Page MenuHomePhabricator

wmi (Wei Mi)
User

Projects

User does not belong to any projects.

User Details

User Since
Feb 20 2015, 10:57 AM (238 w, 6 d)

Recent Activity

Yesterday

wmi updated the diff for D67726: [SampleFDO] Expose an interface to return the size of a section or the size of the profile for profile in ExtBinary format.

Address David's comment.

Thu, Sep 19, 4:11 PM · Restricted Project
wmi added a comment to D67726: [SampleFDO] Expose an interface to return the size of a section or the size of the profile for profile in ExtBinary format.

Sounds a tricky situation to handle. Suppose the profile size increases some limit, what criterial should be used to trim the size? Trim the symbol list or drop more samples?

Thu, Sep 19, 11:51 AM · Restricted Project
wmi added a comment to D67726: [SampleFDO] Expose an interface to return the size of a section or the size of the profile for profile in ExtBinary format.

Is the intended workflow to create a profile with full symbol list and then have an option to trim symbol list according to size impact? Why not do the trimming at the time of the profile creation?

Thu, Sep 19, 11:16 AM · Restricted Project

Wed, Sep 18

wmi updated the diff for D67726: [SampleFDO] Expose an interface to return the size of a section or the size of the profile for profile in ExtBinary format.

Add test to verify the size of the profile matching the size shown by llvm-profdata and add an assertion to verify the size of individual section and header adds up to total file size.

Wed, Sep 18, 4:10 PM · Restricted Project
wmi updated subscribers of D67726: [SampleFDO] Expose an interface to return the size of a section or the size of the profile for profile in ExtBinary format.
Wed, Sep 18, 1:08 PM · Restricted Project
wmi created D67726: [SampleFDO] Expose an interface to return the size of a section or the size of the profile for profile in ExtBinary format.
Wed, Sep 18, 12:17 PM · Restricted Project
wmi committed rG5f7e822dc724: [SampleFDO] Minimize performance impact when profile-sample-accurate is enabled. (authored by wmi).
[SampleFDO] Minimize performance impact when profile-sample-accurate is enabled.
Wed, Sep 18, 9:07 AM
wmi committed rL372232: [SampleFDO] Minimize performance impact when profile-sample-accurate.
[SampleFDO] Minimize performance impact when profile-sample-accurate
Wed, Sep 18, 9:06 AM
wmi closed D67561: [SampleFDO] minimize performance impact when profile-sample-accurate is enabled.
Wed, Sep 18, 9:06 AM · Restricted Project

Tue, Sep 17

wmi updated the diff for D67561: [SampleFDO] minimize performance impact when profile-sample-accurate is enabled.

Address David's comment.

Tue, Sep 17, 11:05 AM · Restricted Project
wmi added inline comments to D67561: [SampleFDO] minimize performance impact when profile-sample-accurate is enabled.
Tue, Sep 17, 11:02 AM · Restricted Project
wmi added inline comments to D67561: [SampleFDO] minimize performance impact when profile-sample-accurate is enabled.
Tue, Sep 17, 10:20 AM · Restricted Project

Mon, Sep 16

wmi added inline comments to D67561: [SampleFDO] minimize performance impact when profile-sample-accurate is enabled.
Mon, Sep 16, 4:04 PM · Restricted Project
wmi updated the diff for D67561: [SampleFDO] minimize performance impact when profile-sample-accurate is enabled.

Create a StringSet for NameTable to avoid linear scanning of the table for each function.

Mon, Sep 16, 4:01 PM · Restricted Project

Sun, Sep 15

wmi added a comment to D67561: [SampleFDO] minimize performance impact when profile-sample-accurate is enabled.

If we use NameTable, is ProfileSymbolList still needed? I assume NameTable is a superset ?

Sun, Sep 15, 9:54 PM · Restricted Project

Fri, Sep 13

wmi created D67561: [SampleFDO] minimize performance impact when profile-sample-accurate is enabled.
Fri, Sep 13, 9:16 AM · Restricted Project

Wed, Sep 11

wmi added a comment to D67359: [IVUser] Limit the iterations to check whether a loop has dedicated exits for extreme large case.

Is there any chance you could provide the test case that shows the behavior? There might be something we could do in hasDedicatedExits to make it faster for large loops with a large number of exit blocks and successors.

