| *** tpb has joined #yosys | 00:00 | |
| *** emeb has quit IRC | 00:15 | |
| *** emeb_mac has joined #yosys | 00:23 | |
| *** N2TOH_ is now known as N2TOH | 00:23 | |
| *** BinaryLust has joined #yosys | 00:47 | |
| *** emeb_mac has quit IRC | 01:14 | |
| *** emeb_mac has joined #yosys | 01:17 | |
| *** Degi has quit IRC | 03:01 | |
| *** Degi has joined #yosys | 03:02 | |
| *** rlee287 has quit IRC | 03:12 | |
| *** bouliiii has quit IRC | 03:19 | |
| *** heijligen has quit IRC | 04:23 | |
| *** heijligen has joined #yosys | 04:24 | |
| cr1901_modern | daveshah: I made a mistake with my PR. Quelle surprise. https://github.com/SymbiFlow/prjtrellis/pull/142 | 04:31 |
|---|---|---|
| tpb | Title: Update environment.sh by cr1901 · Pull Request #142 · SymbiFlow/prjtrellis · GitHub (at github.com) | 04:31 |
| *** dys has quit IRC | 06:05 | |
| *** citypw has joined #yosys | 06:24 | |
| *** parport0 has quit IRC | 06:26 | |
| *** emeb_mac has quit IRC | 06:29 | |
| *** parport0 has joined #yosys | 06:29 | |
| *** dys has joined #yosys | 06:55 | |
| *** jakobwenzel has joined #yosys | 07:27 | |
| *** bouliiii has joined #yosys | 07:58 | |
| *** kraiskil has joined #yosys | 07:59 | |
| *** Asu has joined #yosys | 08:18 | |
| *** jakobwenzel has quit IRC | 08:53 | |
| *** BinaryLust has quit IRC | 09:03 | |
| *** vidbina has joined #yosys | 09:09 | |
| *** bouliiii has quit IRC | 09:14 | |
| *** jakobwenzel has joined #yosys | 09:17 | |
| *** vidbina has quit IRC | 09:31 | |
| *** kgugala has quit IRC | 10:19 | |
| *** kgugala_ has joined #yosys | 10:20 | |
| *** vidbina has joined #yosys | 11:47 | |
| *** az0re has quit IRC | 11:52 | |
| ZirconiumX | echo 'REEBE: NOP qverpgbel vf n ut jbexvat pbcl! Erzbir nop/ naq er-eha "znxr".' | tr 'A-Za-z' 'N-ZA-Mn-za-m' | 12:06 |
| ZirconiumX | Claire has a sense of humour, it seems | 12:06 |
| daveshah | This is actually for a good reason | 12:07 |
| daveshah | It's so "ERROR" doesn't appear when make echos the command, unless there actually is an error | 12:07 |
| ZirconiumX | Mmm | 12:07 |
| ZirconiumX | But you can equally get Make to not echo the command, so | 12:08 |
| daveshah | I remember this being discussed before, but I can't remember where | 12:08 |
| ZirconiumX | (@ prefix) | 12:09 |
| daveshah | Yosys uses this, but it is disabled in verbose mode, iirc | 12:09 |
| *** rswarbrick has joined #yosys | 12:14 | |
| *** X-Scale` has joined #yosys | 12:18 | |
| *** anticw has quit IRC | 12:18 | |
| *** X-Scale has quit IRC | 12:18 | |
| *** X-Scale` is now known as X-Scale | 12:19 | |
| *** anticw has joined #yosys | 12:19 | |
| *** Asu has quit IRC | 12:43 | |
| *** Asu has joined #yosys | 12:43 | |
| *** Asu has quit IRC | 12:51 | |
| *** Asu has joined #yosys | 12:52 | |
| *** anticw has quit IRC | 12:55 | |
| *** anticw has joined #yosys | 12:55 | |
| *** jakobwenzel has quit IRC | 13:05 | |
| *** anticw has quit IRC | 13:07 | |
| *** anticw has joined #yosys | 13:08 | |
| *** rswarbrick has quit IRC | 13:22 | |
| *** citypw has quit IRC | 13:47 | |
| *** emeb has joined #yosys | 13:56 | |
| *** anticw has quit IRC | 13:56 | |
| *** anticw has joined #yosys | 13:57 | |
| *** rswarbrick has joined #yosys | 14:17 | |
| *** jfcaron has joined #yosys | 14:36 | |
| *** anticw has quit IRC | 14:41 | |
| *** anticw has joined #yosys | 14:42 | |
| rswarbrick | ZirconiumX: Hi there! Can we follow up quickly on PR #2083 here? | 14:44 |
| ZirconiumX | Sure | 14:45 |
| rswarbrick | Can you explain exactly what you measured? Those numbers definitely look worse with -DYOSYS_NO_IDS_REFCNT, but that seems slightly surprising to me (with or without the patch in the PR). | 14:46 |
