RPCS3/llvm-mirror
1
0
Fork 0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-11-25 20:23:11 +01:00
Code Issues Projects Releases Wiki Activity

[OpenMP] Add new execution mode for SPMD execution with Generic semantics

Browse Source 
Qualified kernels can be transformed from generic-mode to SPMD mode using an
optimization in OpenMPOpt. This patch introduces a new execution mode to
indicate kernels that have been transformed from generic-mode to SPMD-mode.
These kernels have SPMD-mode execution, but need generic-mode semantics for
scheduling the blocks and threads. Without this far too few blocks will be
scheduled for a generic region as SPMD mode expects the trip count to be
divided by the number of threads.

Reviewed By: ggeorgakoudis

Differential Revision: https://reviews.llvm.org/D106460
This commit is contained in:
Joseph Huber 2021-07-21 12:48:39 -04:00 committed by Huber, Joseph
parent 472a223072
commit 2c3ddf5d6f
3 changed files with 8 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

8
lib/Transforms/IPO/OpenMPOpt.cpp
View File

@ -2886,8 +2886,12 @@ struct AAKernelInfoFunction : AAKernelInfo {
assert(ExecMode->getInitializer() && assert(ExecMode->getInitializer() &&
ExecMode->getInitializer()->isOneValue() && ExecMode->getInitializer()->isOneValue() &&
"Initially non-SPMD kernel has SPMD exec mode!"); "Initially non-SPMD kernel has SPMD exec mode!");
ExecMode->setInitializer(
ConstantInt::get(ExecMode->getInitializer()->getType(), 0)); // Set the global exec mode flag to indicate SPMD-Generic mode.
constexpr int SPMDGeneric = 2;
if (!ExecMode->getInitializer()->isZeroValue())
ExecMode->setInitializer(
ConstantInt::get(ExecMode->getInitializer()->getType(), SPMDGeneric));
// Next rewrite the init and deinit calls to indicate we use SPMD-mode now. // Next rewrite the init and deinit calls to indicate we use SPMD-mode now.
const int InitIsSPMDArgNo = 1; const int InitIsSPMDArgNo = 1;

2
test/Transforms/OpenMP/is_spmd_exec_mode_fold.ll
View File

@ -13,7 +13,7 @@ target triple = "nvptx64"
;. ;.
; CHECK: @[[IS_SPMD_EXEC_MODE:[a-zA-Z0-9_$"\\.-]+]] = weak constant i8 0 ; CHECK: @[[IS_SPMD_EXEC_MODE:[a-zA-Z0-9_$"\\.-]+]] = weak constant i8 0
; CHECK: @[[WILL_BE_SPMD_EXEC_MODE:[a-zA-Z0-9_$"\\.-]+]] = weak constant i8 0 ; CHECK: @[[WILL_BE_SPMD_EXEC_MODE:[a-zA-Z0-9_$"\\.-]+]] = weak constant i8 2
; CHECK: @[[NON_SPMD_EXEC_MODE:[a-zA-Z0-9_$"\\.-]+]] = weak constant i8 1 ; CHECK: @[[NON_SPMD_EXEC_MODE:[a-zA-Z0-9_$"\\.-]+]] = weak constant i8 1
; CHECK: @[[WILL_NOT_BE_SPMD_EXEC_MODE:[a-zA-Z0-9_$"\\.-]+]] = weak constant i8 1 ; CHECK: @[[WILL_NOT_BE_SPMD_EXEC_MODE:[a-zA-Z0-9_$"\\.-]+]] = weak constant i8 1
; CHECK: @[[G:[a-zA-Z0-9_$"\\.-]+]] = external global i8 ; CHECK: @[[G:[a-zA-Z0-9_$"\\.-]+]] = external global i8

2
test/Transforms/OpenMP/spmdization.ll
View File

@ -32,7 +32,7 @@ target triple = "nvptx64"
;. ;.
; CHECK: @[[GLOB0:[0-9]+]] = private unnamed_addr constant [23 x i8] c" ; CHECK: @[[GLOB0:[0-9]+]] = private unnamed_addr constant [23 x i8] c"
; CHECK: @[[GLOB1:[0-9]+]] = private unnamed_addr constant [[STRUCT_IDENT_T:%.*]] { i32 0, i32 2, i32 0, i32 0, i8* getelementptr inbounds ([23 x i8], [23 x i8]* @[[GLOB0]], i32 0, i32 0) }, align 8 ; CHECK: @[[GLOB1:[0-9]+]] = private unnamed_addr constant [[STRUCT_IDENT_T:%.*]] { i32 0, i32 2, i32 0, i32 0, i8* getelementptr inbounds ([23 x i8], [23 x i8]* @[[GLOB0]], i32 0, i32 0) }, align 8
; CHECK: @[[__OMP_OFFLOADING_2C_38C77_SEQUENTIAL_LOOP_L4_EXEC_MODE:[a-zA-Z0-9_$"\\.-]+]] = weak constant i8 0 ; CHECK: @[[__OMP_OFFLOADING_2C_38C77_SEQUENTIAL_LOOP_L4_EXEC_MODE:[a-zA-Z0-9_$"\\.-]+]] = weak constant i8 2
; CHECK: @[[LLVM_COMPILER_USED:[a-zA-Z0-9_$"\\.-]+]] = appending global [1 x i8*] [i8* @__omp_offloading_2c_38c77_sequential_loop_l4_exec_mode], section "llvm.metadata" ; CHECK: @[[LLVM_COMPILER_USED:[a-zA-Z0-9_$"\\.-]+]] = appending global [1 x i8*] [i8* @__omp_offloading_2c_38c77_sequential_loop_l4_exec_mode], section "llvm.metadata"
;. ;.
define weak void @__omp_offloading_2c_38c77_sequential_loop_l4() #0 { define weak void @__omp_offloading_2c_38c77_sequential_loop_l4() #0 {