Page MenuHomePhabricator

fpetrogalli (Francesco Petrogalli)
User

Projects

User does not belong to any projects.

User Details

User Since
Oct 12 2016, 4:50 AM (153 w, 1 d)

Recent Activity

Yesterday

fpetrogalli committed rGcb032aa2c773: [SVFS] Vector Function ABI demangling. (authored by fpetrogalli).
[SVFS] Vector Function ABI demangling.
Thu, Sep 19, 10:48 AM
fpetrogalli committed rL372343: [SVFS] Vector Function ABI demangling..
[SVFS] Vector Function ABI demangling.
Thu, Sep 19, 10:47 AM
fpetrogalli closed D66024: [SVFS] Vector Function ABI name demangler..
Thu, Sep 19, 10:47 AM · Restricted Project
fpetrogalli committed rL372332: Request commit access for fpetrogalli.
Request commit access for fpetrogalli
Thu, Sep 19, 7:59 AM
fpetrogalli added a comment to D67762: [docs] Break long (>80) line. NFC.

I forgot to add the link to the differential review in the commit message. Adding the revision here:

Thu, Sep 19, 7:39 AM · Restricted Project
fpetrogalli committed rGcde4f727fffe: [docs] Break long (>80) line. NFC (authored by fpetrogalli).
[docs] Break long (>80) line. NFC
Thu, Sep 19, 7:23 AM
fpetrogalli committed rL372326: [docs] Break long (>80) line. NFC.
[docs] Break long (>80) line. NFC
Thu, Sep 19, 7:18 AM
fpetrogalli created D67762: [docs] Break long (>80) line. NFC.
Thu, Sep 19, 7:14 AM · Restricted Project

Tue, Sep 17

fpetrogalli updated the diff for D66024: [SVFS] Vector Function ABI name demangler..

Restore the function name to tryParseForVFABI.

Tue, Sep 17, 9:52 AM · Restricted Project
fpetrogalli updated the diff for D66024: [SVFS] Vector Function ABI name demangler..
Tue, Sep 17, 9:10 AM · Restricted Project
fpetrogalli added inline comments to D66024: [SVFS] Vector Function ABI name demangler..
Tue, Sep 17, 8:35 AM · Restricted Project
fpetrogalli updated the diff for D66024: [SVFS] Vector Function ABI name demangler..

I have removed the ParsedData struct as requested by @sdesmalen , my reasoning for keeping it was wrong.

Tue, Sep 17, 8:34 AM · Restricted Project

Mon, Sep 16

fpetrogalli added inline comments to D66024: [SVFS] Vector Function ABI name demangler..
Mon, Sep 16, 8:36 PM · Restricted Project

Fri, Sep 13

fpetrogalli added a comment to D67572: [SVFS] SearchVFSystem interface (WIP)..

This is a follow up patch to https://reviews.llvm.org/D66024

Fri, Sep 13, 2:49 PM · Restricted Project
fpetrogalli created D67572: [SVFS] SearchVFSystem interface (WIP)..
Fri, Sep 13, 2:40 PM · Restricted Project
fpetrogalli updated the diff for D66024: [SVFS] Vector Function ABI name demangler..

I added the Unknown ISA handling in the test ISAIndependentMangling, and updated the comments in the fuzzer.

Fri, Sep 13, 9:35 AM · Restricted Project
fpetrogalli updated the diff for D66024: [SVFS] Vector Function ABI name demangler..

I have split the implementation of VFShape into VFShape and VFInfo, as agreed.

Fri, Sep 13, 9:26 AM · Restricted Project
fpetrogalli added a comment to D66024: [SVFS] Vector Function ABI name demangler..
Fri, Sep 13, 6:51 AM · Restricted Project

Thu, Sep 12

fpetrogalli updated the diff for D66024: [SVFS] Vector Function ABI name demangler..

I have updated the patch according to your feedback. In particular, I have modified all the parsing method to use the ParseRet enum instead of booleans. It looks better now, thanks.

Thu, Sep 12, 2:34 PM · Restricted Project
fpetrogalli added inline comments to D66024: [SVFS] Vector Function ABI name demangler..
Thu, Sep 12, 2:22 PM · Restricted Project
fpetrogalli added inline comments to D66024: [SVFS] Vector Function ABI name demangler..
Thu, Sep 12, 9:36 AM · Restricted Project
fpetrogalli updated the diff for D66024: [SVFS] Vector Function ABI name demangler..

I have address all comments. Thank you for your feedback @jdoerfert @simoll @lebedev.ri !

