init.whole.wave intrinsic are no longer counted. This enables us to allocate only the registers we need in dynamic VGPR mode. --------- Co-authored-by: Thomas Symalla <5754458+tsymalla@users.noreply.github.com>
137 lines
3.4 KiB
LLVM
137 lines
3.4 KiB
LLVM
; RUN: llc -mtriple=amdgcn-amd-amdhsa -verify-machineinstrs -enable-ipra < %s | FileCheck -check-prefix=GCN %s
|
|
; RUN: llc -mtriple=amdgcn-amd-amdhsa -verify-machineinstrs < %s | FileCheck -check-prefix=GCN %s
|
|
|
|
; Kernels are not called, so there is no call preserved mask.
|
|
; GCN-LABEL: {{^}}kernel:
|
|
; GCN: flat_store_dword
|
|
define amdgpu_kernel void @kernel(ptr addrspace(1) %out) #0 {
|
|
entry:
|
|
store i32 0, ptr addrspace(1) %out
|
|
ret void
|
|
}
|
|
|
|
; GCN-LABEL: {{^}}func:
|
|
; GCN: ; NumVgprs: 8
|
|
define hidden void @func() #1 {
|
|
call void asm sideeffect "", "~{v0},~{v1},~{v2},~{v3},~{v4},~{v5},~{v6},~{v7}"() #0
|
|
ret void
|
|
}
|
|
|
|
; GCN-LABEL: {{^}}kernel_call:
|
|
; GCN-NOT: buffer_store
|
|
; GCN-NOT: buffer_load
|
|
; GCN-NOT: readlane
|
|
; GCN-NOT: writelane
|
|
; GCN: flat_load_dword v8
|
|
; GCN: s_swappc_b64
|
|
; GCN-NOT: buffer_store
|
|
; GCN-NOT: buffer_load
|
|
; GCN-NOT: readlane
|
|
; GCN-NOT: writelane
|
|
; GCN: flat_store_dword v{{\[[0-9]+:[0-9]+\]}}, v8
|
|
|
|
; GCN: ; TotalNumSgprs: 37
|
|
; GCN: ; NumVgprs: 9
|
|
define amdgpu_kernel void @kernel_call() #0 {
|
|
%vgpr = load volatile i32, ptr addrspace(1) poison
|
|
tail call void @func()
|
|
store volatile i32 %vgpr, ptr addrspace(1) poison
|
|
ret void
|
|
}
|
|
|
|
; GCN-LABEL: {{^}}func_regular_call:
|
|
; GCN-NOT: buffer_load
|
|
; GCN-NOT: readlane
|
|
; GCN: flat_load_dword v8
|
|
; GCN: s_swappc_b64
|
|
; GCN-NOT: buffer_load
|
|
; GCN-NOT: readlane
|
|
; GCN: flat_store_dword v{{\[[0-9]+:[0-9]+\]}}, v8
|
|
|
|
; GCN: ; TotalNumSgprs: 34
|
|
; GCN: ; NumVgprs: 10
|
|
define void @func_regular_call() #1 {
|
|
%vgpr = load volatile i32, ptr addrspace(1) poison
|
|
tail call void @func()
|
|
store volatile i32 %vgpr, ptr addrspace(1) poison
|
|
ret void
|
|
}
|
|
|
|
; GCN-LABEL: {{^}}func_tail_call:
|
|
; GCN: s_waitcnt
|
|
; GCN-NEXT: s_getpc_b64 s[16:17]
|
|
; GCN-NEXT: s_add_u32 s16,
|
|
; GCN-NEXT: s_addc_u32 s17,
|
|
; GCN-NEXT: s_setpc_b64 s[16:17]
|
|
|
|
; GCN: ; TotalNumSgprs: 18
|
|
; GCN: ; NumVgprs: 8
|
|
define void @func_tail_call() #1 {
|
|
tail call void @func()
|
|
ret void
|
|
}
|
|
|
|
; GCN-LABEL: {{^}}func_call_tail_call:
|
|
; GCN: flat_load_dword v8
|
|
; GCN: s_swappc_b64
|
|
; GCN: flat_store_dword v{{\[[0-9]+:[0-9]+\]}}, v8
|
|
; GCN: s_setpc_b64
|
|
|
|
; GCN: ; TotalNumSgprs: 34
|
|
; GCN: ; NumVgprs: 10
|
|
define void @func_call_tail_call() #1 {
|
|
%vgpr = load volatile i32, ptr addrspace(1) poison
|
|
tail call void @func()
|
|
store volatile i32 %vgpr, ptr addrspace(1) poison
|
|
tail call void @func()
|
|
ret void
|
|
}
|
|
|
|
; GCN-LABEL: {{^}}void_func_void:
|
|
define void @void_func_void() noinline {
|
|
ret void
|
|
}
|
|
|
|
; Make sure we don't get save/restore of FP between calls.
|
|
; GCN-LABEL: {{^}}test_funcx2:
|
|
; GCN: s_getpc_b64
|
|
; GCN-NOT: s32
|
|
; GCN: s_swappc_b64
|
|
; GCN-NOT: s32
|
|
; GCN: s_swappc_b64
|
|
define void @test_funcx2() #0 {
|
|
call void @void_func_void()
|
|
call void @void_func_void()
|
|
ret void
|
|
}
|
|
|
|
; Make sure we save/restore the return address around the call.
|
|
; Function Attrs: norecurse
|
|
define internal void @hoge() #2 {
|
|
bb:
|
|
; GCN-LABEL: {{^}}hoge:
|
|
; GCN-DAG: v_writelane_b32 [[CSR_VGPR:v[0-9]+]], s30,
|
|
; GCN-DAG: v_writelane_b32 [[CSR_VGPR]], s31,
|
|
; GCN: s_swappc_b64 s[30:31]
|
|
; GCN-DAG: v_readlane_b32 s30, [[CSR_VGPR]],
|
|
; GCN-DAG: v_readlane_b32 s31, [[CSR_VGPR]],
|
|
; GCN: s_waitcnt vmcnt(0)
|
|
; GCN: s_setpc_b64 s[30:31]
|
|
call void @eggs()
|
|
ret void
|
|
}
|
|
|
|
; GCN-LABEL: {{^}}wombat:
|
|
define weak amdgpu_kernel void @wombat(ptr %arg, ptr %arg2) {
|
|
bb:
|
|
call void @hoge() #0
|
|
ret void
|
|
}
|
|
|
|
declare dso_local void @eggs()
|
|
|
|
|
|
attributes #0 = { nounwind }
|
|
attributes #1 = { nounwind noinline "amdgpu-no-workitem-id-x" "amdgpu-no-workitem-id-y" "amdgpu-no-workitem-id-z" }
|
|
attributes #2 = { norecurse }
|