I just pushed support for external repositories in ./sync-all, but didn't add async. It makes it easy to do so, however (and the patch is small and lightweight.) José, if async is all you're missing, I can add it to the 'extra' repositories, and then it should be reasonably straightforward to add it to nofib as an extra benchmark, I think. On Fri, Aug 16, 2013 at 1:26 PM, Austin Seipp wrote:

IMO, it's reasonable to allow this, but there's one minor sticky bit.

async's only dependency is stm, and it's also part of the platform, so I expect it will be relatively stable. In this case, perhaps we should just add 'async' to the set of 'extra' libraries for ./sync-all, which can be built with the compiler. Then, it should be easy to add tests for nofib (and even testsuite, if people find bugs.) stm is already one of the 'extra' libraries, and there are a few smp benchmarks that use it too, so this doesn't really change anything in that regard.

The main thing is that async isn't under our normal package structure, so we'll either need to A) mirror it, or B) we need to add support for ./sync-all to sync with an arbitrary HTTP url or something, and point it to Simon's repository as an extra package.

I'm in favor of 2 since then we don't have to maintain an unnecessary mirror, and also, because it might be useful later for similar things.

Thoughts?

On Fri, Aug 16, 2013 at 6:58 AM, Simon Peyton-Jones wrote:

...

(devs: this thread is about adding useful new benchmarks to nofib.)

Oh bother. I’d forgotten about dependencies. I don’t want to make building nofib depend on libraries other those in GHC anyway (bytestring, unix ok, asynch perhaps not). If that makes it tricky, maybe we should give up on the idea.

S

From: José Pedro Magalhães [mailto:jose.pedro.magalhaes@cs.ox.ac.uk] Sent: 05 August 2013 08:41 To: Simon Peyton-Jones Subject: Re: lambda mining

I'm not entirely sure how to do that, though. Do I just add it to the "real" subset? How about dependencies (e.g. bytestring >= 0.9, unix >= 2.5.0, async >= 2.0.0.0, ...)

Cheers, Pedro

On Fri, Aug 2, 2013 at 9:02 AM, Simon Peyton-Jones wrote:

great! Just add it :-)

simon

From: José Pedro Magalhães [mailto:jpm@cs.ox.ac.uk] Sent: 30 July 2013 07:48 To: Simon Peyton-Jones Cc: Nicolas Wu; Wouter Swierstra; Jeroen Bransen Subject: Re: lambda mining

Hi Simon,

(CC-ing co-authors)

Yes, I think it might work fine. Its running time can also be adjusted easily, depending on the maps given as input and some internal parameters. How would we go about adding it to nofib?

Thanks, Pedro

On Tue, Jul 30, 2013 at 7:04 AM, Simon Peyton-Jones wrote:

Pedro

Wandering past your home page I took a look at your “lambda mining” paper. Would it be suitable as a nofib benchmark? Moderate size, authentic code... Would you be interested?

Simon

_______________________________________________ ghc-devs mailing list ghc-devs@haskell.org http://www.haskell.org/mailman/listinfo/ghc-devs