Thu, Sep 12, 9:35 AM · Restricted Project

Wed, Sep 11

fpetrogalli updated the diff for D66024: [SVFS] Vector Function ABI name demangler..

I have deferred the handling of the mask token "M" to the VFParameter struct, introducing a new VFPAramKind that handles global function predication. This should be more in line with the aim of this implementation, which is to make VFShape independent from the Vector Function ABI. In particular, there are use cases that @simoll cares about where predication can be attached to single parameters, and not to the whole function.

Wed, Sep 11, 12:53 PM · Restricted Project
fpetrogalli updated the diff for D66024: [SVFS] Vector Function ABI name demangler..

In this new patch I have:

Wed, Sep 11, 11:07 AM · Restricted Project

Tue, Sep 10

fpetrogalli added inline comments to D66024: [SVFS] Vector Function ABI name demangler..
Tue, Sep 10, 12:53 PM · Restricted Project
fpetrogalli updated the diff for D66024: [SVFS] Vector Function ABI name demangler..
Tue, Sep 10, 12:53 PM · Restricted Project

Mon, Sep 9

fpetrogalli commandeered D66025: SVFS implementation according to RFC: Interface user provided vector functions with the vectorizer..

As with https://reviews.llvm.org/D66024, I am picking up this work from @aranisumedh (thank you!).

Mon, Sep 9, 2:05 PM · Restricted Project
fpetrogalli updated the diff for D67365: [VectorUtils] Remove unused include. [NFC].

I have added the forward declaration.

Mon, Sep 9, 1:50 PM · Restricted Project
fpetrogalli updated the diff for D66024: [SVFS] Vector Function ABI name demangler..

I have done the following updates:

Mon, Sep 9, 1:29 PM · Restricted Project
fpetrogalli added a reviewer for D67365: [VectorUtils] Remove unused include. [NFC]: fhahn.
Mon, Sep 9, 1:05 PM · Restricted Project
fpetrogalli created D67365: [VectorUtils] Remove unused include. [NFC].
Mon, Sep 9, 1:05 PM · Restricted Project

Thu, Sep 5

fpetrogalli added a comment to D66024: [SVFS] Vector Function ABI name demangler..
it would be better to parse the mangled string incrementally, rather than extracting each feature from the string individually.
Thu, Sep 5, 3:26 PM · Restricted Project
fpetrogalli updated the diff for D66024: [SVFS] Vector Function ABI name demangler..

I have updated the code according to the last round of reviews. I have removed pieces of code that were not needed anymore for testing.

Thu, Sep 5, 3:23 PM · Restricted Project

Fri, Aug 30

fpetrogalli updated the diff for D66024: [SVFS] Vector Function ABI name demangler..

Cosmetic changes to a comment.

Fri, Aug 30, 11:52 AM · Restricted Project
fpetrogalli retitled D66024: [SVFS] Vector Function ABI name demangler. from [SVFS] Search vector function system. to [SVFS] Vector Function ABI demangler..
Fri, Aug 30, 11:34 AM · Restricted Project
fpetrogalli updated the diff for D66024: [SVFS] Vector Function ABI name demangler..

I have uploaded a new version with the fuzzer requested by @lebedev.ri .

Fri, Aug 30, 11:34 AM · Restricted Project

Wed, Aug 28

fpetrogalli added inline comments to D66024: [SVFS] Vector Function ABI name demangler..
Wed, Aug 28, 2:04 PM · Restricted Project
fpetrogalli commandeered D66024: [SVFS] Vector Function ABI name demangler..

Hi all,

Wed, Aug 28, 1:55 PM · Restricted Project

Jul 19 2019

fpetrogalli updated subscribers of D64095: SVFS implementation according to RFC: Interface user provided vector functions with the vectorizer..
Jul 19 2019, 3:46 PM · Restricted Project

Jun 4 2019

fpetrogalli added a comment to D60583: [AArch64] Implement Vector Funtion ABI name mangling..

Why/Where did we decide to clobber the attribute list with "non-existent function names"?
I don't think an attribute list like this:
attributes #1 = { "_ZGVsM2v_foo" "_ZGVsM32v_foo" "_ZGVsM4v_foo" "_ZGVsM6v_foo" "_ZGVsM8v_foo" "_ZGVsMxv_foo" ...
is helpful in any way, e.g., this would require us to search through all attributes and interpret them one by one.

