Page MenuHomePhabricator

wmi (Wei Mi)
User

Projects

User does not belong to any projects.

User Details

User Since
Feb 20 2015, 10:57 AM (308 w, 4 d)

Recent Activity

Today

wmi committed rG3729ee893948: Fix Wmissing-field-initializers warnings. (authored by wmi).
Fix Wmissing-field-initializers warnings.
Tue, Jan 19, 3:27 PM
wmi committed rG21b1ad0340a7: [SampleFDO] Add the support to split the function profiles with context into (authored by wmi).
[SampleFDO] Add the support to split the function profiles with context into
Tue, Jan 19, 3:16 PM
wmi closed D94435: [SampleFDO] Add the support to split the function profiles with context into separate sections..
Tue, Jan 19, 3:16 PM · Restricted Project
wmi requested review of D94981: [LiveDebugVariables] Add cache for SkipPHIsLabelsAndDebug to prevent iterating the same set of PHI/LABEL/Debug instructions repeatedly..
Tue, Jan 19, 9:58 AM · Restricted Project

Yesterday

wmi added inline comments to D94435: [SampleFDO] Add the support to split the function profiles with context into separate sections..
Mon, Jan 18, 10:44 PM · Restricted Project

Fri, Jan 15

wmi accepted D93264: [CSSPGO] Introducing distribution factor for pseudo probe..

LGTM. Thanks.

Fri, Jan 15, 9:53 AM · Restricted Project, Restricted Project
wmi added inline comments to D94435: [SampleFDO] Add the support to split the function profiles with context into separate sections..
Fri, Jan 15, 9:38 AM · Restricted Project

Thu, Jan 14

wmi added inline comments to D93264: [CSSPGO] Introducing distribution factor for pseudo probe..
Thu, Jan 14, 10:58 PM · Restricted Project, Restricted Project
wmi added inline comments to D92998: [CSSPGO][llvm-profgen] Pseudo probe based CS profile generation.
Thu, Jan 14, 10:04 PM · Restricted Project
wmi added inline comments to D93556: [CSSPGO][llvm-profgen] Compress recursive cycles in calling context.
Thu, Jan 14, 4:20 PM · Restricted Project
wmi accepted D94111: [CSSPGO][llvm-profgen] Merge and trim profile for cold context to reduce profile size.

LGTM. Thanks.

Thu, Jan 14, 3:31 PM · Restricted Project
wmi added inline comments to D93556: [CSSPGO][llvm-profgen] Compress recursive cycles in calling context.
Thu, Jan 14, 3:15 PM · Restricted Project
wmi added inline comments to D93556: [CSSPGO][llvm-profgen] Compress recursive cycles in calling context.
Thu, Jan 14, 3:10 PM · Restricted Project
wmi added inline comments to D94111: [CSSPGO][llvm-profgen] Merge and trim profile for cold context to reduce profile size.
Thu, Jan 14, 12:09 PM · Restricted Project
wmi added inline comments to D93264: [CSSPGO] Introducing distribution factor for pseudo probe..
Thu, Jan 14, 11:11 AM · Restricted Project, Restricted Project

Wed, Jan 13

wmi added inline comments to D93264: [CSSPGO] Introducing distribution factor for pseudo probe..
Wed, Jan 13, 10:40 PM · Restricted Project, Restricted Project
wmi added inline comments to D94435: [SampleFDO] Add the support to split the function profiles with context into separate sections..
Wed, Jan 13, 10:28 PM · Restricted Project
wmi updated the diff for D94435: [SampleFDO] Add the support to split the function profiles with context into separate sections..

Address Hongtao's comment.

Wed, Jan 13, 10:26 PM · Restricted Project
wmi added a comment to D94001: [CSSPGO] Call site prioritized inlining for sample PGO.
In D94001#2494594, @wmi wrote:

Sorry for the delay in review. It makes a lot of sense to have a priority based inliner for CSSPGO since its profile annotation doesn't rely on replaying the inlining. But I don't understand why we rely on BFS/DFS strategy to expose the hottest callsite for priority based inliner. In my mind, CSSPGO can know the hotness of each callsite with full context information. Can we just sort all the callsites based on CSSPGO profile then try to inline the hottest callsite under a specific context first in a top down manner? We may actually need to inline some parent callsites before we can inline the hottest callsite, but it is all driven by the target to inline the hottest callsite first. If we worry some callsite is too deep and we need to inline through a deep path before we can inline the specific callsite, we may add the depth into priority computation. What do you think?

