| *** ghostmansd[m] <ghostmansd[m]!~ghostmans@broadband-109-173-83-100.ip.moscow.rt.ru> has joined #libre-soc | 00:15 | |
| gnucode | sadoon[m]: good morning. :) | 01:42 |
|---|---|---|
| sadoon[m] | Morning | 01:47 |
| sadoon[m] | It's almost dawn here so :) | 01:47 |
| gnucode | sadoon[m]: it's 2 hours from my bedtime. :) | 02:05 |
| *** gnucode <gnucode!~gnucode@user/jab> has quit IRC | 03:00 | |
| *** ghostmansd[m] <ghostmansd[m]!~ghostmans@broadband-109-173-83-100.ip.moscow.rt.ru> has quit IRC | 06:26 | |
| *** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.42.67> has joined #libre-soc | 06:26 | |
| *** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.42.67> has quit IRC | 08:18 | |
| *** ghostmansd[m] <ghostmansd[m]!~ghostmans@broadband-109-173-83-100.ip.moscow.rt.ru> has joined #libre-soc | 08:19 | |
| lkcl | programmerjake, fiiinally - in extremely dark colours, an extremely small near-invisible "tick" icon can be found next to your name on hackaday :) | 08:44 |
| lkcl | took me over 5 minutes to find it! | 08:45 |
| programmerjake | yup, it worked! thx! | 08:52 |
| *** ghostmansd[m] <ghostmansd[m]!~ghostmans@broadband-109-173-83-100.ip.moscow.rt.ru> has quit IRC | 09:44 | |
| *** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.42.163> has joined #libre-soc | 09:45 | |
| lkcl | excellent | 10:26 |
| lkcl | toshywoshy, about learning SVP64: if you happened to write a tutorial (or improve an existing one) it's easy to justify that as an NLnet "documentation" task | 10:27 |
| lkcl | (i.e. you don't have to be out-of-pocket) | 10:27 |
| lkcl | and if you spot anything in the next 2 weeks, that easily justifies putting you on the "reviewed" list of recipients for an RFP as well | 10:31 |
| lkcl | markos, do you know this guy - https://twitter.com/geofflangdale - just spotted he's in charge of hyperscan | 11:33 |
| markos | yes, he's one of the original authors, he left Intel and then returned | 11:34 |
| markos | we did have an email exchange at one point | 11:34 |
| markos | he said that they were planning to do similar things to hyperscan and port it to other arches and create a microbenchmark to help tuning, etc etc | 11:35 |
| markos | but there were 'objections' within Intel to supporting other architectures, sigh | 11:35 |
| lkcl | sigh | 11:41 |
| markos | "I think your overall approach is sound – I would apply the same approach myself in many cases if I had the same constraints or goals. Not having PMOVMSKB is a big headache." | 11:42 |
| lkcl | we've half a dozen ways of achieving that. grevluti is one - i found it produces 100010001000 and other patterns | 11:45 |
| markos | apparently VSX also has something similar | 11:45 |
| markos | not quite but makes emulation much faster | 11:45 |
| lkcl | 1000s of them, including a5a5a5 aa55aa55 77337733 70000007000000 70007000 70707070 77777777 etc. etc. | 11:46 |
| lkcl | yes i remember you mentioned, it was toshywoshy who found it? | 11:46 |
| markos | vgbbd | 11:47 |
| markos | yup | 11:47 |
| markos | gather bits | 11:47 |
| markos | very useful actually | 11:48 |
| *** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.42.163> has quit IRC | 12:10 | |
| *** ghostmansd[m] <ghostmansd[m]!~ghostmans@broadband-109-173-83-100.ip.moscow.rt.ru> has joined #libre-soc | 12:12 | |
| *** octavius <octavius!~octavius@92.40.217.69.threembb.co.uk> has joined #libre-soc | 12:54 | |
| *** ghostmansd[m] <ghostmansd[m]!~ghostmans@broadband-109-173-83-100.ip.moscow.rt.ru> has quit IRC | 13:02 | |
| *** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.42.255> has joined #libre-soc | 13:04 | |