Wed, Sep 11, 2:58 PM · Restricted Project

Tue, Sep 10

wmi added a comment to D67359: [IVUser] Limit the iterations to check whether a loop has dedicated exits for extreme large case.
In D67359#1664242, @wmi wrote:

I wonder whether the problem is in this function or in analysis itself?
If it is a problem of analysis it is better to restrict it but not an utility function.

I feel the problem is in the utility function. Currently the problem is exposed in IVUser pass but the same can happen anywhere else if only the function is called for a huge loop with many exits and with each exit having many predecessors.

Can you do some measurements to ensure that the issue is really in this utility function but not in the user of it?

Tue, Sep 10, 10:27 AM · Restricted Project

Mon, Sep 9

wmi added a comment to D67359: [IVUser] Limit the iterations to check whether a loop has dedicated exits for extreme large case.

I wonder whether the problem is in this function or in analysis itself?
If it is a problem of analysis it is better to restrict it but not an utility function.

Mon, Sep 9, 9:20 PM · Restricted Project
wmi created D67359: [IVUser] Limit the iterations to check whether a loop has dedicated exits for extreme large case.
Mon, Sep 9, 9:39 AM · Restricted Project

Fri, Aug 30

wmi committed rG198009ae8db1: Fix some errors introduced by rL370563 which were not exposed on my local… (authored by wmi).
Fix some errors introduced by rL370563 which were not exposed on my local…
Fri, Aug 30, 8:17 PM
wmi committed rL370564: Fix some errors introduced by rL370563 which were not exposed on my local….
Fix some errors introduced by rL370563 which were not exposed on my local…
Fri, Aug 30, 8:17 PM
wmi committed rG798e59b81fb7: [SampleFDO] Add profile symbol list section to discriminate function being cold… (authored by wmi).
[SampleFDO] Add profile symbol list section to discriminate function being cold…
Fri, Aug 30, 7:30 PM
wmi committed rL370563: [SampleFDO] Add profile symbol list section to discriminate function being.
[SampleFDO] Add profile symbol list section to discriminate function being
Fri, Aug 30, 7:26 PM
wmi closed D66766: [SampleFDO] Add symbol whitelist section to discriminate function being cold versus function being newly added.
Fri, Aug 30, 7:26 PM · Restricted Project
wmi committed rG5ef5829fb02a: [GVN] Verify value equality before doing phi translation for call instruction (authored by wmi).
[GVN] Verify value equality before doing phi translation for call instruction
Fri, Aug 30, 4:04 PM
wmi committed rL370547: [GVN] Verify value equality before doing phi translation for call instruction.
[GVN] Verify value equality before doing phi translation for call instruction
Fri, Aug 30, 4:04 PM
wmi closed D67013: [GVN] Verify value equality before doing phi translation for call instruction.
Fri, Aug 30, 4:04 PM · Restricted Project
wmi added a comment to D66766: [SampleFDO] Add symbol whitelist section to discriminate function being cold versus function being newly added.

lgtm (please consolidate two test cases by factoring the common source code out and have two wrapper tests referencing it).

Fri, Aug 30, 3:43 PM · Restricted Project
wmi updated the diff for D66766: [SampleFDO] Add symbol whitelist section to discriminate function being cold versus function being newly added.

Extract the common part of compressed-profile-symbol-list.ll and uncompressed-profile-symbol-list.ll to a separate .ll file under Inputs directory.

Fri, Aug 30, 3:42 PM · Restricted Project
wmi added a comment to D66766: [SampleFDO] Add symbol whitelist section to discriminate function being cold versus function being newly added.

Suggesting on the naming: instead of using SymbolWhiteList, use 'ProfileSymbolList', or just 'SymbolList'. Otherwise the patch looks ok

Fri, Aug 30, 2:19 PM · Restricted Project
wmi updated the diff for D66766: [SampleFDO] Add symbol whitelist section to discriminate function being cold versus function being newly added.

Address David's comment.

Fri, Aug 30, 2:16 PM · Restricted Project
wmi added a comment to D66909: [GVN] Disallow phi translation for expressions with calls..

Thanks Wei, I think the patch you sent looks like a good path forward. I'd have a few comments on it.
Do you want to send it for review and pick up the test from this patch? I'll be happy to review it.

Fri, Aug 30, 10:52 AM · Restricted Project
wmi updated the diff for D67013: [GVN] Verify value equality before doing phi translation for call instruction.

Add the missing testcase.

