February 2012 - petsc-dev - mailman3.cels.anl.gov

Re: [petsc-dev] write access to petsc-dev bitbucket repo
by Barry Smith 10 Feb '12

10 Feb '12

On Feb 10, 2012, at 11:25 AM, Matthew Knepley wrote: > On Fri, Feb 10, 2012 at 11:16 AM, Barry Smith <bsmith(a)mcs.anl.gov> wrote: > > On Feb 10, 2012, at 11:06 AM, Matthew Knepley wrote: > > > On Fri, Feb 10, 2012 at 11:04 AM, Satish Balay <balay(a)mcs.anl.gov> wrote: > > sorry about the conflicting messages. We haven't yet completely > > switched to bitbucket. So please continue to push to iit repo. > > > > We have not decided that yet. > > We never decided to switch to bitbucket but suddenly we haven't decided to use petsc.cs.iit.edu? > > I remember you telling Sean to put up the BB site, which happened before iit was back up. Yes, I reminded him that since all hg repositories are equivalent that he could set up a petsc-dev that we could use since petsc.cs.itt.edu was down. No one decided at that instant it was to become the permanent one. If you remember we did this once or twice before in the past couple of years when petsc.cs.itt.edu was down. BTW: I have found no records of bitbucket uptimes on their site or elsewhere so what evidence do you have to make your claim that it is more reliable (reliable meaning minutes per year where one can pull or push)? Barry > > So, we have reduced the argument to: > > directories vs. better management, reliability, availability > > If having directories is honestly a must-have for you, delete the BB site. If instead we are weighing > pros and cons, I think the choice is clear. > > Matt > > > > > Matt > > > > thanks, > > Satish > > > > On Fri, 10 Feb 2012, Brad Aagaard wrote: > > > > > I have write access to the iit petsc-dev repo so that I can push fixes to > > > Matt's Sieve stuff. Matt had me switch over to the bitbucket repo and now I > > > don't have write access. I assume I need a bitbucket login, so I created one > > > (baagaard) and added my public ssh key. If one of you could enable write > > > access to the petsc-dev repo, I would appreciate it. > > > > > > Thanks, > > > Brad > > > > > > > > > > > > > -- > > What most experimenters take for granted before they begin their experiments is infinitely more interesting than any results to which their experiments lead. > > -- Norbert Wiener > > > > > -- > What most experimenters take for granted before they begin their experiments is infinitely more interesting than any results to which their experiments lead. > -- Norbert Wiener

1 0

Re: [petsc-dev] petsc-dev on bitbucket
by Jed Brown 10 Feb '12

10 Feb '12

One other point is integrated issue tracking. If someone makes a feature request, we currently manage it purely by not losing the email. With bitbucket issues, users can "follow" issues so when we include "closes #123" (and other status updates) in a commit message, interested users get notified. I know we try to fix all bugs and "good" feature requests within the hour, but there are topics that take longer. projects.html isn't an optimal solution.

1 0

Re: [petsc-dev] write access to petsc-dev bitbucket repo
by Matthew Knepley 10 Feb '12

10 Feb '12

On Fri, Feb 10, 2012 at 11:16 AM, Barry Smith <bsmith(a)mcs.anl.gov> wrote: > > On Feb 10, 2012, at 11:06 AM, Matthew Knepley wrote: > > > On Fri, Feb 10, 2012 at 11:04 AM, Satish Balay <balay(a)mcs.anl.gov> > wrote: > > sorry about the conflicting messages. We haven't yet completely > > switched to bitbucket. So please continue to push to iit repo. > > > > We have not decided that yet. > > We never decided to switch to bitbucket but suddenly we haven't decided > to use petsc.cs.iit.edu? > I remember you telling Sean to put up the BB site, which happened before iit was back up. So, we have reduced the argument to: directories vs. better management, reliability, availability If having directories is honestly a must-have for you, delete the BB site. If instead we are weighing pros and cons, I think the choice is clear. Matt > > > > Matt > > > > thanks, > > Satish > > > > On Fri, 10 Feb 2012, Brad Aagaard wrote: > > > > > I have write access to the iit petsc-dev repo so that I can push fixes > to > > > Matt's Sieve stuff. Matt had me switch over to the bitbucket repo and > now I > > > don't have write access. I assume I need a bitbucket login, so I > created one > > > (baagaard) and added my public ssh key. If one of you could enable > write > > > access to the petsc-dev repo, I would appreciate it. > > > > > > Thanks, > > > Brad > > > > > > > > > > > > > -- > > What most experimenters take for granted before they begin their > experiments is infinitely more interesting than any results to which their > experiments lead. > > -- Norbert Wiener > > -- What most experimenters take for granted before they begin their experiments is infinitely more interesting than any results to which their experiments lead. -- Norbert Wiener

