Optimize LLVM JIT for large circuits #1261
Merged
Conversation
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
This makes a big difference for large UCCSD cicruits when using llvm::CodeGenOpt::None
|
CUDA Quantum Docs Bot: A preview of the documentation can be found here. |
|
CUDA Quantum Docs Bot: A preview of the documentation can be found here. |
|
CUDA Quantum Docs Bot: A preview of the documentation can be found here. |
|
CUDA Quantum Docs Bot: A preview of the documentation can be found here. |
bmhowe23
changed the title
1tnguyen
reviewed
Feb 22, 2024
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
3 participants
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Outdated suggestions cannot be applied.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.
This makes a big difference for large UCCSD circuits when using
llvm::CodeGenOpt::None, which we do in multiple places.For example, the JIT time for one of @marwafar's benchmark circuits went from ~260 seconds to ~45 seconds. (Most of the remaining time is spent in our Canonicalizer pass.)
Additional info: prior to this PR, we were spending a large amount of time in https://github.com/llvm/llvm-project/blob/llvmorg-16.0.6/llvm/lib/CodeGen/SelectionDAG/SelectionDAGISel.cpp#L1581 for large circuits. This change remedies that by disabling
-fast-isel. The "fast" instruction selection was supposed to be faster than the alternatives, but for us, it is actually slower for large circuits where it matters.Also note that I believe https://github.com/llvm/llvm-project/blob/llvmorg-16.0.6/llvm/lib/CodeGen/TargetPassConfig.cpp#L997 is preventing us from doing this in cleaner way where we specify this option via
TargetMachine::setFastISel(false)andTargetMachine::setO0WantsFastISel(false).