Home liberachat/#haskell: Logs Calendar

Logs: liberachat/#haskell

←Prev  Next→ 1,789,536 events total
2026-04-02 18:10:03 <haskellbridge> <sm> I've never used it, but that sounds like an excellent update
2026-04-02 18:11:10 humasect joins (~humasect@dyn-192-249-132-90.nexicom.net)
2026-04-02 18:12:37 × merijn quits (~merijn@host-cl.cgnat-g.v4.dfn.nl) (Ping timeout: 265 seconds)
2026-04-02 18:13:01 <[exa]> tomsmeding: cooooooool
2026-04-02 18:13:18 × target_i quits (~target_i@user/target-i/x-6023099) (Ping timeout: 246 seconds)
2026-04-02 18:14:18 target_i joins (~target_i@user/target-i/x-6023099)
2026-04-02 18:14:18 × haskellbridge quits (~hackager@96.28.224.214) (Read error: Connection reset by peer)
2026-04-02 18:14:44 × humasect quits (~humasect@dyn-192-249-132-90.nexicom.net) (Client Quit)
2026-04-02 18:16:35 <tomsmeding> Athas: that seems to be blocked only on a bounds bump in linear-accelerate; I've pinged the maintainer. :)
2026-04-02 18:19:40 haskellbridge joins (~hackager@96.28.224.214)
2026-04-02 18:19:40 ChanServ sets mode +v haskellbridge
2026-04-02 18:26:25 × Pozyomka quits (~pyon@user/pyon) (Ping timeout: 245 seconds)
2026-04-02 18:32:23 × somemathguy quits (~somemathg@user/somemathguy) (Quit: WeeChat 4.1.1)
2026-04-02 18:33:13 × arandombit quits (~arandombi@user/arandombit) (Ping timeout: 248 seconds)
2026-04-02 18:43:41 merijn joins (~merijn@host-cl.cgnat-g.v4.dfn.nl)
2026-04-02 18:48:18 × merijn quits (~merijn@host-cl.cgnat-g.v4.dfn.nl) (Ping timeout: 248 seconds)
2026-04-02 18:48:50 × alter2000 quits (~alter2000@user/alter2000) (Ping timeout: 248 seconds)
2026-04-02 18:50:26 alter2000 joins (~alter2000@user/alter2000)
2026-04-02 18:54:14 Pozyomka joins (~pyon@user/pyon)
2026-04-02 18:55:28 tessier_ joins (~tessier@ip68-8-117-219.sd.sd.cox.net)
2026-04-02 18:57:07 × tessier quits (~tessier@ec2-184-72-149-67.compute-1.amazonaws.com) (Ping timeout: 276 seconds)
2026-04-02 18:59:05 merijn joins (~merijn@host-cl.cgnat-g.v4.dfn.nl)
2026-04-02 19:03:24 arandombit joins (~arandombi@2a02:2455:8656:7100:a19d:edd6:ba09:3a91)
2026-04-02 19:03:24 × arandombit quits (~arandombi@2a02:2455:8656:7100:a19d:edd6:ba09:3a91) (Changing host)
2026-04-02 19:03:24 arandombit joins (~arandombi@user/arandombit)
2026-04-02 19:04:11 <Athas> tomsmeding: will it be five years untli the next update?
2026-04-02 19:04:50 <tomsmeding> I don't know!
2026-04-02 19:05:02 <tomsmeding> I hope not, but that's out of my hands :)
2026-04-02 19:05:35 <tomsmeding> This was my last act of vengeance here in Utrecht before I leave for a postdoc :)
2026-04-02 19:05:54 × merijn quits (~merijn@host-cl.cgnat-g.v4.dfn.nl) (Ping timeout: 248 seconds)
2026-04-02 19:09:08 <Athas> A productive end.
2026-04-02 19:11:02 <tomsmeding> At least, Ivo's intent is to have a release earlier than five years from now, but then, previously there was also an intent.