1 0

Re: [petsc-dev] write access to petsc-dev bitbucket repo
by Barry Smith 10 Feb '12

10 Feb '12

On Feb 10, 2012, at 11:06 AM, Matthew Knepley wrote: > On Fri, Feb 10, 2012 at 11:04 AM, Satish Balay <balay(a)mcs.anl.gov> wrote: > sorry about the conflicting messages. We haven't yet completely > switched to bitbucket. So please continue to push to iit repo. > > We have not decided that yet. We never decided to switch to bitbucket but suddenly we haven't decided to use petsc.cs.iit.edu? > > Matt > > thanks, > Satish > > On Fri, 10 Feb 2012, Brad Aagaard wrote: > > > I have write access to the iit petsc-dev repo so that I can push fixes to > > Matt's Sieve stuff. Matt had me switch over to the bitbucket repo and now I > > don't have write access. I assume I need a bitbucket login, so I created one > > (baagaard) and added my public ssh key. If one of you could enable write > > access to the petsc-dev repo, I would appreciate it. > > > > Thanks, > > Brad > > > > > > > -- > What most experimenters take for granted before they begin their experiments is infinitely more interesting than any results to which their experiments lead. > -- Norbert Wiener

1 0

Re: [petsc-dev] write access to petsc-dev bitbucket repo
by Matthew Knepley 10 Feb '12

10 Feb '12

On Fri, Feb 10, 2012 at 11:04 AM, Satish Balay <balay(a)mcs.anl.gov> wrote: > sorry about the conflicting messages. We haven't yet completely > switched to bitbucket. So please continue to push to iit repo. > We have not decided that yet. Matt > thanks, > Satish > > On Fri, 10 Feb 2012, Brad Aagaard wrote: > > > I have write access to the iit petsc-dev repo so that I can push fixes to > > Matt's Sieve stuff. Matt had me switch over to the bitbucket repo and > now I > > don't have write access. I assume I need a bitbucket login, so I created > one > > (baagaard) and added my public ssh key. If one of you could enable write > > access to the petsc-dev repo, I would appreciate it. > > > > Thanks, > > Brad > > > > -- What most experimenters take for granted before they begin their experiments is infinitely more interesting than any results to which their experiments lead. -- Norbert Wiener

1 0

Re: [petsc-dev] write access to petsc-dev bitbucket repo
by Satish Balay 10 Feb '12

10 Feb '12

sorry about the conflicting messages. We haven't yet completely switched to bitbucket. So please continue to push to iit repo. thanks, Satish On Fri, 10 Feb 2012, Brad Aagaard wrote: > I have write access to the iit petsc-dev repo so that I can push fixes to > Matt's Sieve stuff. Matt had me switch over to the bitbucket repo and now I > don't have write access. I assume I need a bitbucket login, so I created one > (baagaard) and added my public ssh key. If one of you could enable write > access to the petsc-dev repo, I would appreciate it. > > Thanks, > Brad >

1 0

Re: [petsc-dev] write access to petsc-dev bitbucket repo
by Sean Farley 10 Feb '12

10 Feb '12

> > I have write access to the iit petsc-dev repo so that I can push fixes to > Matt's Sieve stuff. Matt had me switch over to the bitbucket repo and now I > don't have write access. I assume I need a bitbucket login, so I created > one (baagaard) and added my public ssh key. If one of you could enable > write access to the petsc-dev repo, I would appreciate it. OK, done.

1 0

write access to petsc-dev bitbucket repo
by Brad Aagaard 10 Feb '12

10 Feb '12

I have write access to the iit petsc-dev repo so that I can push fixes to Matt's Sieve stuff. Matt had me switch over to the bitbucket repo and now I don't have write access. I assume I need a bitbucket login, so I created one (baagaard) and added my public ssh key. If one of you could enable write access to the petsc-dev repo, I would appreciate it. Thanks, Brad

