ROCm/hcc là dự án mã nguồn mở trên GitHub, viết chủ yếu bằng C++, với 443 sao. HCC is an Open Source, Optimizing C++ Compiler for Heterogeneous Compute currently for the ROCm GPU Computing Platform
Tóm tắt dựng từ metadata GitHub của chính dự án — chưa có bài review TopGit. Trang sẽ tự động cập nhật khi bài review đầy đủ được xuất bản.
VÌ SAO CHƯA CÓ REVIEW
TopGit viết bài đầy đủ cho repo có nhiều sao nhất và được yêu cầu nhiều nhất. Trang này là snapshot trong thời gian chờ — xem README gốc ở tab READ ME.
HCC : An open source C++ compiler for heterogeneous devices
This repository hosts the HCC compiler implementation project. The goal is to
implement a compiler that takes a program that conforms to a parallel programming
standard such as HC, C++ 17 ParallelSTL and transforms it
into the AMD GCN ISA.
The project is based on LLVM+CLANG. For more information, please visit the
hcc wiki:
https://github.com/RadeonOpenCompute/hcc/wiki
Deprecation Notice
AMD is deprecating HCC to put more focus on HIP development and on other languages supporting heterogeneous compute. We will no longer develop any new feature in HCC and we will stop maintaining HCC after its final release, which is planned for June 2019. If your application was developed with the hc C++ API, we would encourage you to transition it to other languages supported by AMD, such as HIP or OpenCL. HIP and hc language share the same compiler technology, so many hc kernel language features (including inline assembly) are also available through the HIP compilation path.
Download HCC
The project now employs git submodules to manage external components it depends
upon. It it advised to add --recursive when you clone the project so all
submodules are fetched automatically.
For more information about git submodules, please refer to git documentation.
Build HCC from source
To configure and build HCC from source, use the following steps:
mkdir -p build; cd build
cmake -DCMAKE_BUILD_TYPE=Release ..
make
To install it, use the following steps:
sudo make install
Use HCC
For HC source codes:
hcc -hc foo.cpp -o foo
Multiple ISA
HCC now supports having multiple GCN ISAs in one executable file. You can do it
in different ways:
use --amdgpu-target= command line option
It's possible to specify multiple --amdgpu-target= option. Example:
# ISA for Fiji(gfx803) and Vega10(gfx900) would
# be produced
hcc -hc \
--amdgpu-target=gfx803 \
--amdgpu-target=gfx900 \
foo.cpp
configure HCC use CMake HSA_AMDGPU_GPU_TARGET variable
If you build HCC from source, it's possible to configure it to automatically
produce multiple ISAs via HSA_AMDGPU_GPU_TARGET CMake variable.
Use ; to delimit each AMDGPU target. Example:
# ISA for Fiji(gfx803) and Vega10(gfx900) would
# be produced by default
cmake \
-DCMAKE_BUILD_TYPE=Release \
-DHSA_AMDGPU_GPU_TARGET="gfx803;gfx900" \
../hcc
CodeXL Activity Logger
To enable the CodeXL Activity Logger, use the USE_CODEXL_ACTIVITY_LOGGER
environment variable.
In your application compiled using hcc, include the CodeXL Activity Logger header:
#include <CXLActivityLogger.h>
For information about the usage of the Activity Logger for profiling, please
refer to its documentation.
HCC with ThinLTO Linking
To enable the ThinLTO link time, use the KMTHINLTO environment variable.
Set up your environment in the following way:
export KMTHINLTO=1
ThinLTO Phase 1 - Implemented
For applications compiled using hcc, ThinLTO could significantly improve link-time
performance. This implementation will maintain kernels in their .bc file format,
create module-summaries for each, perform llvm-lto's cross-module function importing
and then perform clamp-device (which uses opt and llc tools) on each of the
kernel files. These files are linked with lld into one .hsaco per target specified.
ThinLTO Phase 2 - Under development
This ThinLTO implementation which will use llvm-lto LLVM tool to replace
clamp-device bash script. It adds an optllc option into ThinLTOGenerator,
which will perform in-program opt and codegen in parallel.
To use HCC Printf Functions
Set up environmental variable:
export HCC_ENABLE_PRINTF=1
Then compile the printf kernel with HCC_ENABLE_ACCELERATOR_PRINTF macro defined.
Trang TopGit này là một snapshot — tab "Readme" hiển thị nguyên văn README của repo (đã bỏ link, giữ ảnh). Repo GitHub ở github.com/ROCm/hcc là nguồn chính thức.
ROCm/hcc có phải mã nguồn mở không?
TopGit chưa ghi nhận license cho ROCm/hcc. Phần lớn repo public trên GitHub là mã nguồn mở, nhưng điều khoản khác nhau từng repo — mở file LICENSE để xác nhận.
ROCm/hcc là gì?
ROCm/hcc (ROCm/hcc) là dự án C++ trên GitHub. Theo mô tả gốc: HCC is an Open Source, Optimizing C++ Compiler for Heterogeneous Compute currently for the ROCm GPU Computing Platform
Đọc đầy đủ README ở tab phía trên.
hcc có đáng để bạn bỏ thời gian?
ChatGPT, Claude và Perplexity đều đọc được trang này. Hỏi thử xem họ nghĩ gì về hcc.