GHC build time graphs - ghc-devs - Haskell.org

newer
Potentially confusing syntax for...

GHC build time graphs

older
GHC performance dashboard question

Edward Z. Yang

28 Jan 2016 28 Jan '16

1:30 p.m.

Joachim, Anthony Cowley was recently complaining on Twitter that building GHC takes 39.5% longer now than it did for 7.10.1: https://twitter.com/a_cowley/status/692481516744634368 Do we have graphs for these metrics, e.g. can we see the breakdown of which commits increased the build time? Thanks, Edward

Reply

Sign in to reply online Use email software

Show replies by date

Joachim Breitner

28 Jan 28 Jan

5:16 p.m.

Hi, Am Donnerstag, den 28.01.2016, 10:30 -0800 schrieb Edward Z. Yang:

Anthony Cowley was recently complaining on Twitter that building GHC takes 39.5% longer now than it did for 7.10.1: https://twitter.com/a_cowley/status/692481516744634368

Do we have graphs for these metrics, e.g. can we see the breakdown of which commits increased the build time?

well, we have https://perf.haskell.org/ghc/#graph/buildtime/make In theory, there would be a button to expand the graph on the left. In practice, that is not yet implemented... See https://github.com/nomeata/gipeda/issues/14 Greetings, Joachim -- Joachim “nomeata” Breitner mail@joachim-breitner.de • http://www.joachim-breitner.de/ Jabber: nomeata@joachim-breitner.de • GPG-Key: 0xF0FBF51F Debian Developer: nomeata@debian.org

Reply

Sign in to reply online Use email software

Oleg Grenrus

5:22 p.m.

Is the same compiler used to build HEAD and 7.10,1? - Oleg

On 28 Jan 2016, at 20:30, Edward Z. Yang wrote:

Joachim,

Anthony Cowley was recently complaining on Twitter that building GHC takes 39.5% longer now than it did for 7.10.1: https://twitter.com/a_cowley/status/692481516744634368

Do we have graphs for these metrics, e.g. can we see the breakdown of which commits increased the build time?

Thanks, Edward _______________________________________________ ghc-devs mailing list ghc-devs@haskell.org http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs

Reply

Sign in to reply online Use email software

Joachim Breitner

5:26 p.m.

Hi Oleg, Am Freitag, den 29.01.2016, 00:22 +0200 schrieb Oleg Grenrus:

Is the same compiler used to build HEAD and 7.10,1?

Good call. In fact, no: 7.10.1 is built with 7.6.3, while HEAD is built with 7.10.3. Anthony’s link, i.e. https://perf.haskell.org/ghc/#compare/ca00def1d7093d6b5b2a937ddfc8a01c152038... has links to the build logs of either build; there I could find that information. That might be (part) of the problem. But if it is, it is even worse, as it would mean not only building the compiler got slower, but the compiler itself... Greetings, Joachim -- Joachim “nomeata” Breitner mail@joachim-breitner.de • http://www.joachim-breitner.de/ Jabber: nomeata@joachim-breitner.de • GPG-Key: 0xF0FBF51F Debian Developer: nomeata@debian.org

Reply

Sign in to reply online Use email software

Ben Gamari

5:34 p.m.

Joachim Breitner writes:

Hi Oleg,

Am Freitag, den 29.01.2016, 00:22 +0200 schrieb Oleg Grenrus:

...
Is the same compiler used to build HEAD and 7.10,1?

Good call. In fact, no: 7.10.1 is built with 7.6.3, while HEAD is built with 7.10.3.

Anthony’s link, i.e. https://perf.haskell.org/ghc/#compare/ca00def1d7093d6b5b2a937ddfc8a01c152038... has links to the build logs of either build; there I could find that information.

That might be (part) of the problem. But if it is, it is even worse, as it would mean not only building the compiler got slower, but the compiler itself...

I can verify that the build itself is indeed slower. Validating the current state of ghc-7.10 takes 19 minutes, whereas ghc-8.0 takes 25.5 minutes. This isn't entirely unexpected but the change is quite a bit larger than I had thought. It would be nice to know which commits are responsible. Cheers, - Ben

Reply

Sign in to reply online Use email software

Karel Gardas

9 Feb 9 Feb

11:20 a.m.

On 01/28/16 11:34 PM, Ben Gamari wrote:

Joachim Breitner writes:

...
Hi Oleg,

Am Freitag, den 29.01.2016, 00:22 +0200 schrieb Oleg Grenrus:

...
Is the same compiler used to build HEAD and 7.10,1?

Good call. In fact, no: 7.10.1 is built with 7.6.3, while HEAD is built with 7.10.3.

Anthony’s link, i.e. https://perf.haskell.org/ghc/#compare/ca00def1d7093d6b5b2a937ddfc8a01c152038... has links to the build logs of either build; there I could find that information.

That might be (part) of the problem. But if it is, it is even worse, as it would mean not only building the compiler got slower, but the compiler itself...