Jun 4 2019, 2:08 PM · Restricted Project, Restricted Project

Apr 15 2019

fpetrogalli updated the diff for D60583: [AArch64] Implement Vector Funtion ABI name mangling..
Apr 15 2019, 11:51 AM · Restricted Project, Restricted Project
fpetrogalli added inline comments to D60583: [AArch64] Implement Vector Funtion ABI name mangling..
Apr 15 2019, 11:51 AM · Restricted Project, Restricted Project
fpetrogalli updated the diff for D60583: [AArch64] Implement Vector Funtion ABI name mangling..
Apr 15 2019, 9:22 AM · Restricted Project, Restricted Project
fpetrogalli added inline comments to D60583: [AArch64] Implement Vector Funtion ABI name mangling..
Apr 15 2019, 9:22 AM · Restricted Project, Restricted Project

Apr 12 2019

fpetrogalli updated the summary of D60583: [AArch64] Implement Vector Funtion ABI name mangling..
Apr 12 2019, 1:43 PM · Restricted Project, Restricted Project
fpetrogalli updated the diff for D60583: [AArch64] Implement Vector Funtion ABI name mangling..
Apr 12 2019, 1:43 PM · Restricted Project, Restricted Project
fpetrogalli updated the diff for D60583: [AArch64] Implement Vector Funtion ABI name mangling..
Apr 12 2019, 1:28 PM · Restricted Project, Restricted Project
fpetrogalli added inline comments to D60583: [AArch64] Implement Vector Funtion ABI name mangling..
Apr 12 2019, 1:28 PM · Restricted Project, Restricted Project
fpetrogalli updated the diff for D60583: [AArch64] Implement Vector Funtion ABI name mangling..
Apr 12 2019, 12:08 PM · Restricted Project, Restricted Project
fpetrogalli added a comment to D60583: [AArch64] Implement Vector Funtion ABI name mangling..

Thank you @ABataev.

Apr 12 2019, 12:05 PM · Restricted Project, Restricted Project
fpetrogalli updated the diff for D60583: [AArch64] Implement Vector Funtion ABI name mangling..
Apr 12 2019, 12:05 PM · Restricted Project, Restricted Project

Apr 11 2019

fpetrogalli added a reviewer for D60583: [AArch64] Implement Vector Funtion ABI name mangling.: ABataev.
Apr 11 2019, 3:05 PM · Restricted Project, Restricted Project
fpetrogalli created D60583: [AArch64] Implement Vector Funtion ABI name mangling..
Apr 11 2019, 2:08 PM · Restricted Project, Restricted Project

Apr 10 2019

fpetrogalli added a comment to D59952: [VPLAN] Minor improvement to testing and debug messages..

Hi! One of the tests was missing "REQUIRES: asserts", so it failed on a non-assert build. I took the liberty to add that to the test in r358057. I hope that was okay!

Apr 10 2019, 6:15 AM · Restricted Project

Apr 9 2019

fpetrogalli added a comment to D59952: [VPLAN] Minor improvement to testing and debug messages..

Thank you all.

Apr 9 2019, 10:07 AM · Restricted Project
fpetrogalli added a comment to D59952: [VPLAN] Minor improvement to testing and debug messages..

@dcaballe , gentle ping.

Apr 9 2019, 8:34 AM · Restricted Project

Mar 29 2019

fpetrogalli updated the diff for D59952: [VPLAN] Minor improvement to testing and debug messages..

I have addressed all concern raised in this review.

Mar 29 2019, 2:38 PM · Restricted Project
fpetrogalli added a comment to D59952: [VPLAN] Minor improvement to testing and debug messages..
Mar 29 2019, 2:21 PM · Restricted Project
fpetrogalli added a comment to D59952: [VPLAN] Minor improvement to testing and debug messages..

Should we add a small lit test that covers the new expected behavior? Probably reusing one from D57598 with the stress testing flag would suffice.

Mar 29 2019, 6:49 AM · Restricted Project

Mar 28 2019

fpetrogalli added a comment to D59952: [VPLAN] Minor improvement to testing and debug messages..

I have updated the patch to address all the concerns. Now VF = 4 for stress testing only if the computed VF does not produce vector code (VF < 2).

Mar 28 2019, 2:13 PM · Restricted Project
fpetrogalli updated the diff for D59952: [VPLAN] Minor improvement to testing and debug messages..
Mar 28 2019, 2:07 PM · Restricted Project
fpetrogalli added a comment to D59952: [VPLAN] Minor improvement to testing and debug messages..

Yep, I agree on that we should keep the stress testing mechanism. It will be very useful to make sure that the construction and predication (and maybe other transformation) are robust enough since we can run it on loop nests that are not necessarily vectorizable.
Something important, though, is that we shouldn't use this mechanism to bypass legality or pragma simd requirements to vectorize a loop, i.e., we shouldn't use it to generate actual vector code.
What are you trying to achieve, Francesco?

