Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Add shuffle support to TVM #3633

Merged
merged 12 commits into from
Aug 1, 2019
Merged

Add shuffle support to TVM #3633

merged 12 commits into from
Aug 1, 2019

Conversation

were
Copy link
Contributor

@were were commented Jul 26, 2019

Actually, TVM's IR support is incomplete.
Shuffle's IRVisitor dispatch is missing.
It has no LLVM generator either.

@were
Copy link
Contributor Author

were commented Jul 26, 2019

@tqchen

@were
Copy link
Contributor Author

were commented Jul 26, 2019

@kazum

@were
Copy link
Contributor Author

were commented Jul 26, 2019

@ZihengJiang

@were were requested review from tqchen, ZihengJiang and kazum July 26, 2019 20:13
src/codegen/llvm/codegen_llvm.cc Outdated Show resolved Hide resolved
src/codegen/llvm/codegen_llvm.cc Outdated Show resolved Hide resolved
src/codegen/llvm/codegen_llvm.cc Outdated Show resolved Hide resolved
src/codegen/llvm/codegen_llvm.cc Show resolved Hide resolved
tests/python/unittest/test_codegen_llvm.py Show resolved Hide resolved
src/pass/ir_visitor.cc Outdated Show resolved Hide resolved
@were
Copy link
Contributor Author

were commented Jul 26, 2019

@zhiics I have resolved most of your concerns and answered your question.

@zhiics
Copy link
Member

zhiics commented Jul 28, 2019

@kazum @tqchen @ZihengJiang Can any of you take a look as well?

@tqchen
Copy link
Member

tqchen commented Jul 28, 2019

@were please also use pattern-match to add most of the support(concat and subsetting) to CodgenC(CUDA/OpenCL/Metal)

src/codegen/llvm/codegen_llvm.cc Outdated Show resolved Hide resolved
tests/python/unittest/test_codegen_llvm.py Show resolved Hide resolved
tests/python/unittest/test_codegen_cuda.py Outdated Show resolved Hide resolved
tests/python/unittest/test_codegen_llvm.py Outdated Show resolved Hide resolved
@were
Copy link
Contributor Author

were commented Jul 29, 2019

@tqchen I added CUDA backend.
@kazum I checked the existence and bound of the extracted IntImm and UIntImm.

@were
Copy link
Contributor Author

were commented Jul 30, 2019

@kazum Can you take a look again?
@tqchen Can we merge this for now? I am in a hurry to use Shuffle.

Copy link
Contributor

@kazum kazum left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I still wonder if the test code is correct, but it works correctly after vectorization and the shuffle support looks good to me. Approved.

tests/python/unittest/test_codegen_cuda.py Outdated Show resolved Hide resolved
@were were requested a review from junrushao July 31, 2019 03:39
@tqchen tqchen merged commit a279dd0 into apache:master Aug 1, 2019
@were were deleted the llvm-shuffle branch August 2, 2019 05:15
wweic pushed a commit to wweic/tvm that referenced this pull request Aug 9, 2019
wweic pushed a commit to neo-ai/tvm that referenced this pull request Sep 6, 2019
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

Successfully merging this pull request may close these issues.

4 participants