1
0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-11-23 11:13:28 +01:00
llvm-mirror/test/CodeGen/AMDGPU/extload-private.ll
Tony ccb53c0a97 [AMDGPU] Add volatile support to SIMemoryLegalizer
Treat a non-atomic volatile load and store as a relaxed atomic at
system scope for the address spaces accessed. This will ensure all
relevant caches will be bypassed.

A volatile atomic is not changed and still only bypasses caches upto
the level specified by the SyncScope operand.

Differential Revision: https://reviews.llvm.org/D94214
2021-01-09 00:52:33 +00:00

47 lines
1.7 KiB
LLVM

; RUN: llc -march=amdgcn -mattr=-promote-alloca -amdgpu-sroa=0 -verify-machineinstrs < %s | FileCheck -check-prefix=SI -check-prefix=FUNC %s
; RUN: llc -march=amdgcn -mcpu=tonga -mattr=-promote-alloca -amdgpu-sroa=0 -verify-machineinstrs < %s | FileCheck -check-prefix=SI -check-prefix=FUNC %s
; FUNC-LABEL: {{^}}load_i8_sext_private:
; SI: buffer_load_sbyte v{{[0-9]+}}, off, s[{{[0-9]+:[0-9]+}}], 0 offset:4{{$}}
define amdgpu_kernel void @load_i8_sext_private(i32 addrspace(1)* %out) {
entry:
%tmp0 = alloca i8, addrspace(5)
%tmp1 = load i8, i8 addrspace(5)* %tmp0
%tmp2 = sext i8 %tmp1 to i32
store i32 %tmp2, i32 addrspace(1)* %out
ret void
}
; FUNC-LABEL: {{^}}load_i8_zext_private:
; SI: buffer_load_ubyte v{{[0-9]+}}, off, s[{{[0-9]+:[0-9]+}}], 0 offset:4{{$}}
define amdgpu_kernel void @load_i8_zext_private(i32 addrspace(1)* %out) {
entry:
%tmp0 = alloca i8, addrspace(5)
%tmp1 = load i8, i8 addrspace(5)* %tmp0
%tmp2 = zext i8 %tmp1 to i32
store i32 %tmp2, i32 addrspace(1)* %out
ret void
}
; FUNC-LABEL: {{^}}load_i16_sext_private:
; SI: buffer_load_sshort v{{[0-9]+}}, off, s[{{[0-9]+:[0-9]+}}], 0 offset:4{{$}}
define amdgpu_kernel void @load_i16_sext_private(i32 addrspace(1)* %out) {
entry:
%tmp0 = alloca i16, addrspace(5)
%tmp1 = load i16, i16 addrspace(5)* %tmp0
%tmp2 = sext i16 %tmp1 to i32
store i32 %tmp2, i32 addrspace(1)* %out
ret void
}
; FUNC-LABEL: {{^}}load_i16_zext_private:
; SI: buffer_load_ushort v{{[0-9]+}}, off, s[{{[0-9]+:[0-9]+}}], 0 offset:4 glc{{$}}
define amdgpu_kernel void @load_i16_zext_private(i32 addrspace(1)* %out) {
entry:
%tmp0 = alloca i16, addrspace(5)
%tmp1 = load volatile i16, i16 addrspace(5)* %tmp0
%tmp2 = zext i16 %tmp1 to i32
store i32 %tmp2, i32 addrspace(1)* %out
ret void
}