I suggested removing setting VF = 4 for VPlanStressTest now that we programmatically determine the VF, to streamline things a bit. In a way, just have VPlanStressTest mean: build a VPlan for any loop you can, with the automatically chosen VF.

Mar 28 2019, 1:47 PM · Restricted Project
fpetrogalli added a comment to D59952: [VPLAN] Minor improvement to testing and debug messages..

Hi all,

Mar 28 2019, 1:12 PM · Restricted Project
fpetrogalli created D59952: [VPLAN] Minor improvement to testing and debug messages..
Mar 28 2019, 11:51 AM · Restricted Project

Mar 27 2019

fpetrogalli added a comment to D57598: [VPLAN] Determine Vector Width programmatically..

Thanks Francesco! I'll commit the change tomorrow, unless @hsaito does it today :)

Mar 27 2019, 2:21 PM · Restricted Project
fpetrogalli updated the diff for D57598: [VPLAN] Determine Vector Width programmatically..

I forgot to update the comment in the test as requested from @fhahn . Now it is done.

Mar 27 2019, 12:35 PM · Restricted Project
fpetrogalli added inline comments to D57598: [VPLAN] Determine Vector Width programmatically..
Mar 27 2019, 12:33 PM · Restricted Project
fpetrogalli updated the diff for D57598: [VPLAN] Determine Vector Width programmatically..

I have addressed last round of comments from @fhahn .

Mar 27 2019, 12:33 PM · Restricted Project

Mar 22 2019

fpetrogalli added a comment to D57598: [VPLAN] Determine Vector Width programmatically..
Mar 22 2019, 7:02 PM · Restricted Project
fpetrogalli added a comment to D57598: [VPLAN] Determine Vector Width programmatically..

@npanchen , @fhahn , gentle ping :)

Mar 22 2019, 3:33 PM · Restricted Project

Mar 14 2019

fpetrogalli added inline comments to D57598: [VPLAN] Determine Vector Width programmatically..
Mar 14 2019, 1:49 PM · Restricted Project
fpetrogalli updated the diff for D57598: [VPLAN] Determine Vector Width programmatically..
Mar 14 2019, 1:49 PM · Restricted Project
fpetrogalli added inline comments to D57598: [VPLAN] Determine Vector Width programmatically..
Mar 14 2019, 12:43 PM · Restricted Project
fpetrogalli added inline comments to D57598: [VPLAN] Determine Vector Width programmatically..
Mar 14 2019, 11:50 AM · Restricted Project
fpetrogalli updated the diff for D57598: [VPLAN] Determine Vector Width programmatically..
Mar 14 2019, 11:50 AM · Restricted Project
fpetrogalli added inline comments to D57598: [VPLAN] Determine Vector Width programmatically..
Mar 14 2019, 9:57 AM · Restricted Project

Mar 13 2019

fpetrogalli updated the diff for D57598: [VPLAN] Determine Vector Width programmatically..

Addressed second round of comments from @hsaito.

Mar 13 2019, 8:30 PM · Restricted Project
fpetrogalli updated the diff for D57598: [VPLAN] Determine Vector Width programmatically..

I mostly changes to code to use the infrastructure that LLVM already provides to determine the vectorization factor.

Mar 13 2019, 1:25 PM · Restricted Project
fpetrogalli added a comment to D57598: [VPLAN] Determine Vector Width programmatically..

Hi all,

Mar 13 2019, 1:24 PM · Restricted Project

Feb 4 2019

fpetrogalli added a comment to D57598: [VPLAN] Determine Vector Width programmatically..

Thank you for looking into this.

Feb 4 2019, 8:38 AM · Restricted Project

Feb 1 2019

fpetrogalli created D57598: [VPLAN] Determine Vector Width programmatically..
Feb 1 2019, 9:51 AM · Restricted Project

Dec 11 2018

fpetrogalli added a comment to D54412: [RFC] Re-implementing -fveclib with OpenMP.

Hello again - I have posted the RFC on llvm-dev: http://lists.llvm.org/pipermail/llvm-dev/2018-December/128426.html

Dec 11 2018, 7:51 PM
fpetrogalli added inline comments to D53927: [AArch64] Enable libm vectorized functions via SLEEF.
Dec 11 2018, 2:07 PM · Restricted Project
fpetrogalli added a comment to D54412: [RFC] Re-implementing -fveclib with OpenMP.

Is there a document that explains how to submit an RFC?