Yeah, that's one step further than what I have in this patch. The key here is the prioritized work list, and BFS is just a natural by product of using the priority queue. I called out BFS to emphasize the more balanced inlining, but it's not super accurate because it is only a true BFS when priorities are all identical. What you suggested is essentially increasing the scope of call sites to prioritize - currently it's the immediate callees. I agreed that increasing the scope to multiple levels or the entire sub call tree may be beneficial. But I don't think it conflicts the approach in the patch, instead it can be added leveraging this implementation.

With context profiles organized in a trie, it's relatively easy to get the hottest call sites in sub call tree (assuming no stale profile). I think we can assign total order/priority for these hot call sites as well as the call sites in the middle leading to those hot ones (call sites that lead to hottest call site need to have their priority bumped up), then the priority queue will lead us to these hottest call sites first even if they're several levels away.. I can give it a try to increase the scope of call site for prioritization as a follow up.

Wed, Jan 13, 9:57 PM · Restricted Project
wmi updated the diff for D94435: [SampleFDO] Add the support to split the function profiles with context into separate sections..

Address Hongtao's comment.

Wed, Jan 13, 4:15 PM · Restricted Project
wmi committed rG86341247c4a2: [NFC] Rename ThinLTOPhase to ThinOrFullLTOPhase and move it from PassBuilder.h (authored by wmi).
[NFC] Rename ThinLTOPhase to ThinOrFullLTOPhase and move it from PassBuilder.h
Wed, Jan 13, 3:56 PM
wmi closed D94613: [NFC] Rename ThinLTOPhase to PhaseInAllLTO and move it from PassBuilder.h to Pass.h.
Wed, Jan 13, 3:56 PM · Restricted Project
wmi added a comment to D94613: [NFC] Rename ThinLTOPhase to PhaseInAllLTO and move it from PassBuilder.h to Pass.h.
In D94613#2496792, @hoy wrote:

Thanks for the refactoring!

BTW, looks like there is a use of ThinLTOPhase::None to be replaced in polly/lib/Support/RegisterPasses.cpp. Otherwise, LGTM.

Wed, Jan 13, 3:06 PM · Restricted Project
wmi added inline comments to D94613: [NFC] Rename ThinLTOPhase to PhaseInAllLTO and move it from PassBuilder.h to Pass.h.
Wed, Jan 13, 1:56 PM · Restricted Project
wmi updated the diff for D94613: [NFC] Rename ThinLTOPhase to PhaseInAllLTO and move it from PassBuilder.h to Pass.h.

Address Wenlei and David's comments.

Wed, Jan 13, 1:55 PM · Restricted Project
wmi added inline comments to D94435: [SampleFDO] Add the support to split the function profiles with context into separate sections..
Wed, Jan 13, 10:01 AM · Restricted Project
wmi requested review of D94613: [NFC] Rename ThinLTOPhase to PhaseInAllLTO and move it from PassBuilder.h to Pass.h.
Wed, Jan 13, 9:59 AM · Restricted Project
wmi added a comment to D94435: [SampleFDO] Add the support to split the function profiles with context into separate sections..

Thanks for the change.

If the profile is large, it is better to split the profile into two parts, one with context and one without, so the profile reading in postlink phase only has to read the part with context.

I guess the speed up is more visible with merged partial profile with explicit profile flattening? or did you see it speed up FDO with regular full profile too? For a regular FDO profile, I'd expect most function profiles to have some context except for really small functions..

Wed, Jan 13, 9:45 AM · Restricted Project

Tue, Jan 12

wmi added a comment to D94435: [SampleFDO] Add the support to split the function profiles with context into separate sections..

I guess there will be a separate patch to turn on this work, i.e, setting CtxSplitLayout?

Tue, Jan 12, 9:11 PM · Restricted Project
wmi accepted D92896: [CSSPGO][llvm-profgen] Virtual unwinding with pseudo probe.

LGTM.

Tue, Jan 12, 5:45 PM · Restricted Project
wmi added inline comments to D93556: [CSSPGO][llvm-profgen] Compress recursive cycles in calling context.
Tue, Jan 12, 4:43 PM · Restricted Project
wmi added inline comments to D92896: [CSSPGO][llvm-profgen] Virtual unwinding with pseudo probe.
Tue, Jan 12, 4:25 PM · Restricted Project
wmi accepted D92334: [CSSPGO][llvm-profgen] Pseudo probe decoding and disassembling.

LGTM.

