WIP: [SYCL] Basic SYCL device compiler support #3071
Closed
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
bader
added
the
upstream
Contributor
Author
While I was looking at this item I found a couple of other PRs, which are intended to improve this patch:
In addition to these two patches, we need another patch to emit SPIR metadata, calling convention, etc., which are required for successful translation to SPIR-V format. |
Summary:
All SYCL memory objects shared between host and device (buffers/images, these
objects map to OpenCL buffers and images) must be accessed through special
accessor classes. The "device" side implementation of these classes contain
pointers to the device memory. As there is no way in OpenCL to pass
structures with pointers inside as kernel arguments, all memory objects
shared between host and device must be passed to the kernel as raw
pointers. SYCL also has a special mechanism for passing kernel arguments
from host to the device. In OpenCL kernel arguments are set by calling
`clSetKernelArg` function for each kernel argument, meanwhile in SYCL all the
kernel arguments are fields of "SYCL kernel function" which can be defined
as a lambda function or a named function object and passed as an argument
to SYCL function for invoking kernels (such as `parallel_for` or `single_task`).
To facilitate the mapping of SYCL kernel data members to OpenCL kernel
arguments and overcome OpenCL limitations we added the generation of an
OpenCL kernel function inside the compiler. An OpenCL kernel function
contains the body of the SYCL kernel function, receives OpenCL-like
parameters and additionally does some manipulation to initialize SYCL
kernel data members with these parameters. In some pseudo code the OpenCL
kernel function can look like this:
```
// SYCL kernel is defined in SYCL headers:
template <typename KernelName, typename KernelType/*, ...*/>
__attribute__((sycl_kernel)) void sycl_kernel_function(KernelType KernelFuncObj) {
// ...
KernelFuncObj();
}
// Generated OpenCL kernel function
__kernel KernelName(global int* a) {
KernelType KernelFuncObj; // Actually kernel function object declaration
// doesn't have a name in AST.
// Let the kernel function object have one captured field - accessor A.
// We need to init it with global pointer from arguments:
KernelFuncObj.A.__init(a);
// Body of the SYCL kernel from SYCL headers:
{
KernelFuncObj();
}
}
```
OpenCL kernel function is generated by the compiler inside the Sema
using AST nodes.
Reviewers: bader, Naghasan, ABataev, keryell
Subscribers: agozillon, mgorny, yaxunl, jfb, ebevhan, Anastasia, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D71016
The CodeGen patch is outdated and it doesn't work with SYCL 2020 syntax used in this test.
This patch implements a few of the OpenCL types for SYCL, however doesn't bother handling semantic analysis as these are not intended to be used anywhere but in the SYCL implementation. Signed-off-by: Erich Keane <[email protected]> Differential Revision: https://reviews.llvm.org/D77220
* Make the OpenCL Builtin lookup agnostic to programming models Rename the OpenCL Builtin lookup system to make it agnostic. The Builtin emitter is now outputting the builtin information into a class so that different programming models can be produced and used. * Add SPIR-V variants of TypeSampledImage as clang builtin type. This patch adds SPIR-V sampled image types as derivative of the builtin OpenCL Image types. For each OpenCL image type, clang defines a Sampled<image type> variant and lowered as a "spirv.SampledImage.<image>" llvm opaque type. * Enable SPIR-V builtin lookup Add flag -fdeclare-spirv-builtins to enable lookup of SPIR-V builtins. If -fdeclare-spirv-builtins is passed to clang, the compiler will try to lookup for the builtin described in SPIRVBuiltins.td for any match. If a match is found, overloads are build for the match. This will enforce a stable way to express SPIR-V builtins and make them closer to how the translator mangles them. This will help ensuring builtin for CUDA does not break easily. This will also support any changes suggested by the SPIRV-LLVM people on how to represent builtins. Define __SPIRV_BUILTIN_DECLARATIONS__ when passing -fdeclare-spirv-builtins to clang. Added OpenCL SPIR-V extended set builtins bindings and part of the core SPIR-V (mostly missing Images and Pipes) TODO: Known vendor extensions are not implemented yet. Signed-off-by: Victor Lomuller <[email protected]> Differential Revision: https://reviews.llvm.org/D108034
bader
force-pushed
the
sycl-device-compiler
branch
from
379642d to
e90c927
Compare
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This PR includes a set of commits from required to reach milestone 1 for "SYCL support in upstream clang" project, which implies successful compilation of basic SYCL application like simple-vector-add .
Here is a list of patch with preliminary TODOs left to finalize the solution.