[Codegen][CUDA] Fix: cuda codegen vectorize cast #7561
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
kongroo
force-pushed
the
fix_cuda_vectorized_cast
branch
from
0ecc450
to
4a9aeec
Compare
Laurawly
reviewed
Mar 2, 2021
| } | ||
| if (!fail) { | ||
| return; | ||
| } |
| @@ -511,8 +511,8 @@ def check(t0, t1): | |||
|
|
|||
| # schedule | |||
| s = tvm.te.create_schedule(C.op) | |||
| ob, ib = s[C].split(s[C].op.axis[0], nparts=32) | |||
| _, iib = s[C].split(ib, factor=4) | |||
| ob, ib = s[C].split(s[C].op.axis[0], nparts=n // factor) | |||
There was a problem hiding this comment.
We can also directly say factor=factor here.
There was a problem hiding this comment.
like this?
ob, ib = s[C].split(s[C].op.axis[0], factor=factor)
# _, iib = s[C].split(ib, factor=factor)
s[C].vectorize(ib)
vinx13
approved these changes
Mar 2, 2021
trevor-m
pushed a commit
to trevor-m/tvm
that referenced
this pull request
May 6, 2021
* fix: cuda codegen vectorize cast * style: fix python coding style * fix: missing break * refactor: directly split by factor Co-authored-by: jiangchengquan <jiangchengquan@bytedance.com>
trevor-m
pushed a commit
to neo-ai/tvm
that referenced
this pull request
May 11, 2021
* fix: cuda codegen vectorize cast * style: fix python coding style * fix: missing break * refactor: directly split by factor Co-authored-by: jiangchengquan <jiangchengquan@bytedance.com>
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Outdated suggestions cannot be applied.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.
Data types such as float32x8 and int32x8 are not supported in CUDA, which will result in errors like "Cannot convert type float32x8 to CUDA type!" in code generation. I tried to fix this by storing 2 32-bits values in 1 64-bits value.
Could you please help review this fix? @Laurawly