Fri, Aug 30, 10:43 AM · Restricted Project
wmi created D67013: [GVN] Verify value equality before doing phi translation for call instruction.
Fri, Aug 30, 10:43 AM · Restricted Project

Thu, Aug 29

wmi added a comment to D66909: [GVN] Disallow phi translation for expressions with calls..
In D66909#1651640, @wmi wrote:

Thanks Alina to analyze the problem and write the fix.

I agree with your analysis and I believe the patch could fix the problem. But I am little worried the blunt hammer fix could pontentially hurt performance. Maybe we can use MD to do a finegrain check? I draft a fix attached. Please take a look whether it is correct and free to change it. {F9879631}

Thu, Aug 29, 1:50 PM · Restricted Project
wmi added a comment to D66909: [GVN] Disallow phi translation for expressions with calls..

Thanks Alina to analyze the problem and write the fix.

Thu, Aug 29, 1:46 PM · Restricted Project

Wed, Aug 28

wmi updated the diff for D66766: [SampleFDO] Add symbol whitelist section to discriminate function being cold versus function being newly added.

Address David and Mandeep's comments.

Wed, Aug 28, 5:11 PM · Restricted Project

Mon, Aug 26

wmi added a comment to D66374: [SampleFDO] Add symbol whitelist in the profile and use it when profile-sample-accurate is enabled.
In D66374#1638720, @wmi wrote:

Split the patch of format extension into https://reviews.llvm.org/D66513. The symbol whitelist patch is on top of D66513 and will be sent out later.

Mon, Aug 26, 12:51 PM · Restricted Project
wmi created D66766: [SampleFDO] Add symbol whitelist section to discriminate function being cold versus function being newly added.
Mon, Aug 26, 12:27 PM · Restricted Project
wmi committed rL369919: [SampleFDO] Extract the code calling each section reader to readOneSection..
[SampleFDO] Extract the code calling each section reader to readOneSection.
Mon, Aug 26, 9:03 AM
wmi committed rG077a9c7053dc: [SampleFDO] Extract the code calling each section reader to readOneSection. (authored by wmi).
[SampleFDO] Extract the code calling each section reader to readOneSection.
Mon, Aug 26, 8:55 AM
wmi closed D66693: [SampleFDO] Extract the code calling each section reader to addOneSection.
Mon, Aug 26, 8:55 AM · Restricted Project

Sun, Aug 25

wmi added a comment to D36562: [Bitfield] Make the bitfield a separate location if it has width of legal integer type and its bit offset is naturally aligned for the type.
In D36562#1641930, @wmi wrote:

Shouldn't we disable OPT_ffine_grained_bitfield_accesses only if TSAN is active?

I don't remember why it is disabled for all sanitizer modes. Seems you are right that the disabling the option is only necessary for TSAN. Do you have actual needs for the option to be functioning on other sanitizer modes?

Well, yes and no. We have the option enabled by default and it causes a warning when we use it together with -fsanitize=memtag (we aren't really concerned with other sanitizers). That warning broke a few builds (e.g. CMake doing tests and not wanting to see *any* diagnostics. We can work around that in a number of ways, e.g. we can leave the default off for AArch64.

I'd prefer though to have an upstream solution, if that's considered beneficial for all LLVM users and this one seems like such a case: let anyone use the option with sanitizers, unless it's known that some sanitizers'utility is affected negatively (as with TSAN).

Sun, Aug 25, 10:40 PM · Restricted Project
wmi created D66726: [Bitfield] Disable -ffine-grained-bitfield-accesses only when thread sanitizer is enabled.
Sun, Aug 25, 10:36 PM · Restricted Project
wmi updated subscribers of D66726: [Bitfield] Disable -ffine-grained-bitfield-accesses only when thread sanitizer is enabled.
Sun, Aug 25, 10:36 PM · Restricted Project

Fri, Aug 23

wmi created D66693: [SampleFDO] Extract the code calling each section reader to addOneSection.
Fri, Aug 23, 5:11 PM · Restricted Project
wmi committed rGf491d8639387: Fix some warnings introduced by r369798. (authored by wmi).
Fix some warnings introduced by r369798.
Fri, Aug 23, 12:39 PM
wmi committed rL369799: Fix some warnings introduced by r369798..
Fix some warnings introduced by r369798.
Fri, Aug 23, 12:38 PM
wmi committed rGbe9073249e34: [SampleFDO] Add ExtBinary format to support extension of binary profile. (authored by wmi).
[SampleFDO] Add ExtBinary format to support extension of binary profile.
Fri, Aug 23, 12:06 PM
wmi committed rL369798: [SampleFDO] Add ExtBinary format to support extension of binary profile..
[SampleFDO] Add ExtBinary format to support extension of binary profile.
Fri, Aug 23, 12:04 PM
wmi closed D66513: [SampleFDO] Add ExtBinary format to support extension of binary profile.
Fri, Aug 23, 12:04 PM · Restricted Project

Thu, Aug 22

wmi updated the diff for D66513: [SampleFDO] Add ExtBinary format to support extension of binary profile.