2026-04-02 19:11:19 <davean> tomsmeding: haha, why has it failed to get updates? Its lack of maintanership really hurts it.
2026-04-02 19:12:27 ss4 joins (~wootehfoo@user/wootehfoot)
2026-04-02 19:12:30 × ss4 quits (~wootehfoo@user/wootehfoot) (Remote host closed the connection)
2026-04-02 19:12:47 wootehfoot joins (~wootehfoo@user/wootehfoot)
2026-04-02 19:12:52 <tomsmeding> davean: the people "connected" with accelerate fall into 3 categories: 1. moved on (other job, other interests, etc.); 2. is a full professor and is busy; 3. is a PhD student and is busy
2026-04-02 19:13:32 <tomsmeding> I wasn't even officially connected to accelerate, really, but I dragged myself into it (for better or for worse), and am now transitioning from 3 to 1
2026-04-02 19:14:03 <davean> Well what is your postdoc?
2026-04-02 19:14:05 <EvanR> someone write "AI" somewhere on accelerate to get more funding and support
2026-04-02 19:14:10 merijn joins (~merijn@host-cl.cgnat-g.v4.dfn.nl)
2026-04-02 19:14:19 <tomsmeding> currently forming is a new category containing 1 person who is not moving on and yet not a PhD student any more. We'll see how that category develops
2026-04-02 19:14:20 <Athas> Also, Accelerate comprises many packages, so release engineering is harder than a tag (I suppose).
2026-04-02 19:15:00 <tomsmeding> davean: I something in an area that could involve Accelerate, but likely will not for practical reasons
2026-04-02 19:15:27 <tomsmeding> Athas: kinda yes, but also what's mostly slow is getting everyone to respond to an email.
2026-04-02 19:15:43 <tomsmeding> also fixing all the bugs in the GPU backend. That took us a year or so
2026-04-02 19:15:52 <tomsmeding> (There's probably more!)
2026-04-02 19:16:12 <Athas> I have heard that some released software has bugs, so maybe fixing them all is not necessary before tagging a release.
2026-04-02 19:16:42 <tomsmeding> Athas: having the thing segfault when running a program with more than a few kernels is not something that I'd want to release on my name
2026-04-02 19:16:58 <davean> Athas: People who release buggy software should be ashamed
2026-04-02 19:17:11 <davean> Its kinda a plauge of this era that people really are so sloppy with the software they release.
2026-04-02 19:17:12 <tomsmeding> (and I was the only one willing to do the work to make an actual release, so the known-buggy release indeed didn't happen)
2026-04-02 19:18:37 <tomsmeding> davean: if you replace "buggy" with "known-buggy", then I can get on board
2026-04-02 19:18:45 × merijn quits (~merijn@host-cl.cgnat-g.v4.dfn.nl) (Ping timeout: 244 seconds)
2026-04-02 19:18:47 <EvanR> just don't make mistakes smh
2026-04-02 19:18:57 <davean> tomsmeding: the problem with that is people aren't putting in the effort to look much lately.
2026-04-02 19:18:57 <tomsmeding> well it's haskell already, so what can go wrong
2026-04-02 19:19:07 <EvanR> this tbh
2026-04-02 19:19:10 <EvanR> xD
2026-04-02 19:19:15 <tomsmeding> davean: if you haven't made an effort to look, it's kinda known-buggy, isn't it?
2026-04-02 19:19:18 <davean> I'd have said "known-buggy" even 15 years ago.
2026-04-02 19:19:26 <davean> tomsmeding: They sure don't think so1
2026-04-02 19:19:33 <tomsmeding> :p
2026-04-02 19:20:05 <davean> If we can agree on a reasonably extensive QA process I'm happy to only say known bugs.
2026-04-02 19:20:19 <tomsmeding> EvanR: tangentially, turns out that "when it compiles it works" comes crashing down HARD once you do weird stuff with a GPU driver
2026-04-02 19:20:21 <davean> But I know a lot of people who ship it when it runs at all and think its fine.