| *** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.42.255> has quit IRC | 14:23 | |
| *** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.162.182> has joined #libre-soc | 14:24 | |
| *** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.162.182> has quit IRC | 14:47 | |
| *** ghostmansd[m] <ghostmansd[m]!~ghostmans@broadband-109-173-83-100.ip.moscow.rt.ru> has joined #libre-soc | 14:48 | |
| lkcl | does anyone know the twin 64-bit multiply instructions in VSX? | 19:15 |
| lkcl | toshywoshy, markos ? | 19:15 |
| markos | twin? | 19:16 |
| markos | 64-bit doubles or ints? | 19:16 |
| lkcl | ints | 19:17 |
| markos | scratch that, there is none for 64-bit ints, but for doubles it's xvmuldp | 19:18 |
| lkcl | 2x 64-in producing 2x 64-out (low-half) | 19:18 |
| lkcl | yyeah, i found vmuleud VRT,VRA,VRB | 19:19 |
| lkcl | p345 | 19:19 |
| lkcl | really odd design | 19:19 |
| markos | strange | 19:19 |
| markos | in the intrinsics reference vec_mul says that 64-bit multiplication is [scalarized] | 19:20 |
| markos | and it's not in 3.0 ISA manual | 19:21 |
| lkcl | i know what they're doing - vmuleud is 64-in x 64-in -> 128-out | 19:21 |
| markos | but it's in 3.1B | 19:21 |
| lkcl | where vmuloud takes the higher-halves | 19:21 |
| markos | so Power10 | 19:21 |
| lkcl | yes | 19:21 |
| markos | right | 19:21 |
| *** octavius <octavius!~octavius@92.40.217.69.threembb.co.uk> has quit IRC | 19:22 | |
| markos | that was news to me, I always assumed that just because vec_mul has a prototype for 64-bit ints, that Power9/VSX *actually* supported 64-bit multiplication and I *have* been using it in the past | 19:26 |
| markos | so that's why performance sucked | 19:26 |
| markos | moral of the story: don't trust the intrinsics, read the assembly! | 19:26 |
| lkcl | :) | 19:30 |
| lkcl | just found vmsumudm VRT,VRA,VRB,VRC | 19:30 |
| lkcl | p. 356 | 19:30 |
| lkcl | ahh they're for performing 256-bit horizontal multiply-and-accumulate | 19:32 |
| lkcl | (of 64-bit dual-packed-SIMD source operands) | 19:33 |
| programmerjake | lkcl: https://info.arxiv.org/help/submit.html | 21:56 |
| programmerjake | we could probably ask jean-paul to endorse us when we have a paper to submit | 21:58 |
| *** octavius <octavius!~octavius@92.40.217.69.threembb.co.uk> has joined #libre-soc | 22:02 | |
| programmerjake | hmm, maybe not, can't find him on arxiv | 22:04 |
| toshywoshy | meeting now? | 22:05 |
| toshywoshy | are you coming lkcl and djac? | 22:05 |
| programmerjake | it was canceled, see the email luke sent | 22:06 |
| markos | ? I didn't receive anything | 22:07 |
| programmerjake | LibreSOC / Microwatt / OpenPOWER Sync MkIII | 22:08 |
| markos | lol | 22:08 |
| programmerjake | afaict you're cc-ed on the email | 22:08 |
| markos | it was on top of the body | 22:08 |
| markos | but it seems everyone in the meeting room read the agenda but *not* the cancelation :D | 22:09 |
| octavius | Heheh, needs to be in **bold** next time | 22:10 |
| programmerjake | lkcl: maybe change the subject to include "Canceled" next time? | 22:10 |
| programmerjake | since not everyone may read the email before the meeting, assuming someone will just paste the contents into chat like usual | 22:11 |
| programmerjake | I checked all the people from lip6 in utils.git, afaict none of them can endorse on arxiv | 22:22 |
| *** octavius <octavius!~octavius@92.40.217.69.threembb.co.uk> has quit IRC | 22:22 | |
| toshywoshy | apparently we were so focussed on the agende nobody read the first line | 22:22 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!