replica
/
hanchenye-llvm-project
2
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

GPGPU: add host iterators to kernel arguments 

llvm-svn: 275954
This commit is contained in:
Tobias Grosser 2016-07-19 07:32:55 +00:00
parent 472f9654c8
commit f6044bd0ef
2 changed files with 26 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

15
polly/lib/CodeGen/PPCGCodeGeneration.cpp
View File

@ -140,7 +140,7 @@ private:
/// The kernel function takes as arguments: /// The kernel function takes as arguments:
/// ///
/// - One i8 pointer for each external array reference used in the kernel. /// - One i8 pointer for each external array reference used in the kernel.
/// - Host iterators (TODO) /// - Host iterators
/// - Parameters (TODO) /// - Parameters (TODO)
/// - Other LLVM Value references (TODO) /// - Other LLVM Value references (TODO)
/// ///
@ -223,6 +223,11 @@ Function *GPUNodeBuilder::createKernelFunctionDecl(ppcg_kernel *Kernel) {
Args.push_back(Builder.getInt8PtrTy()); Args.push_back(Builder.getInt8PtrTy());
} }
int NumHostIters = isl_space_dim(Kernel->space, isl_dim_set);
for (long i = 0; i < NumHostIters; i++)
Args.push_back(Builder.getInt64Ty());
auto *FT = FunctionType::get(Builder.getVoidTy(), Args, false); auto *FT = FunctionType::get(Builder.getVoidTy(), Args, false);
auto *FN = Function::Create(FT, Function::ExternalLinkage, Identifier, auto *FN = Function::Create(FT, Function::ExternalLinkage, Identifier,
GPUModule.get()); GPUModule.get());
@ -237,6 +242,14 @@ Function *GPUNodeBuilder::createKernelFunctionDecl(ppcg_kernel *Kernel) {
Arg++; Arg++;
} }
for (long i = 0; i < NumHostIters; i++) {
isl_id *Id = isl_space_get_dim_id(Kernel->space, isl_dim_set, i);
Arg->setName(isl_id_get_name(Id));
IDToValue[Id] = &*Arg;
KernelIDs.insert(std::unique_ptr<isl_id, IslIdDeleter>(Id));
Arg++;
}
return FN; return FN;
} }

12
polly/test/GPGPU/host-control-flow.ll
View File

@ -1,6 +1,9 @@
; RUN: opt %loadPolly -polly-codegen-ppcg -disable-output \ ; RUN: opt %loadPolly -polly-codegen-ppcg -disable-output \
; RUN: -polly-acc-dump-code < %s | FileCheck %s -check-prefix=CODE ; RUN: -polly-acc-dump-code < %s | FileCheck %s -check-prefix=CODE
; RUN: opt %loadPolly -polly-codegen-ppcg -disable-output \
; RUN: -polly-acc-dump-kernel-ir < %s | FileCheck %s -check-prefix=KERNEL-IR
; RUN: opt %loadPolly -polly-codegen-ppcg \ ; RUN: opt %loadPolly -polly-codegen-ppcg \
; RUN: -S < %s | FileCheck %s -check-prefix=IR ; RUN: -S < %s | FileCheck %s -check-prefix=IR
; void foo(float A[2][100]) { ; void foo(float A[2][100]) {
@ -31,6 +34,15 @@
; IR-NEXT: %polly.loop_cond = icmp sle i64 %polly.indvar, 98 ; IR-NEXT: %polly.loop_cond = icmp sle i64 %polly.indvar, 98
; IR-NEXT: br i1 %polly.loop_cond, label %polly.loop_header, label %polly.loop_exit ; IR-NEXT: br i1 %polly.loop_cond, label %polly.loop_header, label %polly.loop_exit
; KERNEL-IR: define ptx_kernel void @kernel_0(i8* %MemRef_A, i64 %c0) {
; KERNEL-IR-NEXT: entry:
; KERNEL-IR-NEXT: %0 = call i32 @llvm.nvvm.read.ptx.sreg.ctaid.x()
; KERNEL-IR-NEXT: %b0 = zext i32 %0 to i64
; KERNEL-IR-NEXT: %1 = call i32 @llvm.nvvm.read.ptx.sreg.tid.x()
; KERNEL-IR-NEXT: %t0 = zext i32 %1 to i64
; KERNEL-IR-NEXT: ret void
; KERNEL-IR-NEXT: }
target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128" target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
define void @foo([100 x float]* %A) { define void @foo([100 x float]* %A) {