Tue, Jan 12, 4:16 PM · Restricted Project
wmi added a comment to D94001: [CSSPGO] Call site prioritized inlining for sample PGO.

Sorry for the delay in review. It makes a lot of sense to have a priority based inliner for CSSPGO since its profile annotation doesn't rely on replaying the inlining. But I don't understand why we rely on BFS/DFS strategy to expose the hottest callsite for priority based inliner. In my mind, CSSPGO can know the hotness of each callsite with full context information. Can we just sort all the callsites based on CSSPGO profile then try to inline the hottest callsite under a specific context first in a top down manner? We may actually need to inline some parent callsites before we can inline the hottest callsite, but it is all driven by the target to inline the hottest callsite first. If we worry some callsite is too deep and we need to inline through a deep path before we can inline the specific callsite, we may add the depth into priority computation. What do you think?

Tue, Jan 12, 3:38 PM · Restricted Project
wmi added inline comments to D93264: [CSSPGO] Introducing distribution factor for pseudo probe..
Tue, Jan 12, 11:07 AM · Restricted Project, Restricted Project

Mon, Jan 11

wmi requested review of D94435: [SampleFDO] Add the support to split the function profiles with context into separate sections..
Mon, Jan 11, 11:06 AM · Restricted Project

Sun, Jan 10

wmi added inline comments to D94111: [CSSPGO][llvm-profgen] Merge and trim profile for cold context to reduce profile size.
Sun, Jan 10, 11:20 PM · Restricted Project
wmi added inline comments to D93556: [CSSPGO][llvm-profgen] Compress recursive cycles in calling context.
Sun, Jan 10, 9:29 PM · Restricted Project
wmi added a comment to D94110: [CSSPGO][llvm-profgen] Aggregate samples on call frame trie to speed up profile generation.

This change speeds up this by grouping all the call frame within one LBR sample into a trie and aggregating the result(sample counter) on it.

Sun, Jan 10, 9:28 PM · Restricted Project

Fri, Jan 8

wmi added inline comments to D92896: [CSSPGO][llvm-profgen] Virtual unwinding with pseudo probe.
Fri, Jan 8, 5:18 PM · Restricted Project
wmi added a comment to D92334: [CSSPGO][llvm-profgen] Pseudo probe decoding and disassembling.

Sorry for the delay in review. I will try to catch up.

Fri, Jan 8, 3:59 PM · Restricted Project

Thu, Jan 7

wmi accepted D92584: [CSSPGO][llvm-profgen] Refactor to unify hashable interface for trace sample and context-sensitive counter.

Sorry for the delay. LGTM.

Thu, Jan 7, 4:56 PM · Restricted Project

Dec 17 2020

wmi updated subscribers of D93254: [NFC][SampleFDO] Preparation to support multiple sections with the same type in ExtBinary format..

Thanks to Simon Pilgrim who helped to fix it at
https://reviews.llvm.org/rG4bb10be9a6e06a6c51cc1695ff5dc9d68c953334.

Dec 17 2020, 9:04 AM · Restricted Project

Dec 16 2020

wmi committed rGa906e3eccd1e: [NFC][SampleFDO] Preparation to support multiple sections with the same type (authored by wmi).
[NFC][SampleFDO] Preparation to support multiple sections with the same type
Dec 16 2020, 10:29 PM
wmi closed D93254: [NFC][SampleFDO] Preparation to support multiple sections with the same type in ExtBinary format..
Dec 16 2020, 10:29 PM · Restricted Project
wmi updated the diff for D93254: [NFC][SampleFDO] Preparation to support multiple sections with the same type in ExtBinary format..

Address Hongtao's comment.

Dec 16 2020, 5:16 PM · Restricted Project
wmi added inline comments to D93254: [NFC][SampleFDO] Preparation to support multiple sections with the same type in ExtBinary format..
Dec 16 2020, 5:14 PM · Restricted Project
wmi accepted D92347: [CSSPGO] Consume pseudo-probe-based AutoFDO profile.

LGTM.

Dec 16 2020, 12:16 PM · Restricted Project
wmi updated the diff for D93254: [NFC][SampleFDO] Preparation to support multiple sections with the same type in ExtBinary format..

Address Hongtao's comment.

Dec 16 2020, 12:15 PM · Restricted Project
wmi added inline comments to D93254: [NFC][SampleFDO] Preparation to support multiple sections with the same type in ExtBinary format..
Dec 16 2020, 12:05 PM · Restricted Project

Dec 15 2020

wmi added inline comments to D92347: [CSSPGO] Consume pseudo-probe-based AutoFDO profile.
Dec 15 2020, 8:31 PM · Restricted Project
wmi updated the diff for D93254: [NFC][SampleFDO] Preparation to support multiple sections with the same type in ExtBinary format..