I can verify that the build itself is indeed slower. Validating the current state of ghc-7.10 takes 19 minutes, whereas ghc-8.0 takes 25.5 minutes. This isn't entirely unexpected but the change is quite a bit larger than I had thought. It would be nice to know which commits are responsible.

btw, just recent experience on ARM64 (X-gene board): bootstrapping 7.10.1 with 7.6.x took: ~120 minutes bootstrapping 8.0.1 RC2 with 7.10.1 took: ~446 minutes both run as: ./configure; time make -j8 although board is shared I've tried to look at it from time to time and have not observed anyone else there so I can assume this is indeed GHC slowness on unregisterised -fvia-C ports... Karel

Reply

Sign in to reply online Use email software

Reid Barton

12 Feb 12 Feb

3:02 p.m.

On Tue, Feb 9, 2016 at 11:20 AM, Karel Gardas wrote:

On 01/28/16 11:34 PM, Ben Gamari wrote:

...
Joachim Breitner writes:

Hi Oleg,

...
Am Freitag, den 29.01.2016, 00:22 +0200 schrieb Oleg Grenrus:

...
Is the same compiler used to build HEAD and 7.10,1?

Good call. In fact, no: 7.10.1 is built with 7.6.3, while HEAD is built with 7.10.3.

Anthony’s link, i.e.

https://perf.haskell.org/ghc/#compare/ca00def1d7093d6b5b2a937ddfc8a01c152038... has links to the build logs of either build; there I could find that information.

That might be (part) of the problem. But if it is, it is even worse, as it would mean not only building the compiler got slower, but the compiler itself...

I can verify that the build itself is indeed slower. Validating the

current state of ghc-7.10 takes 19 minutes, whereas ghc-8.0 takes 25.5 minutes. This isn't entirely unexpected but the change is quite a bit larger than I had thought. It would be nice to know which commits are responsible.

btw, just recent experience on ARM64 (X-gene board):

bootstrapping 7.10.1 with 7.6.x took: ~120 minutes bootstrapping 8.0.1 RC2 with 7.10.1 took: ~446 minutes

both run as: ./configure; time make -j8

It would be interesting to have the time for bootstrapping 7.10.1 with 7.10.1 too, for comparison. Regards, Reid Barton

Reply

Sign in to reply online Use email software

Karel Gardas

4:43 p.m.

On 02/12/16 09:02 PM, Reid Barton wrote:

btw, just recent experience on ARM64 (X-gene board):

bootstrapping 7.10.1 with 7.6.x took: ~120 minutes bootstrapping 8.0.1 RC2 with 7.10.1 took: ~446 minutes

both run as: ./configure; time make -j8

It would be interesting to have the time for bootstrapping 7.10.1 with 7.10.1 too, for comparison.

Good idea, build is running, but is sluggish just from the first sight. Running already for 1h20m and still building stage1. This is more in line with 8.0.1 I'm afraid. Anyway, I'll update with proper time once it is at the end. Looks like I'll also need to do 7.8.x build to see where we are on this... Karel

Reply

Sign in to reply online Use email software

Karel Gardas

13 Feb 13 Feb

1:57 p.m.

On 02/12/16 09:02 PM, Reid Barton wrote:

btw, just recent experience on ARM64 (X-gene board):

bootstrapping 7.10.1 with 7.6.x took: ~120 minutes bootstrapping 8.0.1 RC2 with 7.10.1 took: ~446 minutes

both run as: ./configure; time make -j8

It would be interesting to have the time for bootstrapping 7.10.1 with 7.10.1 too, for comparison.

boostrapping 7.10.1 with 7.10.1 took: ~212 minutes -- but this is not comparable directly with numbers above since for whatever reason haddock and doc is not build although I've used previous ./configure; time make -j8. Anyway, parallel make itself is not good to use here since the build process may also contain some noise of paralellel make fixes and such. I think more reliable benchmark may be simple build (single-threaded) of ghc-cabal here. I'll see what I can do about that. Cheers, Karel

Reply

Sign in to reply online Use email software

Karel Gardas

6:10 p.m.

On 02/13/16 07:57 PM, Karel Gardas wrote:

On 02/12/16 09:02 PM, Reid Barton wrote:

...
btw, just recent experience on ARM64 (X-gene board):

bootstrapping 7.10.1 with 7.6.x took: ~120 minutes bootstrapping 8.0.1 RC2 with 7.10.1 took: ~446 minutes

both run as: ./configure; time make -j8

It would be interesting to have the time for bootstrapping 7.10.1 with 7.10.1 too, for comparison.

boostrapping 7.10.1 with 7.10.1 took: ~212 minutes -- but this is not comparable directly with numbers above since for whatever reason haddock and doc is not build although I've used previous ./configure; time make -j8.

Anyway, parallel make itself is not good to use here since the build process may also contain some noise of paralellel make fixes and such.

I think more reliable benchmark may be simple build (single-threaded) of ghc-cabal here. I'll see what I can do about that.

OK, so not everything is that bad. Benchmarking compilation of GHC 7.10.1 ghc-cabal tool reveals: GHC 7.6.3: ~12 minutes GHC 7.10.1: ~24 minutes GHC 8.0.1 RC2: ~14 minutes Please note that GHC 7.6.3 compiles 89 files while GHC 7.10.1 and 8.0.1 RC2 compile 90 files. Test done on the same x-gene board like the tests above for the reference... Karel

Reply

Sign in to reply online Use email software

3384

Age (days ago)

3400

Last active (days ago)

Download

9 comments

6 participants

tags

participants (6)

Ben Gamari
Edward Z. Yang
Joachim Breitner
Karel Gardas
Oleg Grenrus
Reid Barton