-- Regards, Austin - PGP: 4096R/0x91384671

-- Regards, Austin - PGP: 4096R/0x91384671

I don't think it's worth dragging all that in for one benchmark. That said, Nofib is our main performance regression suite, so the more representative it is the better, and the less likely we are to accidently optimise for the wrong thing. If we had a bunch of new benchmark programs (which I would love) then it might well make sense to install more libraries with them -- if anyone feels like helping with that. Simon From: josepedromagalhaes@gmail.com [mailto:josepedromagalhaes@gmail.com] On Behalf Of José Pedro Magalhães Sent: 19 August 2013 14:17 To: Austin Seipp Cc: Simon Peyton-Jones; ghc-devs Subject: Re: lambda mining Hi, Here are the complete dependencies: Build-depends: base >= 4 && < 5, array >= 0.4, mtl >= 2.0.0.0, ansi-terminal, containers, old-time >= 1.1, directory >= 1.1, filepath >= 1.3, attoparsec >= 0.10, bytestring >= 0.9, parallel-io >= 0.3.2, hashable >= 1.1, unordered-containers >= 0.2, pqueue >= 1.2.0, unix >= 2.5.0, async >= 2.0.0.0, deepseq >= 1.1.0.0 Anything too problematic? Thanks, Pedro On Sun, Aug 18, 2013 at 6:48 PM, Austin Seipp mailto:aseipp@pobox.com> wrote: I just pushed support for external repositories in ./sync-all, but didn't add async. It makes it easy to do so, however (and the patch is small and lightweight.) José, if async is all you're missing, I can add it to the 'extra' repositories, and then it should be reasonably straightforward to add it to nofib as an extra benchmark, I think. On Fri, Aug 16, 2013 at 1:26 PM, Austin Seipp mailto:aseipp@pobox.com> wrote:

IMO, it's reasonable to allow this, but there's one minor sticky bit.

async's only dependency is stm, and it's also part of the platform, so I expect it will be relatively stable. In this case, perhaps we should just add 'async' to the set of 'extra' libraries for ./sync-all, which can be built with the compiler. Then, it should be easy to add tests for nofib (and even testsuite, if people find bugs.) stm is already one of the 'extra' libraries, and there are a few smp benchmarks that use it too, so this doesn't really change anything in that regard.

The main thing is that async isn't under our normal package structure, so we'll either need to A) mirror it, or B) we need to add support for ./sync-all to sync with an arbitrary HTTP url or something, and point it to Simon's repository as an extra package.

I'm in favor of 2 since then we don't have to maintain an unnecessary mirror, and also, because it might be useful later for similar things.

Thoughts?

On Fri, Aug 16, 2013 at 6:58 AM, Simon Peyton-Jones mailto:simonpj@microsoft.com> wrote:

...

(devs: this thread is about adding useful new benchmarks to nofib.)

Oh bother. I'd forgotten about dependencies. I don't want to make building nofib depend on libraries other those in GHC anyway (bytestring, unix ok, asynch perhaps not). If that makes it tricky, maybe we should give up on the idea.

S

From: José Pedro Magalhães [mailto:jose.pedro.magalhaes@cs.ox.ac.ukmailto:jose.pedro.magalhaes@cs.ox.ac.uk] Sent: 05 August 2013 08:41 To: Simon Peyton-Jones Subject: Re: lambda mining

I'm not entirely sure how to do that, though. Do I just add it to the "real" subset? How about dependencies (e.g. bytestring >= 0.9, unix >= 2.5.0, async >= 2.0.0.0, ...)

Cheers, Pedro

On Fri, Aug 2, 2013 at 9:02 AM, Simon Peyton-Jones mailto:simonpj@microsoft.com> wrote:

great! Just add it :-)

simon

From: José Pedro Magalhães [mailto:jpm@cs.ox.ac.ukmailto:jpm@cs.ox.ac.uk] Sent: 30 July 2013 07:48 To: Simon Peyton-Jones Cc: Nicolas Wu; Wouter Swierstra; Jeroen Bransen Subject: Re: lambda mining

Hi Simon,

(CC-ing co-authors)

Yes, I think it might work fine. Its running time can also be adjusted easily, depending on the maps given as input and some internal parameters. How would we go about adding it to nofib?

Thanks, Pedro

On Tue, Jul 30, 2013 at 7:04 AM, Simon Peyton-Jones mailto:simonpj@microsoft.com> wrote:

Pedro

Wandering past your home page I took a look at your "lambda mining" paper. Would it be suitable as a nofib benchmark? Moderate size, authentic code... Would you be interested?

Simon

_______________________________________________ ghc-devs mailing list ghc-devs@haskell.orgmailto:ghc-devs@haskell.org http://www.haskell.org/mailman/listinfo/ghc-devs

-- Regards, Austin - PGP: 4096R/0x91384671

-- Regards, Austin - PGP: 4096R/0x91384671