Hi Francesco,

It's not that complex for a document. The text you have here would be more than fine, and the discussion that happened is exactly what we look for, but Phab reviews only copy llvm-commits, not llvm-dev, and that is the main problem.

Dec 11 2018, 1:55 PM
fpetrogalli added inline comments to D53927: [AArch64] Enable libm vectorized functions via SLEEF.
Dec 11 2018, 1:39 PM · Restricted Project
fpetrogalli added a comment to D54412: [RFC] Re-implementing -fveclib with OpenMP.

Just to clarify - to me it is not clear whether I am required to create a new review on phabricator to be able to refer to this proposal as an RFC, and if so, which fields have to be set, and with which values, to make sure that the proposal is recognized as an official RFC by the LLVM community.

Dec 11 2018, 1:25 PM

Dec 10 2018

fpetrogalli added a comment to D54412: [RFC] Re-implementing -fveclib with OpenMP.
Dec 10 2018, 8:34 AM
fpetrogalli added a comment to D54412: [RFC] Re-implementing -fveclib with OpenMP.

I feel like pointing out that this RFC did *NOT* go to the lists, which makes zero sense, since it's RFC.

Dec 10 2018, 8:18 AM

Nov 29 2018

fpetrogalli updated the diff for D54412: [RFC] Re-implementing -fveclib with OpenMP.
Nov 29 2018, 12:50 PM
fpetrogalli added inline comments to D54412: [RFC] Re-implementing -fveclib with OpenMP.
Nov 29 2018, 12:46 PM

Nov 27 2018

fpetrogalli added a comment to D54412: [RFC] Re-implementing -fveclib with OpenMP.

@hsaito , there are another couple of reasons for which I'd prefer to keep the -fveclib option.

Nov 27 2018, 1:14 PM

Nov 26 2018

fpetrogalli added inline comments to D54412: [RFC] Re-implementing -fveclib with OpenMP.
Nov 26 2018, 7:36 PM

Nov 25 2018

fpetrogalli added inline comments to D54412: [RFC] Re-implementing -fveclib with OpenMP.
Nov 25 2018, 8:29 PM
fpetrogalli updated the diff for D54412: [RFC] Re-implementing -fveclib with OpenMP.

I think i have addressed all your comments. Please let me know if I have missed anything.

Nov 25 2018, 8:26 PM

Nov 20 2018

fpetrogalli accepted D53927: [AArch64] Enable libm vectorized functions via SLEEF.
Nov 20 2018, 10:05 AM · Restricted Project
fpetrogalli added a comment to D53927: [AArch64] Enable libm vectorized functions via SLEEF.

I am sorry for having kept you waiting for the patch. I was not fully aware of the approval process upstream (didn’t do much contributions upstream myself), and before being able to approve I had to understand the process.

Nov 20 2018, 7:36 AM · Restricted Project

Nov 19 2018

fpetrogalli added a comment to D54412: [RFC] Re-implementing -fveclib with OpenMP.

IMPORTANT NOTICE: The contents of this email and any attachments are confidential and may also be privileged. If you are not the intended recipient, please notify the sender immediately and do not disclose the contents to any other person, use it for any purpose, or store or copy the information in any medium. Thank you.

Nov 19 2018, 2:47 PM
fpetrogalli added a comment to D54412: [RFC] Re-implementing -fveclib with OpenMP.

I still would like to see the ability to differentiate quasi-intrinsic veclib functions versus user-written vector functions and enable veclib w/o enabling user-written vector functions, even if -fopenmp-simd becomes default. Some of our ICC customers take perf data with and w/o openmp simd (and proprietary simd pragma) and they expect SVML to be used for both cases. I don't have much preferences in how to accomplish it. For example, if we'd like to keep the header file contents totally standard based, we could use something like a command line flag -fclang-vector-intrinsic=file1,file2,..,fileN to indicate that "declare simd" in those header file is "quasi-intrinsic". We could silently do this for the veclibs shipped with clang (e.g., check if "declare simd" is coming from clang include directory) ---- I'm fine with this ----, but that's a bit unfair to the 3rd party veclibs.

Nov 19 2018, 2:44 PM
fpetrogalli added a comment to D54412: [RFC] Re-implementing -fveclib with OpenMP.

I would avoid having a clang-specific pragma. By just using the OpenMP one, we provide the library vendors with a mechanism that is based on a standard and therefore can be used by any compiler.

Nov 19 2018, 7:18 AM

Nov 16 2018

fpetrogalli added inline comments to D54412: [RFC] Re-implementing -fveclib with OpenMP.
Nov 16 2018, 1:14 PM