Address David's comment. Rename some section enums and reserve some enum values.

Thu, Aug 22, 4:40 PM · Restricted Project
wmi added a comment to D36562: [Bitfield] Make the bitfield a separate location if it has width of legal integer type and its bit offset is naturally aligned for the type.

Shouldn't we disable OPT_ffine_grained_bitfield_accesses only if TSAN is active?

Thu, Aug 22, 3:15 PM · Restricted Project
wmi updated the diff for D66513: [SampleFDO] Add ExtBinary format to support extension of binary profile.

Address David's comment.
Add ExtBinaryBase class for the basics of extensible format and ExtBinary inherits from it, so we can add other types of profiles in extensible format separately.
Some refactoring and comment changes.

Thu, Aug 22, 11:25 AM · Restricted Project

Wed, Aug 21

wmi added a comment to D66513: [SampleFDO] Add ExtBinary format to support extension of binary profile.

llvm-profdata should also have a round trip test:

text format --> ext-format-> text format --> compare text format before and after

text-format-> binary-format->ext-format-> text format --> compare text format.

Wed, Aug 21, 11:38 AM · Restricted Project

Aug 20 2019

wmi added a comment to D66374: [SampleFDO] Add symbol whitelist in the profile and use it when profile-sample-accurate is enabled.

Split the patch of format extension into https://reviews.llvm.org/D66513. The symbol whitelist patch is on top of D66513 and will be sent out later.

Aug 20 2019, 9:14 PM · Restricted Project
wmi created D66513: [SampleFDO] Add ExtBinary format to support extension of binary profile.
Aug 20 2019, 9:06 PM · Restricted Project
wmi accepted D66191: [AutoFDO] Make call targets order deterministic for sample profil.

LGTM. Thanks!

Aug 20 2019, 11:28 AM · Restricted Project

Aug 19 2019

wmi added a comment to D66374: [SampleFDO] Add symbol whitelist in the profile and use it when profile-sample-accurate is enabled.

Can you split the format extension change into a separate patch?

Aug 19 2019, 8:45 AM · Restricted Project

Aug 16 2019

wmi created D66374: [SampleFDO] Add symbol whitelist in the profile and use it when profile-sample-accurate is enabled.
Aug 16 2019, 4:56 PM · Restricted Project
wmi added inline comments to D66191: [AutoFDO] Make call targets order deterministic for sample profil.
Aug 16 2019, 4:34 PM · Restricted Project

Aug 15 2019

wmi added a comment to D66191: [AutoFDO] Make call targets order deterministic for sample profil.

Thanks for working on it.

Aug 15 2019, 10:39 AM · Restricted Project

Aug 13 2019

wmi added a comment to D66013: [X86] Emit correct unwind info for split-stack prologue.

I see. You already take the adjustment in insertCFIInstrs into consideration and assume the CFIs in the layout order will match CFIs in control flow in the end. In this way, .cfi_remember_state at the beginning of current BB is always assumed to remember the outgoing CFI information of its predecessor BB.

Aug 13 2019, 3:29 PM · Restricted Project
wmi added inline comments to D66013: [X86] Emit correct unwind info for split-stack prologue.
Aug 13 2019, 12:36 PM · Restricted Project
wmi added a comment to D66013: [X86] Emit correct unwind info for split-stack prologue.

I suspect there is some problem caused by interaction of .cfi_remember_state/.cfi_restore_state and CFIInstrInserter pass.

Aug 13 2019, 11:53 AM · Restricted Project

Aug 12 2019

wmi added a reviewer for D66013: [X86] Emit correct unwind info for split-stack prologue: djokov.
Aug 12 2019, 10:07 AM · Restricted Project
wmi added a comment to D32201: [RALLOC] Increase CSR cost in RegAllocGreedy to favour splitting over CSR first use.

Should this good patch be rebased and continued to move on? @wmi
it'd be better to move on this work because D27366 is abandoned and D34608 is not updated for a long time, and those 3 patches are all trying to fix same issue and not landed successfully.
It would be great If it can land.

