JIT: Use successor edges instead of block targets for BBJ_EHFINALLYRET #98563
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
ghost
assigned 
dotnet-issue-labeler
bot
added
the
area-CodeGen-coreclr
|
Tagging subscribers to this area: @JulieLeeMSFT, @jakobbotsch Issue DetailsPart of #93020. This is a first step in a broader effort to replace
|
51 tasks
Diff results for #98563Throughput diffsThroughput diffs for linux/arm64 ran on windows/x64Overall (+0.02% to +0.03%)
MinOpts (+0.02% to +0.06%)
FullOpts (+0.02% to +0.03%)
Throughput diffs for linux/x64 ran on windows/x64Overall (+0.03% to +0.04%)
MinOpts (+0.02% to +0.09%)
FullOpts (+0.03% to +0.04%)
Throughput diffs for osx/arm64 ran on windows/x64Overall (+0.02% to +0.03%)
MinOpts (+0.02% to +0.07%)
FullOpts (+0.02% to +0.03%)
Throughput diffs for windows/arm64 ran on windows/x64Overall (+0.02% to +0.03%)
MinOpts (+0.02% to +0.08%)
FullOpts (+0.02% to +0.03%)
Throughput diffs for windows/x64 ran on windows/x64Overall (+0.03% to +0.04%)
MinOpts (+0.02% to +0.09%)
FullOpts (+0.03% to +0.04%)
Details here Throughput diffs for windows/x86 ran on linux/x86Overall (+0.02% to +0.03%)
MinOpts (+0.02% to +0.09%)
FullOpts (+0.02% to +0.03%)
Details here |
|
cc @dotnet/jit-contrib, @AndyAyersMS PTAL. No diffs, though there is some TP impact -- CI failure is #98500. |
Diff results for #98563Throughput diffsThroughput diffs for linux/arm ran on windows/x86Overall (+0.02% to +0.03%)
MinOpts (+0.02% to +0.06%)
FullOpts (+0.02% to +0.03%)
Details here Throughput diffs for linux/arm64 ran on linux/x64MinOpts (+0.00% to +0.01%)
Throughput diffs for linux/x64 ran on linux/x64MinOpts (+0.00% to +0.01%)
Details here |
AndyAyersMS
approved these changes
Feb 16, 2024
| bool found = false; | ||
|
|
||
| // Walk the successor table looking for the specified successor block. | ||
| // Search succTab for succEdge so we can splice it out of the table. |
There was a problem hiding this comment.
Do you know how frequently we remove a successor like this?
If it happens somewhat frequently, I wonder if we'd be better off allowing there to be voids in the array and just skipping over them.
There was a problem hiding this comment.
I'm not sure how often we encounter this, though I imagine this bit of logic isn't cheap TP-wise. There are quite a few places where we, for example, convert a BBJ_ALWAYS to a BBJ_COND (and vice versa), and we have to remove the current successor edge(s) from the successor block's pred list (and then add in new edges to the pred list -- it would be nice to see just how much #93636 helps with this). I can take a look at this locally and see how helpful it would be to skip this maintenance.
| // | ||
| BasicBlock* succBlock = edge->getDestinationBlock(); | ||
| assert(succBlock != nullptr); | ||
| succBlock->ensurePredListOrder(this); |
There was a problem hiding this comment.
Speaking of pred list order, we might want to revisit #93636 sometime.
|
CI failure is #98500. |
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
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.
Part of #93020. This is a first step in a broader effort to replace
BasicBlock's jump targets (bbTarget,bbFalseTarget, etc.) with successor flow edges. This PR also adds overloaded implementations for commonly used edge maintenance methods (fgReplacePred,fgRemoveRefPred, etc.) that can take advantage of the cheap access to successor edges (i.e. without callingfgGetPredForBlock-- hopefully we'll be able to get rid of that at some point).