I haven't seen much about the round tables at the dev meeting, on the list or the web page, anyone keeping track of those? I propose we do the usual one for VPlan to see who is working on what in the future, and one on SVE, so that we can sync on what's missing for this stage (IR) and what's the priorities for the next stages. I'd also like to discuss transformations in MLIR, like parallel extensions, polyhedral, threading, open mp, etc. How feasible would be to use it as an intermediate-IR for some LLVM passes, in addition to being an input-IR to the middle end. Should we track this somewhere? At least having a list would be nice, so that people know what to look for, leaving the times to be defined later, if necessary. cheers, --renato
Hi, On Wed, Sep 18, 2019 at 7:16 AM Renato Golin via llvm-dev < llvm-dev at lists.llvm.org> wrote:> I haven't seen much about the round tables at the dev meeting, on the > list or the web page, anyone keeping track of those? > > I propose we do the usual one for VPlan to see who is working on what > in the future, and one on SVE, so that we can sync on what's missing > for this stage (IR) and what's the priorities for the next stages. > > I'd also like to discuss transformations in MLIR, like parallel > extensions, polyhedral, threading, open mp, etc. How feasible would be > to use it as an intermediate-IR for some LLVM passes, in addition to > being an input-IR to the middle end. >Yeah we're really interested in topics about MLIR related thing! The one we had at EuroLLVM was very interesting. I'm interested in a round-table to discuss possible evolutions of GitHub workflow (post-transition) and discussing experience with pre-merge testing, pull-request, etc.> > Should we track this somewhere? At least having a list would be nice, > so that people know what to look for, leaving the times to be defined > later, if necessary. >I asked Tanya about this last week and she was wrapping up a bunch of things with the program, but we should have before the end of the month a form to accept round table suggestions before the event, that include picking time slots (hence having the schedule for the talks before to help picking convenient time slots for a round-table topic). Best, -- Mehdi -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20190918/bcbf387f/attachment.html>
Op do 19 sep. 2019 om 08:56 schreef Mehdi AMINI via llvm-dev < llvm-dev at lists.llvm.org>:> Hi, > > On Wed, Sep 18, 2019 at 7:16 AM Renato Golin via llvm-dev < > llvm-dev at lists.llvm.org> wrote: > >> I haven't seen much about the round tables at the dev meeting, on the >> list or the web page, anyone keeping track of those? >> >> I propose we do the usual one for VPlan to see who is working on what >> in the future, and one on SVE, so that we can sync on what's missing >> for this stage (IR) and what's the priorities for the next stages. >> >> I'd also like to discuss transformations in MLIR, like parallel >> extensions, polyhedral, threading, open mp, etc. How feasible would be >> to use it as an intermediate-IR for some LLVM passes, in addition to >> being an input-IR to the middle end. >> > > Yeah we're really interested in topics about MLIR related thing! The one > we had at EuroLLVM was very interesting. > > I'm interested in a round-table to discuss possible evolutions of GitHub > workflow (post-transition) and discussing experience with pre-merge > testing, pull-request, etc. >I'm interested in a round-table to discuss the pros and cons of potentially moving away from bugzilla to another issue tracking system, with the GitHub issue tracking system being the most obvious alternative. I guess it may be best for that to be a different round table to a Github workflow round-table, but there may be some overlap with a workflow roundtable? Thanks, Kristof -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20190919/0b2f9fbe/attachment.html>
> On Sep 18, 2019, at 11:55 PM, Mehdi AMINI <joker.eph at gmail.com> wrote: > > Hi, > >> On Wed, Sep 18, 2019 at 7:16 AM Renato Golin via llvm-dev <llvm-dev at lists.llvm.org> wrote: >> I haven't seen much about the round tables at the dev meeting, on the >> list or the web page, anyone keeping track of those? >> >> I propose we do the usual one for VPlan to see who is working on what >> in the future, and one on SVE, so that we can sync on what's missing >> for this stage (IR) and what's the priorities for the next stages. >> >> I'd also like to discuss transformations in MLIR, like parallel >> extensions, polyhedral, threading, open mp, etc. How feasible would be >> to use it as an intermediate-IR for some LLVM passes, in addition to >> being an input-IR to the middle end. > > Yeah we're really interested in topics about MLIR related thing! The one we had at EuroLLVM was very interesting. > > I'm interested in a round-table to discuss possible evolutions of GitHub workflow (post-transition) and discussing experience with pre-merge testing, pull-request, etc. > >> >> Should we track this somewhere? At least having a list would be nice, >> so that people know what to look for, leaving the times to be defined >> later, if necessary. > > I asked Tanya about this last week and she was wrapping up a bunch of things with the program, but we should have before the end of the month a form to accept round table suggestions before the event, that include picking time slots (hence having the schedule for the talks before to help picking convenient time slots for a round-table topic). >Yes, this is dependent upon the schedule being fixed. The timeline is still about right. I have a form people submit to schedule slots. -Tanya> Best, > > -- > Mehdi > >-------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20190919/23d290f4/attachment.html>
Possibly Parallel Threads
- Writing loop transformations on the right representation is more productive
- Writing loop transformations on the right representation is more productive
- [RFC] Writing loop transformations on the right representation is more productive
- Writing loop transformations on the right representation is more productive
- [RFC] Polly Status and Integration