2026-04-02 19:21:33 <davean> Its a bit harder to test GPU compilers, esp since a lot of what you're debugging is the GPU stack
2026-04-02 19:22:42 <EvanR> I blame the compiler in the driver
2026-04-02 19:24:15 × alter2000 quits (~alter2000@user/alter2000) (Ping timeout: 272 seconds)
2026-04-02 19:24:27 <davean> I probably am impacted by more software bugs each month now than I dealt with in the entir 2000-2005 period.
2026-04-02 19:24:55 × tusko quits (~uwu@user/tusko) (Remote host closed the connection)
2026-04-02 19:25:01 <davean> Software was not good then, not at all, but its complexity vs. the effort put in was a bit more in balance.
2026-04-02 19:25:14 tusko joins (~uwu@user/tusko)
2026-04-02 19:25:16 <tomsmeding> davean: I can say with certainty that accelerate's current testing regime is beneath your standards, but at least it's slightly better than it was a few years ago
2026-04-02 19:25:16 <davean> I was impacted a lot more by it just not having features.
2026-04-02 19:25:48 <davean> tomsmeding: I kinda assume for accelerate you need a lot of users to actually start testing well because of the complexity of the stack you sit on top of.
2026-04-02 19:26:01 alter2000 joins (~alter2000@user/alter2000)
2026-04-02 19:26:02 <davean> the problem with that is it so often is broken no one can rely on it to start using it,.
2026-04-02 19:26:13 <tomsmeding> right
2026-04-02 19:26:22 <tomsmeding> well, I think it works well enough now for people to start testing it :)
2026-04-02 19:26:39 <davean> Sure, the problem is will it work in a year? :)
2026-04-02 19:26:40 <tomsmeding> the CPU backend worked well enough for a long time already, by the way; it was the GPU backend that was unstable
2026-04-02 19:26:46 <tomsmeding> I dunno!
2026-04-02 19:27:18 <tomsmeding> Mostly depends on how backwards-compatible LLVM's IR format is, and how backwards-compatible CUDA turns out to be next year round
2026-04-02 19:27:31 <davean> I get hesitant every time I consider designing something that relies on accelerate for this reason :)
2026-04-02 19:27:40 <tomsmeding> the former seems to be fine so far
2026-04-02 19:28:08 <davean> Hum, so CUDA software I have from 2008 still works fine.
2026-04-02 19:28:13 <davean> There is something more
2026-04-02 19:28:39 <davean> (I recently found out, because someone got me to fix the compile on an ancient game project)
2026-04-02 19:28:40 <tomsmeding> if you try to bind CUDA functions from Haskell, you're more exposed to the function name shuffling they do under the hood
2026-04-02 19:28:58 <tomsmeding> doing this (instead of writing the CUDA glue code in C) was probably a mistake, and one that is in the plans of being fixed
2026-04-02 19:29:30 <davean> Ah yah if you're trying to hook it yourself instead of use the compiler which is what their spec actually is ... good luck!
2026-04-02 19:29:32 <tomsmeding> davean: and very old stuff is more likely to still work than the fancy thing from 2 years ago
2026-04-02 19:29:33 merijn joins (~merijn@host-cl.cgnat-g.v4.dfn.nl)
2026-04-02 19:29:36 <davean> and I don't mean that encuragingly.
2026-04-02 19:29:41 <tomsmeding> yeah
2026-04-02 19:30:23 <tomsmeding> in any case, the people hacking on accelerate are aware of all this, but there's little incentive in the academic system to spend the hours to fix it
2026-04-02 19:33:48 × merijn quits (~merijn@host-cl.cgnat-g.v4.dfn.nl) (Ping timeout: 246 seconds)
2026-04-02 19:35:23 <davean> tomsmeding: does it have an AMD backend these days that works?
2026-04-02 19:35:33 <tomsmeding> there's plans for one

All times are in UTC.