RPCS3/llvm-mirror
1
0
Fork 0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2025-01-25 13:27:52 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity

AMDGPU: Split private memory tests

Browse Source 
Reenable R600 testing

reviewer: arsenm

Differential Revision: http://reviews.llvm.org/D20031

llvm-svn: 269207
This commit is contained in:
Jan Vesely 2016-05-11 17:24:45 +00:00
parent cc684634d9
commit 6b5644eeff
3 changed files with 57 additions and 24 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

30
test/CodeGen/AMDGPU/amdgcn.private-memory.ll Normal file
View File

@ -0,0 +1,30 @@
; RUN: llc -show-mc-encoding -mattr=+promote-alloca -verify-machineinstrs -march=amdgcn < %s | FileCheck %s -check-prefix=SI-PROMOTE -check-prefix=SI -check-prefix=FUNC
; RUN: llc -show-mc-encoding -mattr=+promote-alloca -verify-machineinstrs -mtriple=amdgcn--amdhsa -mcpu=kaveri < %s | FileCheck %s -check-prefix=SI-PROMOTE -check-prefix=SI -check-prefix=FUNC -check-prefix=HSA-PROMOTE
; RUN: llc -show-mc-encoding -mattr=-promote-alloca -verify-machineinstrs -march=amdgcn < %s | FileCheck %s -check-prefix=SI-ALLOCA -check-prefix=SI -check-prefix=FUNC
; RUN: llc -show-mc-encoding -mattr=-promote-alloca -verify-machineinstrs -mtriple=amdgcn-amdhsa -mcpu=kaveri < %s | FileCheck %s -check-prefix=SI-ALLOCA -check-prefix=SI -check-prefix=FUNC -check-prefix=HSA-ALLOCA
; RUN: llc -show-mc-encoding -mattr=+promote-alloca -verify-machineinstrs -march=amdgcn -mcpu=tonga < %s | FileCheck %s -check-prefix=SI-PROMOTE -check-prefix=SI -check-prefix=FUNC
; RUN: llc -show-mc-encoding -mattr=-promote-alloca -verify-machineinstrs -march=amdgcn -mcpu=tonga < %s | FileCheck %s -check-prefix=SI-ALLOCA -check-prefix=SI -check-prefix=FUNC
declare i32 @llvm.amdgcn.workitem.id.x() nounwind readnone
; Make sure we don't overwrite workitem information with private memory
; FUNC-LABEL: {{^}}work_item_info:
; SI-NOT: v_mov_b32_e{{(32|64)}} v0
define void @work_item_info(i32 addrspace(1)* %out, i32 %in) {
entry:
%0 = alloca [2 x i32]
%1 = getelementptr [2 x i32], [2 x i32]* %0, i32 0, i32 0
%2 = getelementptr [2 x i32], [2 x i32]* %0, i32 0, i32 1
store i32 0, i32* %1
store i32 1, i32* %2
%3 = getelementptr [2 x i32], [2 x i32]* %0, i32 0, i32 %in
%4 = load i32, i32* %3
%5 = call i32 @llvm.amdgcn.workitem.id.x()
%6 = add i32 %4, %5
store i32 %6, i32 addrspace(1)* %out
ret void
}

25
test/CodeGen/AMDGPU/private-memory.ll → test/CodeGen/AMDGPU/amdgpu.private-memory.ll
View File

@ -1,3 +1,4 @@
; RUN: llc -march=r600 -mcpu=cypress < %s | FileCheck %s -check-prefix=R600 -check-prefix=FUNC
; RUN: llc -show-mc-encoding -mattr=+promote-alloca -verify-machineinstrs -march=amdgcn < %s | FileCheck %s -check-prefix=SI-PROMOTE -check-prefix=SI -check-prefix=FUNC
; RUN: llc -show-mc-encoding -mattr=+promote-alloca -verify-machineinstrs -mtriple=amdgcn--amdhsa -mcpu=kaveri < %s | FileCheck %s -check-prefix=SI-PROMOTE -check-prefix=SI -check-prefix=FUNC -check-prefix=HSA-PROMOTE
; RUN: llc -show-mc-encoding -mattr=-promote-alloca -verify-machineinstrs -march=amdgcn < %s | FileCheck %s -check-prefix=SI-ALLOCA -check-prefix=SI -check-prefix=FUNC
@ -8,8 +9,6 @@
; RUN: opt -S -mtriple=amdgcn-unknown-amdhsa -mcpu=kaveri -amdgpu-promote-alloca < %s | FileCheck -check-prefix=HSAOPT -check-prefix=OPT %s
; RUN: opt -S -mtriple=amdgcn-unknown-unknown -mcpu=kaveri -amdgpu-promote-alloca < %s | FileCheck -check-prefix=NOHSAOPT -check-prefix=OPT %s
declare i32 @llvm.amdgcn.workitem.id.x() nounwind readnone
; HSAOPT: @mova_same_clause.stack = internal unnamed_addr addrspace(3) global [256 x [5 x i32]] undef, align 4
; HSAOPT: @high_alignment.stack = internal unnamed_addr addrspace(3) global [256 x [8 x i32]] undef, align 16
@ -264,28 +263,6 @@ entry:
}
; Make sure we don't overwrite workitem information with private memory
; FUNC-LABEL: {{^}}work_item_info:
; R600-NOT: MOV T0.X
; Additional check in case the move ends up in the last slot
; R600-NOT: MOV * TO.X
; SI-NOT: v_mov_b32_e{{(32|64)}} v0
define void @work_item_info(i32 addrspace(1)* %out, i32 %in) {
entry:
%0 = alloca [2 x i32]
%1 = getelementptr [2 x i32], [2 x i32]* %0, i32 0, i32 0
%2 = getelementptr [2 x i32], [2 x i32]* %0, i32 0, i32 1
store i32 0, i32* %1
store i32 1, i32* %2
%3 = getelementptr [2 x i32], [2 x i32]* %0, i32 0, i32 %in
%4 = load i32, i32* %3
%5 = call i32 @llvm.amdgcn.workitem.id.x()
%6 = add i32 %4, %5
store i32 %6, i32 addrspace(1)* %out
ret void
}
; Test that two stack objects are not stored in the same register
; The second stack object should be in T3.X

26
test/CodeGen/AMDGPU/r600.private-memory.ll Normal file
View File

@ -0,0 +1,26 @@
; RUN: llc -march=r600 -mcpu=cypress < %s | FileCheck %s -check-prefix=R600 -check-prefix=FUNC
declare i32 @llvm.r600.read.tidig.x() nounwind readnone
; Make sure we don't overwrite workitem information with private memory
; FUNC-LABEL: {{^}}work_item_info:
; R600-NOT: MOV T0.X
; Additional check in case the move ends up in the last slot
; R600-NOT: MOV * TO.X
define void @work_item_info(i32 addrspace(1)* %out, i32 %in) {
entry:
%0 = alloca [2 x i32]
%1 = getelementptr [2 x i32], [2 x i32]* %0, i32 0, i32 0
%2 = getelementptr [2 x i32], [2 x i32]* %0, i32 0, i32 1
store i32 0, i32* %1
store i32 1, i32* %2
%3 = getelementptr [2 x i32], [2 x i32]* %0, i32 0, i32 %in
%4 = load i32, i32* %3
%5 = call i32 @llvm.r600.read.tidig.x()
%6 = add i32 %4, %5
store i32 %6, i32 addrspace(1)* %out
ret void
}