| *** gnucode <gnucode!~gnucode@user/jab> has quit IRC | 03:20 | |
| markos | lkcl, did you read the comments about the software ecosystem? | 10:01 |
|---|---|---|
| markos | they basically shot themselves in the foot, it's MIPS but not entirely MIPS so everything is totally incompatible | 10:01 |
| markos | so you can't reuse debian mips ports on this cpu | 10:02 |
| markos | our approach is much more sane | 10:02 |
| markos | it will save us *years* of porting development | 10:02 |
| lkcl | toshywoshy, this one's really short and easy to do https://bugs.libre-soc.org/show_bug.cgi?id=1058 | 10:07 |
| lkcl | this one's longer and will need slightly more investment of time in the unit tests https://bugs.libre-soc.org/show_bug.cgi?id=1034 | 10:07 |
| lkcl | but what jacob wrote for crternlogi can be used as a template | 10:08 |
| lkcl | although bmask might be easier as a next candidate after cprop? | 10:09 |
| *** ghostmansd[m] <ghostmansd[m]!~ghostmans@broadband-109-173-83-100.ip.moscow.rt.ru> has quit IRC | 10:58 | |
| *** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.166.89> has joined #libre-soc | 10:58 | |
| lkcl | ls008 v1 is in, after i added appendices for the element stepping functions | 11:00 |
| *** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.166.89> has quit IRC | 11:00 | |
| *** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.170.158> has joined #libre-soc | 11:01 | |
| *** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.170.158> has quit IRC | 11:07 | |
| *** ghostmansd[m] <ghostmansd[m]!~ghostmans@broadband-109-173-83-100.ip.moscow.rt.ru> has joined #libre-soc | 11:17 | |
| *** ghostmansd[m] <ghostmansd[m]!~ghostmans@broadband-109-173-83-100.ip.moscow.rt.ru> has quit IRC | 11:39 | |
| *** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.170.253> has joined #libre-soc | 11:40 | |
| lkcl | markos, you have some warm words about hyperscan? | 11:59 |
| lkcl | https://git.libre-soc.org/?p=libreriscv.git;a=commitdiff;h=f829e7af88c17bb0b3904c5477ab6abef4dec493 | 12:00 |
| lkcl | email me them, or add them directly to the latex i'm happy with either | 12:00 |
| lkcl | https://git.libre-soc.org/?p=libreriscv.git;a=blob;f=conferences/ngisearch2023/ngisearch2023.tex;hb=HEAD | 12:01 |
| lkcl | texstudio does a great job on that btw | 12:01 |
| *** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.170.253> has quit IRC | 12:20 | |
| *** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.56.223> has joined #libre-soc | 12:21 | |
| *** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.56.223> has quit IRC | 12:32 | |
| *** ghostmansd[m] <ghostmansd[m]!~ghostmans@broadband-109-173-83-100.ip.moscow.rt.ru> has joined #libre-soc | 12:32 | |
| *** ghostmansd[m] <ghostmansd[m]!~ghostmans@broadband-109-173-83-100.ip.moscow.rt.ru> has quit IRC | 12:47 | |
| *** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.40.13> has joined #libre-soc | 12:47 | |
| markos | lkcl, I will prepare something and send it to you in the next minutes | 14:49 |
| markos | again, we cannot do anything on *hyperscan* | 14:49 |
| markos | it's always vectorscan | 14:50 |
| markos | hyperscan is Intel's and I have zero control over it, I tried | 14:50 |
| markos | it's not exactly nitpicking but I'm careful not to promise something I cannot deliver | 14:50 |
| markos | it's API and ABI compatible so for end users it should not make a difference | 14:51 |
| markos | sent | 15:10 |
| *** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.40.13> has quit IRC | 16:12 | |
| *** ghostmansd[m] <ghostmansd[m]!~ghostmans@91.205.168.99> has joined #libre-soc | 16:14 | |
| lkcl | thx - mainly, what do we plan to do, what's the goal? | 17:07 |
| lkcl | i.e. what's the goal of the project? in one line: "the plan is to port VectorScan to SVP64" | 17:08 |
| lkcl | and to see if estimates of power reduction and performance can be achieved | 17:09 |
| lkcl | markos, ok let me know if you're happy with this https://git.libre-soc.org/?p=libreriscv.git;a=blob;f=conferences/ngisearch2023/ngisearch2023.tex;hb=HEAD | 17:21 |
| lkcl | very very brief, it's literally "what is it, what does it do, and what are we going to attempt" | 17:21 |
| lkcl | it'll be like under a minute available per slide, maybe only 40 seconds. | 17:24 |
| lkcl | there's 10 projects, total time is 60 minutes, that's 6 minutes per project *including Q&A* | 17:24 |
| markos | yes, that's correct | 17:47 |
| markos | basically, it's as you said, port it to SVP64 | 17:47 |
| markos | ideally we are aiming for full porting of every algorithm, but if that's not possible then at least it should be usable even in scalar form in some algorithms | 17:48 |
| markos | remember it doesn't even compile currently if no SIMD is available | 17:48 |
| markos | so one of the intermediate goals is to provide normal scalar implementations of algorithms and port those to SVP64 | 17:48 |
| markos | which is a good thing in the long run, it will run on all arches with working C compiler | 17:49 |
| lkcl | bleugh! :) | 17:51 |
| lkcl | i remember you mentioned about going back to an older version | 17:51 |
| markos | to look for the scalar versions yes, but I wouldn't actually "revert" the whole project to an older version | 18:01 |
| markos | so this would be the first step in my approach, check the original scalar functions, make sure it runs with NO SIMD unit available | 18:02 |
| markos | eg on SFFS :) | 18:02 |
| markos | and then start porting each function to SVP64 through the emulator | 18:02 |
| markos | the nice thing is that we have a huge unit test suite so it really means we can get a full 100% usable port | 18:03 |
| markos | btw, what's the video call medium? jitsi/zoom/other? | 18:05 |
| markos | I don't remember receiving any such link | 18:05 |
| markos | ah it's jitsi | 18:05 |
| markos | nm | 18:06 |
| lkcl | the one on hivane.net | 21:05 |
| *** DevHONKIntel486S <DevHONKIntel486S!~allonsenf@2001:470:69fc:105::2:4883> has joined #libre-soc | 21:05 | |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!