264ac2d3af
This PR adds support for HIP on macOS: Mach-O section naming, Darwin host toolchain initialization guards, and HIPSPV behavior when Darwin is the host. This has been verified using chipStar on MacOS via the PoCL OpenCL implementation. ## Uninitialized target workaround Darwin’s toolchain is only initialized when its own TranslateArgs runs. For HIP/CUDA device jobs, Darwin is used as the HostTC and never gets its args translated, so its target stays uninitialized. The new checks avoid asserting on that uninitialized state. A better long-term fix is to initialize Darwin earlier (see the FIXME in Driver.cpp BuildJobsForAction). - [ ] Initialize Darwin toolchain during construction instead of lazily in TranslateArgs. See Driver.cpp BuildJobsForAction FIXME. - [x] In Darwin’s addClangTargetOptions, skip host-stdlib flags when DeviceOffloadKind != OFK_None so HIPSPV can safely delegate to the host.
29 lines
1.2 KiB
Plaintext
29 lines
1.2 KiB
Plaintext
// Verify that HIP fat binary sections use Mach-O "segment,section" format on Darwin.
|
|
|
|
// RUN: echo -n "GPU binary would be here." > %t
|
|
// RUN: %clang_cc1 -triple x86_64-apple-macosx10.15.0 -emit-llvm %s \
|
|
// RUN: -fcuda-include-gpubinary %t -o - -x hip \
|
|
// RUN: | FileCheck %s --check-prefix=HIPEF
|
|
// RUN: %clang_cc1 -cuid=123 -triple x86_64-apple-macosx10.15.0 -emit-llvm %s \
|
|
// RUN: -o - -x hip \
|
|
// RUN: | FileCheck %s --check-prefix=HIPNEF
|
|
|
|
#include "Inputs/cuda.h"
|
|
|
|
__device__ int device_var;
|
|
__constant__ int constant_var;
|
|
|
|
// When fat binary is embedded, section names use Mach-O format.
|
|
// HIPEF: @[[FATBIN:.*]] = private constant{{.*}} c"GPU binary would be here.",{{.*}}section "__HIP,__hip_fatbin"{{.*}}align 4096
|
|
// HIPEF: @__hip_fatbin_wrapper = internal constant { i32, i32, ptr, ptr }
|
|
// HIPEF-SAME: section "__HIP,__fatbin"
|
|
|
|
// When fat binary is external (no -fcuda-include-gpubinary), external symbol uses Mach-O section.
|
|
// HIPNEF: @[[FATBIN:__hip_fatbin_[0-9a-f]+]] = external constant i8, section "__HIP,__hip_fatbin"
|
|
// HIPNEF: @__hip_fatbin_wrapper = internal constant { i32, i32, ptr, ptr }
|
|
// HIPNEF-SAME: section "__HIP,__fatbin"
|
|
|
|
__global__ void kernelfunc(int i, int j, int k) {}
|
|
|
|
void hostfunc(void) { kernelfunc<<<1, 1>>>(1, 1, 1); }
|