Aug 12 2019, 9:34 AM · Restricted Project
wmi accepted D65848: [ThinLTO][AutoFDO] Fix memory corruption due to race condition from thin backends.

LGTM.

Aug 12 2019, 8:40 AM · Restricted Project

Aug 8 2019

wmi added a comment to D65848: [ThinLTO][AutoFDO] Fix memory corruption due to race condition from thin backends.

Thanks for the change.

Aug 8 2019, 12:28 PM · Restricted Project
wmi added a comment to D65848: [ThinLTO][AutoFDO] Fix memory corruption due to race condition from thin backends.

Thanks for working on the issue.

Aug 8 2019, 8:44 AM · Restricted Project

Jul 31 2019

wmi committed rGf49c107f06c6: [DAGCombine] Limit the number of times for the same store and root nodes to… (authored by wmi).
[DAGCombine] Limit the number of times for the same store and root nodes to…
Jul 31 2019, 1:00 PM
wmi committed rL367472: [DAGCombine] Limit the number of times for the same store and root nodes.
[DAGCombine] Limit the number of times for the same store and root nodes
Jul 31 2019, 12:59 PM
wmi closed D65174: [DAGCombine] Limit the number of times for a store being considered for merging.
Jul 31 2019, 12:59 PM · Restricted Project
wmi added a comment to D65174: [DAGCombine] Limit the number of times for a store being considered for merging.

Nirav, thanks for all the suggestions and quick responses!

Jul 31 2019, 11:59 AM · Restricted Project
wmi updated the diff for D65174: [DAGCombine] Limit the number of times for a store being considered for merging.

Address Nirav's comment.

Jul 31 2019, 11:32 AM · Restricted Project
wmi added inline comments to D65174: [DAGCombine] Limit the number of times for a store being considered for merging.
Jul 31 2019, 11:32 AM · Restricted Project
wmi updated the diff for D65174: [DAGCombine] Limit the number of times for a store being considered for merging.

Address Nirav's comment.

Jul 31 2019, 10:40 AM · Restricted Project
wmi added a comment to D65174: [DAGCombine] Limit the number of times for a store being considered for merging.

Thanks for the suggestions. I will update patch immediately. Please see if it matches what is in your mind.

Jul 31 2019, 10:35 AM · Restricted Project

Jul 30 2019

wmi updated the diff for D65174: [DAGCombine] Limit the number of times for a store being considered for merging.

Address Nirav's comment.

Jul 30 2019, 8:43 PM · Restricted Project
wmi added a comment to D65174: [DAGCombine] Limit the number of times for a store being considered for merging.

Thanks Nirav. I update the patch. Please take a look.

Jul 30 2019, 8:40 PM · Restricted Project
wmi committed rG888efda2808c: [DAGCombiner] Add an option to control whether or not to enable store merging. (authored by wmi).
[DAGCombiner] Add an option to control whether or not to enable store merging.
Jul 30 2019, 4:16 PM
wmi committed rL367365: [DAGCombiner] Add an option to control whether or not to enable store merging..
[DAGCombiner] Add an option to control whether or not to enable store merging.
Jul 30 2019, 4:16 PM
wmi closed D65482: [DAGCombiner] Add an option to control whether or not to enable store merging.
Jul 30 2019, 4:15 PM · Restricted Project, Restricted Project
wmi created D65482: [DAGCombiner] Add an option to control whether or not to enable store merging.
Jul 30 2019, 3:38 PM · Restricted Project, Restricted Project
wmi added a comment to D65174: [DAGCombine] Limit the number of times for a store being considered for merging.

It may not be easy for us to prune the searching space without missing some merging opportunity. I feel it is easier to achieve that goal by controlling how many times a {RootNode, StoreNode} pair appear in candidate set repeatedly. With the patch, no bit change in the binaries were found for clang and some large application. So the current patch is a safe change.

True. But my concern is that as we prune the search space, we will introduce burrs in the which will be extremely hard to find and notice. I agree that its' not worth it to solve it the ideal way, but if a quick to write but safe pruning of the search space happens to work we should go with that. If not, it's likely it'll be too hard to resolve better than a heuristics. The scheme here is fine modulo the following concerns:

  1. If it's really an issue with the candidate dependence check, we should move the RootStoreCountMap update be modified in the dependency search, not the candidate check. This shouldn't dramatically change the run time.
  2. The RootNode is effectively a function of the candidate store. We really should only have a single RootNode so it'd be better as a map from SDNode* to (SDNode*, int).
Jul 30 2019, 9:45 AM · Restricted Project

Jul 29 2019

wmi added a comment to D65174: [DAGCombine] Limit the number of times for a store being considered for merging.