| ZirconiumX | Execution time | 14:47 |
| ZirconiumX | Unfortunately I'm under WSL, so I can't run it under perf | 14:47 |
| ZirconiumX | (otherwise I would) | 14:47 |
| rswarbrick | That's not what I meant. Did you compile the same code with and without the flag? And was that with or without the patch in the PR? | 14:47 |
| ZirconiumX | Same code with and without the flag | 14:48 |
| ZirconiumX | Didn't touch the patch, since I had a hunch it wasn't necessary | 14:48 |
| ZirconiumX | Number of instructions has little correlation with speed | 14:48 |
| rswarbrick | Ok, so I think you've just proved that this benchmark isn't affected by time spent refcounting. | 14:48 |
| rswarbrick | Re # insns / speed: Yes, until you get an icache miss. yosys is 8MB, so this is actually relevant, I thin | 14:49 |
| rswarbrick | *think | 14:49 |
| ZirconiumX | ...no | 14:49 |
| rswarbrick | Can you explain what the "...no" is in answer to? | 14:50 |
| ZirconiumX | If instruction count had any relation to speed then Rust code would not be competitive with C, for example | 14:50 |
| ZirconiumX | But because we live in a world with superscalar out of order execution, giant caches and latency hiding, I really don't see it as a problem | 14:51 |
| *** anticw has quit IRC | 14:51 | |
| ZirconiumX | The entirety of the Yosys executable could fit in my processor's cache | 14:51 |
| ZirconiumX | And it won't all be used at once | 14:51 |
| *** anticw has joined #yosys | 14:52 | |
| rswarbrick | Oh, so you have an 8MB L1 icache and then a separate dcache? That's quite a bit bigger than my laptop. | 14:52 |
| ZirconiumX | It doesn't need to all fit in L1 | 14:52 |
| rswarbrick | ... but if it doesn't then you get cache misses. Which cost time. | 14:52 |
| ZirconiumX | I feel like you're arguing on what you *think* the problem is | 14:53 |
| rswarbrick | I think you're arguing about what you think I'm trying to improve :-D | 14:53 |
| ZirconiumX | The SNES core is a pretty major workout for Yosys | 14:54 |
| ZirconiumX | At 19.4k LUT4s | 14:54 |
| ZirconiumX | If reference counting had any significant effect on the performance of Yosys, that would surely amplify it | 14:55 |
| ZirconiumX | But nothing changes | 14:55 |
| rswarbrick | Wait a sec... | 14:55 |
| ZirconiumX | 27.2k cells, all of which need names (strings!) | 14:55 |
| rswarbrick | My motivation here is that the code does less stuff with the patch. This may or may not improve wall-clock times today, but it definitely won't hurt, and makes things simpler over time. You're right that wall-clock times are the right thing to balance against "should we make the code more complicated". | 14:55 |
| ZirconiumX | Regardless of my own feelings on the patch, it'll still get vetoed by Claire | 14:56 |
| rswarbrick | I would argue that your benchmarking shows that we should never use YOSYS_NO_IDS_REFCNT (for safety, since the performance cost is tiny) and that we may as well pass the smart pointers by reference to avoid +1 -1 round-trips. | 14:56 |
| rswarbrick | Ok. In which case, I hope she can weigh in on what she thinks the right approach is between do nothing / do something clever. I agree that the patch at the PR at the moment is neither of the two! | 14:57 |
| ZirconiumX | If it was my own code, I would probably merge your patch, tbh | 14:57 |
| rswarbrick | Thanks for your vote of confidence on the PR :-p | 14:58 |
| ZirconiumX | Though I'd be even more in favour of "use less IdStrings" over this | 15:00 |
| *** anticw has quit IRC | 15:01 | |
| *** anticw has joined #yosys | 15:01 | |
