search for: asbirlea

...support > for `-foo-pass`. > Hm, is there any written rationale behind such a decision? > I would have thought that -passes= is the temporary solution, not the > other way around. > This is really a separate issue that's somewhat orthogonal to the original issue, but someone like asbirlea may be able to chime in more. Maybe a new RFC thread? > > > However, until NPM is enabled by default, we still want tests using opt > to use the legacy PM by default. > > We could attempt to make `-passes=` work with the legacy PM and have a > legacy vs new PM flag, but given...

...nclined to favor a tighter deadline, the motivation here being to ensure that working on potential blockers is prioritized with plenty of time to spare, so the switch remains time-boxed. Best, Alina > Thanks, > Sjoerd. > > ------------------------------ > *From:* Alina Sbirlea <asbirlea at google.com> > *Sent:* 24 July 2020 19:51 > *To:* Sjoerd Meijer <Sjoerd.Meijer at arm.com> > *Cc:* Philip Reames <listmail at philipreames.com>; Chandler Carruth < > chandlerc at gmail.com>; Eric Christopher <echristo at gmail.com>; llvm-dev < > llvm-d...

...; ------------------------------ > *From:* llvm-dev <llvm-dev-bounces at lists.llvm.org> on behalf of Eric > Christopher via llvm-dev <llvm-dev at lists.llvm.org> > *Sent:* 23 July 2020 02:05 > *To:* Philip Reames <listmail at philipreames.com>; Alina Sbirlea < > asbirlea at google.com>; Chandler Carruth <chandlerc at gmail.com> > *Cc:* llvm-dev <llvm-dev at lists.llvm.org> > *Subject:* Re: [llvm-dev] New pass manager for optimization pipeline > status and questions > > FWIW I'm in favor of this direction while making sure that we k...

(minor inline additions) On 09/13/2018 01:51 AM, Chandler Carruth wrote: > Haven't had time to dig into this, but wanted to add +Alina Sbirlea > <mailto:asbirlea at google.com> to the thread as she has been working on > promotion and other aspects of LICM for a long time here. Thanks! > On Wed, Sep 12, 2018 at 11:41 PM Philip Reames > <listmail at philipreames.com <mailto:listmail at philipreames.com>> wrote: > > I'm...

...s, Alina On Thu, Sep 13, 2018 at 1:43 PM Philip Reames <listmail at philipreames.com<mailto:listmail at philipreames.com>> wrote: (minor inline additions) On 09/13/2018 01:51 AM, Chandler Carruth wrote: Haven't had time to dig into this, but wanted to add +Alina Sbirlea<mailto:asbirlea at google.com> to the thread as she has been working on promotion and other aspects of LICM for a long time here. Thanks! On Wed, Sep 12, 2018 at 11:41 PM Philip Reames <listmail at philipreames.com<mailto:listmail at philipreames.com>> wrote: I'm thinking about making some sem...

...is complete: Failure [failed test-check-all] Build details are at http://lab.llvm.org:8011/builders/lld-x86_64-win/builds/2255 blamelist: LLVM GN Syncbot <llvmgnsyncbot at gmail.com>, Eric Astor <epastor at google.com>, Craig Topper <craig.topper at intel.com>, Alina Sbirlea <asbirlea at google.com>, Nico Weber <thakis at chromium.org>, Amara I also got email with pointers to: http://green.lab.llvm.org/green//job/clang-stage1-RA/14180/consoleFull#-1417328700a1ca8a51-895e-46c6-af87-ce24fa4cd561 Chances are that there's something genuinely broken somewhere (maybe co...

I'm thinking about making some semi radical changes to load store promotion works in LICM, and figured it would be a good idea to get buy in before I actually started writing code.  :) TLDR: legality of sinking stores to exits is hard, can we separate load handling into a super aggressive form of PRE, and use predicated stores to avoid solving legality question? Background We've

FWIW I'm in favor of this direction while making sure that we keep focus on removing the vestiges of the old pass manager for the code health impact to the project. -eric On Wed, Jul 22, 2020 at 3:15 PM Philip Reames via llvm-dev < llvm-dev at lists.llvm.org> wrote: > (I'm probably going to derail your thread, sorry about that.) > > I think at this point, we should just

...o:Ashutosh.Nema at amd.com] Sent: Wednesday, June 29, 2016 9:50 PM To: Saito, Hideki <hideki.saito at intel.com>; Demikhovsky, Elena <elena.demikhovsky at intel.com>; silviu.baranga at gmail.com; Zaks, Ayal <ayal.zaks at intel.com> Cc: llvm-dev <llvm-dev at lists.llvm.org>; asbirlea at google.com; renato.golin at linaro.org; mssimpso at codeaurora.org; kv.bhat at samsung.com; Shahid, Asghar-ahmad <Asghar-ahmad.Shahid at amd.com>; sanjoy at playingwithpointers.com; mzolotukhin at apple.com; Michael Kuperstein <mkuper at google.com> Subject: RE: [llvm-dev] [Proposal]...

...Thu, Sep 13, 2018 at 1:43 PM Philip Reames <listmail at philipreames.com> > wrote: > >> (minor inline additions) >> >> On 09/13/2018 01:51 AM, Chandler Carruth wrote: >> >> Haven't had time to dig into this, but wanted to add +Alina Sbirlea >> <asbirlea at google.com> to the thread as she has been working on promotion >> and other aspects of LICM for a long time here. >> >> Thanks! >> >> On Wed, Sep 12, 2018 at 11:41 PM Philip Reames <listmail at philipreames.com> >> wrote: >> >>> I'm...

On Tue, Sep 1, 2020 at 3:32 PM David Blaikie <dblaikie at gmail.com> wrote: > On Tue, Sep 1, 2020 at 12:07 PM Nico Weber via cfe-dev < > cfe-dev at lists.llvm.org> wrote: > >> Hi, >> >> llvmbb's job is to inform people of build breaks. However, it seems to >> trigger for a big list of bots, and at least one of them seems to always be >>

After talking with some NPM people, I believe the ultimate goal after NPM is enabled by default is to only support `-passes=`, and remove support for `-foo-pass`. However, until NPM is enabled by default, we still want tests using opt to use the legacy PM by default. We could attempt to make `-passes=` work with the legacy PM and have a legacy vs new PM flag, but given the design/syntax of

One common concern raised for cases where Loop Vectorizer generate bigger types than target supported: Based on VF currently we check the cost and generate the expected set of instruction[s] for bigger type. It has two challenges for bigger types cost is not always correct and code generation may not generate efficient instruction[s]. Probably can depend on the support provided by below RFC by

David, That's right, WebAssembly does not have a way to conditionally use a feature or even do runtime feature testing right now. It's on our roadmap of things to design and standardize, but it is still a long way off. > Another direction would be to require the features to be specified consistently for all components of the build, I guess - if that's the net effect anyway. Would

>Vectorizer's output should be as clean as vector code can be so that analyses and optimizers downstream can >do a great job optimizing. Guess I should clarify this philosophical position of mine. In terms of vector code optimization that complicates the output of vectorizer: If vectorizer is the best place to perform the optimization, it should do so. This includes the cases like