The const indices passed to writeOneSection were incorrect (from a stale patch). Upload the latest version.

Dec 15 2020, 6:02 PM · Restricted Project
wmi added inline comments to D92347: [CSSPGO] Consume pseudo-probe-based AutoFDO profile.
Dec 15 2020, 5:45 PM · Restricted Project
wmi added inline comments to D92584: [CSSPGO][llvm-profgen] Refactor to unify hashable interface for trace sample and context-sensitive counter.
Dec 15 2020, 12:16 PM · Restricted Project

Dec 14 2020

wmi added a comment to D93254: [NFC][SampleFDO] Preparation to support multiple sections with the same type in ExtBinary format..

To reduce the profile loading time during postlink phase when ThinLTO is enabled, we are going to split profile into two parts. One part contains profiles with inline instance and another part contains flattened profiles without inline instance in them.

For the flattened profile, do you explicitly flatten part of raw profile, or it's just the portion of function profile that does not have any inlinee profile?

Dec 14 2020, 9:55 PM · Restricted Project
wmi added a comment to D93254: [NFC][SampleFDO] Preparation to support multiple sections with the same type in ExtBinary format..
In D93254#2453629, @hoy wrote:

Can you please shed light on how multiple same-typed sections will be used? Thanks.

Dec 14 2020, 8:41 PM · Restricted Project
wmi retitled D93254: [NFC][SampleFDO] Preparation to support multiple sections with the same type in ExtBinary format. from [NFC][SampleFDO] Support multiple sections with the same type in ExtBinary format. to [NFC][SampleFDO] Preparation to support multiple sections with the same type in ExtBinary format..
Dec 14 2020, 3:18 PM · Restricted Project
wmi requested review of D93254: [NFC][SampleFDO] Preparation to support multiple sections with the same type in ExtBinary format..
Dec 14 2020, 3:09 PM · Restricted Project
wmi added inline comments to D92584: [CSSPGO][llvm-profgen] Refactor to unify hashable interface for trace sample and context-sensitive counter.
Dec 14 2020, 12:39 PM · Restricted Project
wmi added inline comments to D92584: [CSSPGO][llvm-profgen] Refactor to unify hashable interface for trace sample and context-sensitive counter.
Dec 14 2020, 12:22 PM · Restricted Project
wmi added inline comments to D92347: [CSSPGO] Consume pseudo-probe-based AutoFDO profile.
Dec 14 2020, 10:40 AM · Restricted Project
wmi added inline comments to D92347: [CSSPGO] Consume pseudo-probe-based AutoFDO profile.
Dec 14 2020, 9:34 AM · Restricted Project

Dec 13 2020

wmi added inline comments to D92347: [CSSPGO] Consume pseudo-probe-based AutoFDO profile.
Dec 13 2020, 12:13 PM · Restricted Project

Dec 9 2020

wmi accepted D91878: [CSSPGO] Pseudo probe encoding and emission..

Please add a message for all the assertions. Other than that, LGTM.

Dec 9 2020, 8:29 PM · Restricted Project
wmi added a comment to D92806: Single function compilation mode..
In D92806#2443435, @hoy wrote:

@wmi Can you please recommend a reviewer for this? Thanks!

Dec 9 2020, 4:13 PM · Restricted Project, Restricted Project

Dec 8 2020

wmi added a comment to D91878: [CSSPGO] Pseudo probe encoding and emission..
In D91878#2441363, @hoy wrote:
In D91878#2440875, @wmi wrote:

Sorry for the delay in review. Some general questions.

I remember Wenlei mentioned in the RFC discussion that the overall binary size increase from pseudo probe is ~12% if CFG encoding is added. Is the size increase for object file from pseudo probe about the same as binary? I am asking because the total size of linker input matters for us. Previously we may use fission to separate the debug information from object file so it won't be accounted as linker input.

According to the description, The address range will be split into multiple ranges by the address of pseudo probe. If some BB somehow doesn't have block type pseudo probe, the addresses in the BB will be mistakenly associated with the pseudo probe located before it in the binary. I don't know whether you saw cases like that. A random thought is we may generate some fake pseudo probes for the BBs which don't have them, as a mark of the BB boundary.

Thanks for reviewing this patch!

