Newest at the top
| 2024-11-17 18:23:12 +0100 | <jonkri> | Hi! I'm looking into using GHC.Iface.Load to access the API of packages programmatically. Is there a straightforward way of getting a GHC.Driver.Session.Settings record? Thanks! |
| 2024-11-17 18:18:53 +0100 | <briandaed> | FreeBSD caesar 14.1-RELEASE-p5 FreeBSD 14.1-RELEASE-p5 GENERIC amd64 |
| 2024-11-17 18:18:13 +0100 | <briandaed> | 17m and still going, but got some errors, flags and some atomic operations changed for gcc 4.4 |
| 2024-11-17 18:17:43 +0100 | euleritian | (~euleritia@77.22.252.159) |
| 2024-11-17 18:16:43 +0100 | euleritian | (~euleritia@77.22.252.159) (Read error: Connection reset by peer) |
| 2024-11-17 18:16:43 +0100 | <Hecate> | Build completed in 19m37s on my FreeBSD box with Intel i7-7700 (8 thread) @ 3.600GHz |
| 2024-11-17 18:10:51 +0100 | target_i | (~target_i@user/target-i/x-6023099) (Quit: leaving) |
| 2024-11-17 18:04:38 +0100 | Everything | (~Everythin@178-133-181-60.mobile.vf-ua.net) (Quit: leaving) |
| 2024-11-17 18:03:55 +0100 | machinedgod | (~machinedg@d108-173-18-100.abhsia.telus.net) machinedgod |
| 2024-11-17 18:02:38 +0100 | <Hecate> | :D |
| 2024-11-17 18:02:26 +0100 | <briandaed> | -j killed my junk |
| 2024-11-17 18:01:25 +0100 | jonkri | (~jonkri@90-230-234-124-no600.tbcn.telia.com) jonkri |
| 2024-11-17 18:01:14 +0100 | <briandaed> | looks like new aphex twin track |
| 2024-11-17 17:59:43 +0100 | <Hecate> | noice |
| 2024-11-17 17:59:42 +0100 | <Hecate> | https://i.imgur.com/AcH1Rir.png |
| 2024-11-17 17:57:43 +0100 | <shapr> | ooh, I want to run the benchmark on my laptop |
| 2024-11-17 17:56:49 +0100 | <briandaed> | again agree, some ancient results https://qbaylogic.github.io/benchmark-compilation/ which I believe still make sense |
| 2024-11-17 17:54:36 +0100 | <Hecate> | erm, s/cores/threads/ |
| 2024-11-17 17:54:25 +0100 | <Hecate> | briandaed: there's a couple of rules indeed. One of them being: Plan to use 1GB of RAM for each thread of compilation. If you have more cores than gigabytes of RAM, you *must* restrict the amount of cores used |
| 2024-11-17 17:53:50 +0100 | <briandaed> | wouldn't bet on memory latency, rather on being computing intensive (cpu-bound) |
| 2024-11-17 17:51:17 +0100 | <shapr> | Is that for lower memory latency? |
| 2024-11-17 17:50:46 +0100 | <briandaed> | I agree, and looks high frequency is more important than core count |
| 2024-11-17 17:50:16 +0100 | <Hecate> | or Firefox |
| 2024-11-17 17:50:13 +0100 | <Hecate> | big projects like GHC and LLVM |
| 2024-11-17 17:50:09 +0100 | <Hecate> | laptops are not good compilation machines for big projects anyway, briandaed. They often have not-so-good air flow, which is quite critical |
| 2024-11-17 17:49:29 +0100 | <briandaed> | nice, I gave up after 0.5h and some errors, and looks it's time for change |
| 2024-11-17 17:48:18 +0100 | <Hecate> | built in 11m46s on my Fedora box with a Ryzen 5 7600X (12 threads) @ 5.453GHz |
| 2024-11-17 17:45:04 +0100 | gorignak | (~gorignak@user/gorignak) gorignak |
| 2024-11-17 17:44:33 +0100 | gorignak | (~gorignak@user/gorignak) (Quit: quit) |
| 2024-11-17 17:42:05 +0100 | <briandaed> | maerwald, Hecate: thanks and thanks |
| 2024-11-17 17:40:29 +0100 | acidjnk_new | (~acidjnk@p200300d6e7283f99b0ad6825ae812995.dip0.t-ipconnect.de) acidjnk |
| 2024-11-17 17:40:28 +0100 | <haskellbridge> | <maerwald> It's unclear why I only hit it in 9.6 though |
| 2024-11-17 17:40:26 +0100 | <Hecate> | with GHC 9.10.1 as the boot GHC |
| 2024-11-17 17:40:17 +0100 | <Hecate> | briandaed: |
| 2024-11-17 17:40:16 +0100 | <Hecate> | ./boot && ./configure && ./hadrian/build -j --flavour=Quick |
| 2024-11-17 17:39:17 +0100 | <haskellbridge> | <maerwald> An RTS bug with kqueue |
| 2024-11-17 17:38:54 +0100 | <haskellbridge> | <maerwald> https://gitlab.haskell.org/ghc/ghc/-/issues/24672 |
| 2024-11-17 17:37:43 +0100 | <briandaed> | oops |
| 2024-11-17 17:37:38 +0100 | <haskellbridge> | <maerwald> briandaed: 9.6.6 is broken |
| 2024-11-17 17:36:35 +0100 | <briandaed> | problems... only until recently something with messed flags, with 9.2.7 installed via ghcup, today switched to 9.6.6 and so far so good |
| 2024-11-17 17:35:55 +0100 | <briandaed> | Hecate please also share commands you issue to build it, I followed something found on ghc-devs not sure if it's optimal what I've found |
| 2024-11-17 17:34:32 +0100 | <Hecate> | briandaed: drop me a line at hecate@haskell.foundation if you ever find problems on FreeBSD, I try to help on this front |
| 2024-11-17 17:33:07 +0100 | <briandaed> | hello :D |
| 2024-11-17 17:32:49 +0100 | <Hecate> | briandaed: also hello fellow FreeBSD user! :D |
| 2024-11-17 17:32:39 +0100 | <Hecate> | I can test on my FreeBSD box |
| 2024-11-17 17:32:32 +0100 | <Hecate> | I doubt it |
| 2024-11-17 17:31:59 +0100 | <briandaed> | yeah... looks it's a time to change laptop, 28 minutes here, not sure if compiling on FreeBSD make it slower... |
| 2024-11-17 17:31:45 +0100 | <Hecate> | let me see on mine |
| 2024-11-17 17:31:40 +0100 | <geekosaur> | as such I rarely give it more than 4 |
| 2024-11-17 17:31:31 +0100 | <geekosaur> | it doesn't use many cores; ghc is kinda poor at parallelized compiles |
