Implement JAX pow2_decompose primitive.
#100
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
balancap
force-pushed
the
pow2-decompose-primitive
branch
2 times, most recently
from
669c5a1
to
19904d3
Compare
The primitive `pow2_decompose` is the core decomposition kernel used everywhere in AutoScale/Scalify, meaning it is worth properly formalizing it as a JAX primitive, simplifying the Jaxpr level graph and allowing proper custom kernel optimization on different HW platforms (GPU, IPU, TPU, ...). NOTE: this PR is fixing additional subnormal related bugs, due to inconsistency of jnp.frexp vs Numpy. See: jax-ml/jax#19689
balancap
force-pushed
the
pow2-decompose-primitive
branch
from
19904d3
to
61d62ab
Compare
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Outdated suggestions cannot be applied.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.
The primitive
pow2_decomposeis the core decomposition kernel used everywhere in AutoScale/Scalify, meaning it is worth properly formalizing it as a JAX primitive, simplifying the Jaxpr level graph and allowing proper custom kernel optimization on different HW platforms (GPU, IPU, TPU, ...).NOTE: this PR is fixing additional subnormal related bugs, due to inconsistency of
jnp.frexpvs Numpy.See: jax-ml/jax#19689