[Driver][SYCL] Add support for large device code via a linker script #6584
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
Add -fsycl-huge-device-code as a driver option. When enabled, the driver generates a linker script which directs the linker to place device code later in the binary, which makes it less likely to create a distance larger than that which a PC32 relocation can span. For example, if a 3GB __clang_offload_bundle__ section is placed betwen .text and .rodata, a PC32 relocation to .rodata in the .text is not possible. With this option enabled the device code will be placed after both rather than between them.
mdtoguchi
reviewed
Aug 16, 2022
mdtoguchi
reviewed
Aug 16, 2022
Contributor
Author
|
There is 1 failing LIT test in "SYCL / Linux / OCL x64 LLVM Test Suite" which appears unrelated. Others have mentioned that it should be fixed by intel/llvm-test-suite#1040. |
Contributor
Author
|
Hi @intel/dpcpp-doc-reviewers -- could you please take a look at the |
pvchupin
previously approved these changes
Aug 18, 2022
Contributor
Author
|
@intel/llvm-gatekeepers, I think this PR is ready in terms of reviews. However, it had one unrelated LIT failure in "SYCL / Linux / OCL x64 LLVM Test Suite" which should be fixed by intel/llvm-test-suite#1040. Can the PR be merged with testing in this state, or is it recommended that I re-run it? |
mdtoguchi
approved these changes
Aug 23, 2022
AlexeySachkov
approved these changes
Aug 24, 2022
Contributor
|
Previously approved by intel/dpcpp-doc-reviewers and only minor amendments since. Merging this. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
5 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Outdated suggestions cannot be applied.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.
Add -fsycl-link-huge-device-code as a driver option. When enabled, the driver
generates a linker script which directs the linker to place device code
later in the binary, which makes it less likely to create a distance
larger than that which a PC32 relocation can span.
For example, if a 3GB __clang_offload_bundle__ section is placed betwen
.text and .rodata, a PC32 relocation to .rodata in the .text is not
possible. With this option enabled the device code will be placed after
both rather than between them.