Currently HIP toolchain calls clang to emit bitcode then calls opt/llc for device compilation for the default -fno-gpu-rdc
case, which is unnecessary since clang is able to compile a single source file to ISA.
This patch fixes the HIP action builder and toolchain so that the default -fno-gpu-rdc can be done like a canonical
toolchain, i.e. one clang -cc1 invocation to compile source code to ISA.
This can avoid unnecessary processes to speed up the compilation, and avoid redundant LLVM passes which are
performed in clang -cc1 and opt.
Differential Revision: https://reviews.llvm.org/D81627
This patch renames -f{no-}cuda-rdc to -f{no-}gpu-rdc and keeps the original
options as aliases. When -fgpu-rdc is off,
clang will assume the device code in each translation unit does not call
external functions except those in the device library, therefore it is possible
to compile the device code in each translation unit to self-contained kernels
and embed them in the host object, so that the host object behaves like
usual host object which can be linked by lld.
The benefits of this feature is: 1. allow users to create static libraries which
can be linked by host linker; 2. amortized device code linking time.
This patch modifies HIP action builder to insert actions for linking device
code and generating HIP fatbin, and pass HIP fatbin to host backend action.
It extracts code for constructing command for generating HIP fatbin as
a function so that it can be reused by early finalization. It also modifies
codegen of HIP host constructor functions to embed the device fatbin
when it is available.
Differential Revision: https://reviews.llvm.org/D52377
llvm-svn: 343611
To support separate compile/link and linking across device IR in different source files,
a new HIP action builder is introduced. Basically it compiles/links host and device
code separately, and embed fat binary in host linking stage through linker script.
Differential Revision: https://reviews.llvm.org/D46476
llvm-svn: 333483
Summary:
This patch proposes a new class to generate and record action dependences related with offloading. The builder provides three main functionalities:
- Add device dependences to host actions.
- Add host dependence to device actions.
- Register device top-level actions.
The constructor of the builder detect the programming models that should be supported, and generates a specialized builder for each. If a new programming model is to be added in the future, only a new specialized builder has to be implemented.
When the specialized builder is generated, it produces programming-model-specific diagnostics.
A CUDA specialized builder is proposed in the patch that mostly consists of the partition of the current `buildCudaAction` by the three different functionalities.
Reviewers: tra, echristo, ABataev, jlebar, hfinkel
Subscribers: Hahnfeld, whchung, guansong, jlebar, mehdi_amini, andreybokhanko, tcramer, mkuron, cfe-commits, arpith-jacob, carlo.bertolli, caomhin
Differential Revision: https://reviews.llvm.org/D18172
llvm-svn: 282865
Yaxun (Sam) Liu