1 0

Re: [petsc-dev] A more positive thread
by Paul Mullowney 10 Feb '12

10 Feb '12

I'll write up something about the implementation this weekend and pass it on. The key is that everything is done asynchronously and that communication costs are pretty well hidden for large matrices. The code is in our add on package, txpetscgpu. If many people are working on this code, it may be worthwhile to move this code into the main PETSc source. The example I've been using to test is : src/ksp/ksp/examples/tutorials/ex10.c The matrices that I get 1) 40+ GFlops is af_shell4 from the UFL collection. 2) cfd.2.10 gets around 18 GFlops on 4 GPUs in CSR format. It gets 22 GFlops on 2 GPUs in ELL format. My system has 4 Nvidia 2070s attached to a single node. There are 2 nodes. -Paul > > On Fri, Feb 10, 2012 at 10:11 AM, Paul Mullowney <paulm(a)txcorp.com > <mailto:[email protected]>> wrote: > > Congrats! Given that I have a DoE phase II to continue Petsc GPU > development, we should probably all get on the same page. > > If you're looking to do multi-GPU computing, I've already > effectively redesigned the MatMult to get very good strong scaling > and some excellent performance. > I have some matrices getting 40+ GFlops on 4 GPUs. 20 GFlops on 2 > GPUs, ... > > > Great. Can you check in a benchmark code to src/benchmarks so that I > can run it and reproduce these results? It would > be great to start this sort of thing to make papers easier. I can help > with any problems doing that. > > I will look over the code so I get a better idea what is happening. > > Thanks, > > Matt > > > -Paul > > > Pramod (who did the work) was my student at EPCC last year. We > looked at porting Fluidity (or bits of it) to GPU and as a > result decided that working on PETSc's GPU support was going > to be the most promising angle for the time we had available. > We have never got round to benchmarking Fluidity with Pramod's > extended sparse matrix format support, but it shouldn't be too > much effort. > > If you have any questions about it all, feel free to ask. > > cheers, > Michele > On 10 Feb 2012, at 15:57, Matthew Knepley wrote: > > Has anyone taken a look at this: > > http://www.epcc.ed.ac.uk/wp-content/uploads/2011/11/PramodKumbhar.pdf > > It is the Fluidity people I think. I have just gotten an > NSF award with Dan Negrut > in Wisconsin and Ahmed Sameh at Purdue to port Ahmed's > SPIKE preconditioner > (you may have heard Olaf Schenk talk about this) to PETSc, > and in particular to > PETSc's GPU backend so that Dan can run it on his GPU > cluster. Thus, we will > be seriously stressing this feature in the near future. > > Matt > > -- > What most experimenters take for granted before they begin > their experiments is infinitely more interesting than any > results to which their experiments lead. > -- Norbert Wiener > > > > > > -- > What most experimenters take for granted before they begin their > experiments is infinitely more interesting than any results to which > their experiments lead. > -- Norbert Wiener

1 0

Re: [petsc-dev] petsc-dev on bitbucket
by Matthew Knepley 10 Feb '12

10 Feb '12

On Fri, Feb 10, 2012 at 10:08 AM, Satish Balay <balay(a)mcs.anl.gov> wrote: > On Fri, 10 Feb 2012, Matthew Knepley wrote: > > > > until we move to bitbucket - can you continue to push to petsc.cs.iit > > > - and not bitbucket? [for now - pushes to petsc.cs.iit are > > > automatically pushed to bitbucket] > > > > > > > I thought we were switched already :) I cannot understand why we would > not > > give up this maintenance burden (and directories is not a good enough > > reason). > > For my part - I generally don't like the way the change is pushed. > > - one prblem => suggest change to a new system [not evaluating all the > features we currently have vs what we loose] > I thought I did that in several mails. What do you think I have no addressed? > - and I really get annoyed when a problem is pointed out with the new > sytem - and the solution to that is: the current model is workng - but > antique - lets change it to fix for this new system we are pusing > By this you mean not having directories? I agree that is a shortcoming. My explicit comment was that this is far outweighed by the positives. Everything is a balance, and here I believe the balance is clearly on the side of BB. You disagree on this I see. > Wrt current admin overhead - its not that much. Most of the work is > adding new users [not modifying user keys]. This part of admin work > will is exist with bitbucket. > Every time the user changes keys, we have to do something. I cannot believe Barry is not screaming about this maintenance nightmare. > The other systrem admin part is not much. Jsut reboot the machine when > there is a kernel update. > We clearly had this situation this week, and "just reboot" sucks pretty badly. > Also the reliability aspect of such sites is also verblown. I'm sure > some of them had their outages. [maybe not bit bucket - but some of > them do]. Out outages are more frequent. There is no need to exaggerate. However, I listed several improvements in maintenance in addition to uptime. Matt > > Satish > -- What most experimenters take for granted before they begin their experiments is infinitely more interesting than any results to which their experiments lead. -- Norbert Wiener

1 0