Regarding the binary size, an object file will likely have the same size inflation with an executable since the size of the pseudo probe sections are propositional to the text section. If that is an issue to the linker, we can also perform fission to the pseudo probe data. Pseudo probe encoding is currently done on function (comdat) level (see MCPseudoProbeSection::Emit in MCPseudoProbe.cpp`), which means a function always gets a separate real absolute probe encoded in the first place followed by relative address delta encoded for subsequent probes. Therefore fission on function level should be fine. Fission on block level won't work, as you pointed out. Would function-level fission work for you?

Dec 8 2020, 5:00 PM · Restricted Project
wmi committed rG64e768536889: [SampleFDO] Store fixed length MD5 in NameTable instead of using ULEB128 if (authored by wmi).
[SampleFDO] Store fixed length MD5 in NameTable instead of using ULEB128 if
Dec 8 2020, 4:21 PM
wmi closed D92621: [SampleFDO] Store fixed length MD5 in NameTable instead of using ULEB128 if MD5 is used..
Dec 8 2020, 4:21 PM · Restricted Project
wmi added a comment to D91878: [CSSPGO] Pseudo probe encoding and emission..

Sorry for the delay in review. Some general questions.

Dec 8 2020, 12:42 PM · Restricted Project

Dec 7 2020

wmi accepted D92804: [ICP] Don't promote when target not defined in module.

LGTM.

Dec 7 2020, 5:28 PM · Restricted Project
wmi added a comment to D92621: [SampleFDO] Store fixed length MD5 in NameTable instead of using ULEB128 if MD5 is used..

We find that after this change, the elapsed time to build a large application distributively is reduced by 5% and the accumulative cpu time used for building is also reduced by 5%.

This is great, thanks for the patch! Just to clarify, did you mean 5% of end to end ThinLTO+AutoFDO build? I'm curious what is the total profile reading time your saw in terms of %? and how different that % is between md5 profile vs non-md5 profile? Currently we mostly use non-md5 profile, assuming md5 profile is more about saving profile size.

Dec 7 2020, 11:12 AM · Restricted Project

Dec 4 2020

wmi updated the diff for D92621: [SampleFDO] Store fixed length MD5 in NameTable instead of using ULEB128 if MD5 is used..

Address Hongtao's comment.

Dec 4 2020, 9:55 PM · Restricted Project
wmi added inline comments to D92621: [SampleFDO] Store fixed length MD5 in NameTable instead of using ULEB128 if MD5 is used..
Dec 4 2020, 4:28 PM · Restricted Project
wmi updated the diff for D92621: [SampleFDO] Store fixed length MD5 in NameTable instead of using ULEB128 if MD5 is used..

Address David and Hongtao's comments.

Dec 4 2020, 1:16 PM · Restricted Project

Dec 3 2020

wmi added inline comments to D92621: [SampleFDO] Store fixed length MD5 in NameTable instead of using ULEB128 if MD5 is used..
Dec 3 2020, 10:12 PM · Restricted Project
wmi retitled D92621: [SampleFDO] Store fixed length MD5 in NameTable instead of using ULEB128 if MD5 is used. from [SampleFDO] [SampleFDO] Store fixed length MD5 in NameTable instead of using ULEB128 if MD5 is used. to [SampleFDO] Store fixed length MD5 in NameTable instead of using ULEB128 if MD5 is used..
Dec 3 2020, 5:17 PM · Restricted Project
wmi requested review of D92621: [SampleFDO] Store fixed length MD5 in NameTable instead of using ULEB128 if MD5 is used..
Dec 3 2020, 5:17 PM · Restricted Project

Dec 2 2020

wmi accepted D90125: [CSSPGO] Infrastructure for context-sensitive Sample PGO and Inlining.

LGTM.

Dec 2 2020, 3:09 PM · Restricted Project
wmi added inline comments to D90125: [CSSPGO] Infrastructure for context-sensitive Sample PGO and Inlining.
Dec 2 2020, 3:09 PM · Restricted Project
wmi added inline comments to D90125: [CSSPGO] Infrastructure for context-sensitive Sample PGO and Inlining.
Dec 2 2020, 12:34 PM · Restricted Project
wmi accepted D91756: [CSSPGO] Pseudo probes for function calls..
In D91756#2427795, @hoy wrote:
In D91756#2427759, @wmi wrote:

Another question. Sorry for not bringing it up earlier. When a call with probe metadata attached is inlined, the probe will be gone or it will be kept somehow? I think you want to keep the probe especially for inline instance to reconstruct the context but I didn't figure it out how you did that from the description.

