[CPU] Use static compilation for kernels. #29
Merged
Conversation
This file contains 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
Signed-off-by: Ilya Enkovich <[email protected]>
minjang
approved these changes
Jun 20, 2024
| @@ -48,6 +48,8 @@ def _build(name, src, srcdir, library_dirs, include_dirs, libraries): | |||
| # CPU backend uses C++ (driver.cpp). Some old version compilers need a specific C++17 flag. | |||
| if src.endswith(".cpp") or src.endswith(".cc"): | |||
| cc_cmd += ["-std=c++17", "-fopenmp"] | |||
| if src.endswith(".s"): | |||
| cc_cmd += ["-gdwarf-5"] | |||
There was a problem hiding this comment.
I think we can follow it up later on a different platform and compiler (e.g., Mac/clang), but this makes the problem much simpler: just compiling .s. Looks good.
| ret = llvm.translate_to_bc(src) | ||
| return ret | ||
| def make_asm(src, metadata, options): | ||
| return llvm.translate_to_host_asm(src, options.enable_fp_fusion) |
There was a problem hiding this comment.
Did you measure the compilation time? It tends to be very long in some cases. For sure, we will address it later.
There was a problem hiding this comment.
It should be the same as we had for JIT. It can be quite long when we produce huge basic blocks (like tens or even hundreds of thousands of operations) because some optimizations can be O(basic_block_size^2). We should generate better code for those cases and use smaller block sizes for the CPU.
minjang
pushed a commit
to minjang/triton-cpu
that referenced
this pull request
Jun 22, 2024
Signed-off-by: Ilya Enkovich <[email protected]>
minjang
pushed a commit
that referenced
this pull request
Jun 24, 2024
Signed-off-by: Ilya Enkovich <[email protected]>
Devjiu
pushed a commit
to Devjiu/triton-cpu
that referenced
this pull request
Aug 13, 2024
Signed-off-by: Ilya Enkovich <[email protected]>
int3
pushed a commit
that referenced
this pull request
Aug 29, 2024
Signed-off-by: Ilya Enkovich <[email protected]>
minjang
pushed a commit
that referenced
this pull request
Sep 22, 2024
Signed-off-by: Ilya Enkovich <[email protected]>
minjang
pushed a commit
that referenced
this pull request
Oct 22, 2024
Signed-off-by: Ilya Enkovich <[email protected]>
minjang
pushed a commit
that referenced
this pull request
Oct 24, 2024
Signed-off-by: Ilya Enkovich <[email protected]>
int3
pushed a commit
that referenced
this pull request
Dec 6, 2024
Signed-off-by: Ilya Enkovich <[email protected]>
ienkovich
added a commit
that referenced
this pull request
Dec 6, 2024
Signed-off-by: Ilya Enkovich <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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 patch moves from JIT compilation for kernels to static compiler usage. The modified compilation flow has several advantages:
cpu_utils.soThis should make #25 and a part of #18 redundant.