| rswarbrick | Yeah. I guess I understand the value of interning stuff into a symbol table, though, so I'm not completely sure. | 15:02 |
| ZirconiumX | The main aim is to speed up equality tests | 15:02 |
| rswarbrick | I've worked on a compiler in the past: I do understand what a symbol table does! :-) | 15:03 |
| rswarbrick | You also get a total ordering on symbols (operator<), which can be handy. | 15:03 |
| rswarbrick | e.g. you can compare sets of symbols for equality by first sorting (by the cheap integer comparison) and then ticking them off one-by-one. | 15:04 |
| rswarbrick | In a very different world, Maxima (a computer algebra system) uses this as a trick for efficiently representing multivariate polynomials! | 15:04 |
| ZirconiumX | In my own world - chess - we use sequential probability ratio tests to decide if something is better or not, but I haven't quite got the infrastructure to do that for FPGAs | 15:06 |
| rswarbrick | Hmm, that's a new phrase to me. /me goes googling... | 15:08 |
| sorear | oh you worked on maxima? need | 15:08 |
| rswarbrick | sorear: Not recently, but I spent quite a bit of time on it 10 years ago or so. | 15:09 |
| *** anticw has quit IRC | 15:46 | |
| *** anticw has joined #yosys | 15:46 | |
| *** vidbina has quit IRC | 15:54 | |
| *** anticw has quit IRC | 15:58 | |
| *** kraiskil has quit IRC | 15:59 | |
| *** anticw has joined #yosys | 15:59 | |
| *** dys has quit IRC | 16:06 | |
| *** emeb_mac has joined #yosys | 16:08 | |
| *** kraiskil has joined #yosys | 16:10 | |
| *** anticw has quit IRC | 16:15 | |
| *** anticw has joined #yosys | 16:15 | |
| *** anticw has quit IRC | 16:22 | |
| *** anticw has joined #yosys | 16:23 | |
| *** jakobwenzel has joined #yosys | 16:40 | |
| *** marex-cloud has quit IRC | 16:50 | |
| *** kraiskil has quit IRC | 16:50 | |
| *** rjeli has quit IRC | 16:50 | |
| *** emilazy has quit IRC | 16:55 | |
| *** kraiskil has joined #yosys | 16:55 | |
| *** mithro has quit IRC | 16:57 | |
| *** bubble_buster has quit IRC | 17:00 | |
| *** emilazy has joined #yosys | 17:00 | |
| *** bubble_buster has joined #yosys | 17:02 | |
| *** rswarbrick has quit IRC | 17:03 | |
| *** mithro has joined #yosys | 17:03 | |
| *** anticw has quit IRC | 17:06 | |
| *** anticw has joined #yosys | 17:07 | |
| *** emilazy has quit IRC | 17:07 | |
| *** bubble_buster has quit IRC | 17:10 | |
| *** az0re has joined #yosys | 17:17 | |
| *** mithro has quit IRC | 17:17 | |
| *** bubble_buster has joined #yosys | 17:21 | |
| *** emilazy has joined #yosys | 17:21 | |
| *** rjeli has joined #yosys | 17:22 | |
| *** anticw has quit IRC | 17:23 | |
| *** mithro has joined #yosys | 17:23 | |
| *** anticw has joined #yosys | 17:24 | |
| *** kraiskil has quit IRC | 17:56 | |
| *** anticw has quit IRC | 18:13 | |
| *** anticw has joined #yosys | 18:15 | |
| *** jfcaron has quit IRC | 18:18 | |
| *** jfcaron has joined #yosys | 18:18 | |
| *** BinaryLust has joined #yosys | 18:36 | |
| *** mirage335 has quit IRC | 18:41 | |
| *** mirage335 has joined #yosys | 18:54 | |
| *** mirage335 has quit IRC | 20:18 | |
| *** DaKnig has quit IRC | 20:39 | |
| *** daknig has joined #yosys | 20:45 | |
| *** daknig is now known as DaKnig | 20:46 | |
| *** strobokopp has joined #yosys | 20:48 | |
| *** mirage335 has joined #yosys | 21:03 | |
| *** jakobwenzel has quit IRC | 21:04 | |
| *** simeonm has quit IRC | 22:01 | |
| *** BinaryLust has quit IRC | 22:12 | |
| *** simeonm has joined #yosys | 22:22 | |
| *** Asu has quit IRC | 22:48 | |
| *** jfcaron has quit IRC | 23:15 | |
| *** emeb has left #yosys | 23:27 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!