No problem. Sorry for not clarifying it in the description. When a callee is inlined, the probe metadata will go with the inlined instructions. The !dbg metadata of an inlined instruction is in form of a scope stack. The top of the stack is the instruction's original !dbg metadata and the bottom of the stack is the for the original callsite of the top-level inliner. Except for the top of the stack, all other elements of the stack actually refer to the nested inlined callsites whose discriminator fields (which actually represents a calliste probe) can be used to represent the inline context of an inlined PseudoProbeInst or a CallInst. I'll update the description.

Dec 2 2020, 11:46 AM · Restricted Project, Restricted Project

Dec 1 2020

wmi added a comment to D91756: [CSSPGO] Pseudo probes for function calls..

Another question. Sorry for not bringing it up earlier. When a call with probe metadata attached is inlined, the probe will be gone or it will be kept somehow? I think you want to keep the probe especially for inline instance to reconstruct the context but I didn't figure it out how you did that from the description.

Dec 1 2020, 11:08 PM · Restricted Project, Restricted Project
wmi added inline comments to D91756: [CSSPGO] Pseudo probes for function calls..
Dec 1 2020, 9:16 PM · Restricted Project, Restricted Project
wmi added inline comments to D91756: [CSSPGO] Pseudo probes for function calls..
Dec 1 2020, 3:33 PM · Restricted Project, Restricted Project

Nov 30 2020

wmi added inline comments to D90125: [CSSPGO] Infrastructure for context-sensitive Sample PGO and Inlining.
Nov 30 2020, 11:25 PM · Restricted Project

Nov 25 2020

wmi accepted D89723: [CSSPGO][llvm-profgen] Context-sensitive profile data generation.

LGTM.

Nov 25 2020, 4:32 PM · Restricted Project
wmi accepted D86499: [CSSPGO] Pseudo probe instrumentation pass.

LGTM.

Nov 25 2020, 4:03 PM · Restricted Project

Nov 24 2020

wmi accepted D91676: Avoid redundant work when computing vtable vcall visibility.
Nov 24 2020, 10:45 AM · Restricted Project
wmi accepted D91676: Avoid redundant work when computing vtable vcall visibility.

LGTM.

Nov 24 2020, 9:55 AM · Restricted Project
wmi added inline comments to D91676: Avoid redundant work when computing vtable vcall visibility.
Nov 24 2020, 9:34 AM · Restricted Project

Nov 20 2020

wmi accepted D86502: [CSSPGO] A Clang switch -fpseudo-probe-for-profiling for pseudo-probe instrumentation..

LGTM.

Nov 20 2020, 11:46 AM · Restricted Project, Restricted Project
wmi added inline comments to D86502: [CSSPGO] A Clang switch -fpseudo-probe-for-profiling for pseudo-probe instrumentation..
Nov 20 2020, 10:57 AM · Restricted Project, Restricted Project

Nov 19 2020

wmi accepted D86495: [CSSPGO] MIR target-independent pseudo instruction for pseudo-probe intrinsic.
Nov 19 2020, 10:47 PM · Restricted Project
wmi added a comment to D86490: [CSSPGO] IR intrinsic for pseudo-probe block instrumentation.

By the way, David has gone through the patches and talked to me offline saying he is ok with the patches generally. Thanks for your great work and patience!

Nov 19 2020, 10:07 PM · Restricted Project
wmi accepted D86490: [CSSPGO] IR intrinsic for pseudo-probe block instrumentation.

LGTM.

Nov 19 2020, 10:03 PM · Restricted Project
wmi added inline comments to D89723: [CSSPGO][llvm-profgen] Context-sensitive profile data generation.
Nov 19 2020, 11:22 AM · Restricted Project

Nov 18 2020

wmi accepted D89715: [CSSPGO][llvm-profgen] Instruction symbolization.
Nov 18 2020, 3:32 PM · Restricted Project

Nov 16 2020

wmi accepted D89712: [CSSPGO][llvm-profgen] Disassemble text sections.

LGTM.

Nov 16 2020, 2:20 PM · Restricted Project

Nov 13 2020

wmi added inline comments to D89712: [CSSPGO][llvm-profgen] Disassemble text sections.
Nov 13 2020, 5:14 PM · Restricted Project
wmi accepted D89707: [CSSPGO][llvm-profgen] Parse mmap events from perf script.

LGTM.

Nov 13 2020, 4:10 PM · Restricted Project