Tuesday, 2022-06-21

*** DiffieHellman is now known as GNU\DiffieHellma05:15
*** GNU\DiffieHellma is now known as DiffieHellman05:15
ghostmansdOut of curiosity, why the doc repository is still called libreriscv? Shouldn't it be renamed to something arch-agnostic?13:59
lkclit's legacy14:00
lkcland i have to do a frickin lot at the back-end because it's the source repository for ikiwiki14:00
lkcland change the gitolite3.conf for everyone14:00
ghostmansdah OK14:00
lkcland change the "overlay" system which pulls in 2 other repos14:01
lkcland and and and14:01
ghostmansdjust curiosity, I don't in any way suggest doing it now :-)14:01
ghostmansdI wanted to update my page14:01
ghostmansdand thought about this misnaming14:01
lkclvia the website the wiki's the right name14:01
ghostmansdyeah but still I mostly update the repo14:02
ghostmansdI generally don't edit docs via web page14:02
ghostmansdthat's why I'm aware of the differenct14:02
lkclyehyeh me too14:03
lkcli just did "mv libreriscv libresoc" and forgot about it14:03
lkcl(on the local checkout)14:03
ghostmansdlkcl, had to adjust tor to just check your tweet14:05
ghostmansdgood news!14:05
lkclsigh joy14:05
lkclyes, i asked him to add decoupling capacitors and "funded by NLnet under EU Grants" to the silkscreen14:06
lkclhe ran it at 100 mhz and it maaaaasively dropped the power rails14:06
lkclbut i think that's simply because it's such a massive chip14:06
tplatenlong time ago I worked on the orangecrab port of ls2, last time I tried to get dram working where I got this error17:28
tplatenERROR: DQS group mismatch, port DQSW270 of 'ddrphy.U$$46' in group LDQ41 is driven by DQSBUFM 'ddrphy.U$$45' in group LDQ5317:28
lkclcesar, you were investigating this, weren't you?18:27
ghostmansdThis binutils discussion hangs way too long. I wish GNU folks could spend at least 1% of these efforts on copyright assignment, to be honest. :-)18:30
tplatencross post from #yosys19:12
tplatenno ##openfpga19:12
tplaten(08:05:00 PM) gatecat: might it be this issue? https://github.com/orangecrab-fpga/orangecrab-hardware/issues/2419:13
tplaten(08:05:13 PM) gatecat: (this is just a case of the schematic being wrong aiui and the hw fine)19:13
ghostmansdI'm in the middle of some binutils code refactoring for macros and prefix migration. You wouldn't believe, but I used setjmp/longjmp combo. I haven't had a need to use it for years, but for this part of code, it really fits naturally.19:24
ghostmansdObviously I've used it for exception. When we have some issue in parser, we immediately jump to the failure label some frames before.19:25
ghostmansdIn my defence, I found they already use this combo, and exactly for parsing. So I guess I'm not the only fucking lunatic there.19:26
ghostmansdFor coroutines, I rather prefer such creepy stuff to be stackful. But for exceptions, setjmp/longjmp is OK.19:27
lkclye gods i haven't seen that since dcethread20:00
lkcltplaten, that sounds about right - best to double-check the ECP5 BGA285 datasheet20:02
tplatenI'll do that20:03
lkcli think we all got rev 0.220:10
cesartplaten: I intend to submit a patch to nmigen-boards, aligning it to orangecrab schematics, but also referencing the lpf file in orangecrab-examples, and those generated by litex-boards.21:49
cesarBut, before flashing an ls2 in the orangecrab, I first want to make some Icarus simulations. I'm not very confident with the gram sdram core just yet. Litedram on the other hand seems to work just fine, so I can use it as a reference if needed.21:53
programmerjakewe were mentioned on phoronix again: https://www.phoronix.com/scan.php?page=news_item&px=ThinkSilicon-NEOX-RISC-V-GPU21:55
programmerjakei replied in the comments: https://www.phoronix.com/forums/forum/phoronix/latest-phoronix-articles/1329872-think-silicon-shows-off-first-risc-v-3d-gpu/page3#post133000321:55
programmerjakeafaoct that's 64 Gbaud -- crazy fast21:59
programmerjakei'll be in the meeting in a sec...22:00

Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!