Fix fp16 mma SIGSEGV: bit pattern dereferenced as pointer in mma_impl - #142
Open
Elio-yang wants to merge 1 commit into
Open
Fix fp16 mma SIGSEGV: bit pattern dereferenced as pointer in mma_impl#142Elio-yang wants to merge 1 commit into
Elio-yang wants to merge 1 commit into
Conversation
mma_impl reinterpreted the extracted 16-bit fp16 value as a pointer and dereferenced it (instructions.cc:1952), crashing any PTX-mode fp16 mma/wmma with nonzero data (e.g. fp16 1.0 = 0x3C00 -> load from address 0x3C00). Take the address of the storage instead. Includes a minimal reproducer under reproduce/ (stock SM7_QV100 config + gdb backtrace). Co-authored-by: Lars Christensen <1lbchristensen1@gmail.com>
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
mma_impl reinterpreted the extracted 16-bit fp16 value as a pointer and dereferenced it (instructions.cc:1952), crashing any PTX-mode fp16 mma/wmma with nonzero data (e.g. fp16 1.0 = 0x3C00 -> load from address 0x3C00). Take the address of the storage instead. Includes a minimal reproduce folder/ (SM7_QV100 config + gdb backtrace).