Skip to content

added support for rocm gpu autodetect #549

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 4 commits into from
Oct 13, 2017
Merged

added support for rocm gpu autodetect #549

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
c9da17b
added support for rocm gpu autodetect
Oct 13, 2017
32cc82e
changed type casting from old style to static_cast
Oct 13, 2017
410d0b1
fixed code to generate gfx specific code object
Oct 13, 2017
d9ccda8
fixed namespaces
Oct 13, 2017
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion HalideIR
Open in desktop
Submodule HalideIR updated 3 files
+3 −2 Makefile
+1 −5 src/ir/IRPrinter.cpp
+2 −2 src/tvm/node.h
16 changes: 14 additions & 2 deletions src/codegen/llvm/codegen_amdgpu.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -135,9 +135,21 @@ runtime::Module BuildAMDGPU(Array<LoweredFunc> funcs, std::string target) {
CHECK(target.length(
) >= 4 &&
target.substr(0, 4) == "rocm");

TVMContext tvmCtx;
tvmCtx.device_type = kROCM;
tvmCtx.device_id = 0;
TVMRetValue val;
tvm::runtime::DeviceAPI::Get(tvmCtx)->GetAttr(tvmCtx, tvm::runtime::kExist, &val);
if (val.operator int() == 1) {
tvm::runtime::DeviceAPI::Get(tvmCtx)->GetAttr(tvmCtx, tvm::runtime::kComputeVersion, &val);
} else {
val = 803;
}

llvm::TargetMachine* tm = \
GetLLVMTargetMachine("-mtriple=amdgcn-amd-amdhsa-hcc -mcpu=gfx803" + \
target.substr(4, target.length() - 4));
GetLLVMTargetMachine("-mtriple=amdgcn-amd-amdhsa-hcc -mcpu=gfx" + \
std::to_string(val.operator int())+ target.substr(4, target.length() - 4));

std::unique_ptr<CodeGenAMDGPU> cg(new CodeGenAMDGPU());
std::unique_ptr<llvm::LLVMContext> ctx(new llvm::LLVMContext());
Expand Down
6 changes: 5 additions & 1 deletion src/runtime/rocm/rocm_device_api.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -44,7 +44,11 @@ class ROCMDeviceAPI final : public DeviceAPI {
value = 64;
break;
}
case kComputeVersion: return;
case kComputeVersion:
hipDeviceProp_t prop;
ROCM_CALL(hipGetDeviceProperties(&prop, ctx.device_id));
*rv = prop.gcnArch;
return;
}
*rv = value;
}
Expand Down