compile3: gate FLOAT16/IMAGE/NOLOCALS defaults on QCOM#7
Merged
Conversation
|
This branch currently is behind tinygrad/master. The line count difference bot is disabled. |
Only set FLOAT16=1, IMAGE=2, NOLOCALS=1 defaults when DEV is not explicitly set or is set to QCOM. When callers pass a different DEV (e.g. CPU:LLVM), these defaults are skipped — models like dmonitoring_model don't need them and fail validation with FLOAT16=1 on CPU.
sunnyhaibin
force-pushed
the
fix/remove-compile3-defaults
branch
from
76ca5ae to
89e2b31
Compare
sunnyhaibin
changed the title
sunnyhaibin
changed the title
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
FLOAT16=1,IMAGE=2,NOLOCALS=1env var defaults behind aDEVcheckDEVis not set or is explicitlyQCOMJIT_BATCH_SIZE=0default remains unconditional (matches upstream)Why
When callers pass a non-QCOM device (e.g.
DEV=CPU:LLVM), these defaults shouldn't apply. Specifically,dmonitoring_modelon CPU:LLVM fails the output-sensitivity validation test whenFLOAT16=1is set — doubling inputs produces identical outputs.Behavior matrix
QCOM=1, no DEV)QCOMCPU:LLVMQCOMCPU:LLVM