Ah. That's a tricky one. It sounds like this isn't a case of the candidate stores not being mergable, but rather us giving up before we completely verify that.

I don't suppose the values in the indirect value operands to the candidate stores have a relatively short dependence on a node not too far from the RootNode. Right now, we only prune at the root node (and components through the TokenFactors). but we could also just mark further predecessors of the RootNode as not needing to be searched.

Jul 29 2019, 2:47 PM · Restricted Project
wmi added a comment to D65174: [DAGCombine] Limit the number of times for a store being considered for merging.

Recently we run into a case with compile time problem (27 mins in O3 but a few seconds in O0). In that case, we have a huge block containing many stores. Each time a store is visited, dagcombiner will find a candidate set all chained by the same root node for the store, then try to merge the nodes in the candidate set into a wider store. This is a quadratic algorithm, because for the same candidate set, everytime a store inside of the set is visited, the same set will be considered for merging.

Could you share the test case? I looked into a similar issue and came up with some sort of caching for the lookup. I never shared the patch, because we addressed the issue with a limit. But it might be helpful for your case.

Jul 29 2019, 12:49 PM · Restricted Project
wmi added a comment to D65174: [DAGCombine] Limit the number of times for a store being considered for merging.

I check and the reason the candidates cannot be merged isn't because of the legal size. It is that checkMergeStoreCandidatesForDependencies always returns false so https://reviews.llvm.org/D60133 doesn't help.

Ah! Interesting. I've only seen that dependence triggered from inlined memcpy.

What's the shape of the DAG here? There must be some non-standard data dependence between stores, which is not being checked. Maybe we can change the candidate search to avoid such candidates.

Jul 29 2019, 12:45 PM · Restricted Project
wmi added a comment to D65174: [DAGCombine] Limit the number of times for a store being considered for merging.

Thanks for pointing me to D60133.

Jul 29 2019, 11:23 AM · Restricted Project

Jul 26 2019

wmi committed rG55a68a24003a: [JumpThreading] Stop searching predecessor when the current bb is in a… (authored by wmi).
[JumpThreading] Stop searching predecessor when the current bb is in a…
Jul 26 2019, 2:03 PM
wmi committed rL367154: [JumpThreading] Stop searching predecessor when the current bb is in a.
[JumpThreading] Stop searching predecessor when the current bb is in a
Jul 26 2019, 1:59 PM
wmi closed D65310: [JumpThreading] In updatePredecessorProfileMetadata, stop searching predecessor when the current bb is an unreachable single bb loop .
Jul 26 2019, 1:59 PM · Restricted Project
wmi updated the diff for D65310: [JumpThreading] In updatePredecessorProfileMetadata, stop searching predecessor when the current bb is an unreachable single bb loop .

Address Eli's comment.

Jul 26 2019, 12:15 PM · Restricted Project

Jul 25 2019

wmi updated the diff for D65310: [JumpThreading] In updatePredecessorProfileMetadata, stop searching predecessor when the current bb is an unreachable single bb loop .

Address Eli's comment.

Jul 25 2019, 7:18 PM · Restricted Project
wmi added a comment to D65310: [JumpThreading] In updatePredecessorProfileMetadata, stop searching predecessor when the current bb is an unreachable single bb loop .

Is there any reason to expect that unreachable loops have exactly one block?

Jul 25 2019, 7:18 PM · Restricted Project
wmi created D65310: [JumpThreading] In updatePredecessorProfileMetadata, stop searching predecessor when the current bb is an unreachable single bb loop .
Jul 25 2019, 5:03 PM · Restricted Project

Jul 24 2019

wmi accepted D65162: [llvm-profdata] Profile dump for compact binary format.

LGTM. Thanks!

Jul 24 2019, 1:52 PM · Restricted Project
wmi added a comment to D65162: [llvm-profdata] Profile dump for compact binary format.

Thanks for the patch.

Jul 24 2019, 9:25 AM · Restricted Project

Jul 23 2019

wmi created D65174: [DAGCombine] Limit the number of times for a store being considered for merging.
Jul 23 2019, 4:59 PM · Restricted Project