llvm/test/CodeGen/AMDGPU/load-global-f64.ll
Matt Arsenault df587174eb AMDGPU: Improve load/store of illegal types.
There was a combine before to handle the simple copy case.
Split this into handling loads and stores separately.

We might want to change how this handles some of the vector
extloads, since this can result in large code size increases.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274394 91177308-0d34-0410-b5e6-96231b3b80d8
2016-07-01 22:47:50 +00:00

95 lines
3.3 KiB
LLVM

; RUN: llc -march=amdgcn -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=GCN-NOHSA -check-prefix=FUNC %s
; RUN: llc -mtriple=amdgcn-amdhsa -mcpu=kaveri -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=GCN-HSA -check-prefix=FUNC %s
; RUN: llc -march=amdgcn -mcpu=tonga -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=GCN-NOHSA -check-prefix=FUNC %s
; FUNC-LABEL: {{^}}global_load_f64:
; GCN-NOHSA: buffer_load_dwordx2 [[VAL:v\[[0-9]+:[0-9]+\]]]
; GCN-NOHSA: buffer_store_dwordx2 [[VAL]]
; GCN-HSA: flat_load_dwordx2 [[VAL:v\[[0-9]+:[0-9]+\]]]
; GCN-HSA: flat_store_dwordx2 {{v\[[0-9]+:[0-9]+\]}}, [[VAL]]
define void @global_load_f64(double addrspace(1)* %out, double addrspace(1)* %in) #0 {
%ld = load double, double addrspace(1)* %in
store double %ld, double addrspace(1)* %out
ret void
}
; FUNC-LABEL: {{^}}global_load_v2f64:
; GCN-NOHSA: buffer_load_dwordx4
; GCN-HSA: flat_load_dwordx4
define void @global_load_v2f64(<2 x double> addrspace(1)* %out, <2 x double> addrspace(1)* %in) #0 {
entry:
%ld = load <2 x double>, <2 x double> addrspace(1)* %in
store <2 x double> %ld, <2 x double> addrspace(1)* %out
ret void
}
; FUNC-LABEL: {{^}}global_load_v3f64:
; GCN-NOHSA: buffer_load_dwordx4
; GCN-NOHSA: buffer_load_dwordx4
; GCN-HSA: flat_load_dwordx4
; GCN-HSA: flat_load_dwordx4
define void @global_load_v3f64(<3 x double> addrspace(1)* %out, <3 x double> addrspace(1)* %in) #0 {
entry:
%ld = load <3 x double>, <3 x double> addrspace(1)* %in
store <3 x double> %ld, <3 x double> addrspace(1)* %out
ret void
}
; FUNC-LABEL: {{^}}global_load_v4f64:
; GCN-NOHSA: buffer_load_dwordx4
; GCN-NOHSA: buffer_load_dwordx4
; GCN-HSA: flat_load_dwordx4
; GCN-HSA: flat_load_dwordx4
define void @global_load_v4f64(<4 x double> addrspace(1)* %out, <4 x double> addrspace(1)* %in) #0 {
entry:
%ld = load <4 x double>, <4 x double> addrspace(1)* %in
store <4 x double> %ld, <4 x double> addrspace(1)* %out
ret void
}
; FUNC-LABEL: {{^}}global_load_v8f64:
; GCN-NOHSA: buffer_load_dwordx4
; GCN-NOHSA: buffer_load_dwordx4
; GCN-NOHSA: buffer_load_dwordx4
; GCN-NOHSA: buffer_load_dwordx4
; GCN-HSA: flat_load_dwordx4
; GCN-HSA: flat_load_dwordx4
; GCN-HSA: flat_load_dwordx4
; GCN-HSA: flat_load_dwordx4
define void @global_load_v8f64(<8 x double> addrspace(1)* %out, <8 x double> addrspace(1)* %in) #0 {
entry:
%ld = load <8 x double>, <8 x double> addrspace(1)* %in
store <8 x double> %ld, <8 x double> addrspace(1)* %out
ret void
}
; FUNC-LABEL: {{^}}global_load_v16f64:
; GCN-NOHSA: buffer_load_dwordx4
; GCN-NOHSA: buffer_load_dwordx4
; GCN-NOHSA: buffer_load_dwordx4
; GCN-NOHSA: buffer_load_dwordx4
; GCN-NOHSA: buffer_load_dwordx4
; GCN-NOHSA: buffer_load_dwordx4
; GCN-NOHSA: buffer_load_dwordx4
; GCN-NOHSA: buffer_load_dwordx4
; GCN-HSA: flat_load_dwordx4
; GCN-HSA: flat_load_dwordx4
; GCN-HSA: flat_load_dwordx4
; GCN-HSA: flat_load_dwordx4
; GCN-HSA: flat_load_dwordx4
; GCN-HSA: flat_load_dwordx4
; GCN-HSA: flat_load_dwordx4
; GCN-HSA: flat_load_dwordx4
define void @global_load_v16f64(<16 x double> addrspace(1)* %out, <16 x double> addrspace(1)* %in) #0 {
entry:
%ld = load <16 x double>, <16 x double> addrspace(1)* %in
store <16 x double> %ld, <16 x double> addrspace(1)* %out
ret void
}
attributes #0 = { nounwind }