Saturday, 2021-01-02

*** tpb has joined #tomu00:00
*** CarlFK has joined #tomu00:49
*** CarlFK has quit IRC00:58
*** UDXS has quit IRC02:33
*** TheJJ has quit IRC02:34
*** TheJJ has joined #tomu02:35
*** TheJJ has quit IRC02:42
*** TheJJ has joined #tomu02:52
*** TheJJ has quit IRC02:55
*** TheJJ has joined #tomu03:35
*** srk has quit IRC03:37
*** srk has joined #tomu03:38
im-tomu[tcal-x] Another update -- the donut demo option DOES work, it is just really slow since the VexRiscv used doesn't have a hardware multiplier.06:04
*** craigo has joined #tomu08:09
*** whatnick has joined #tomu12:25
*** Jay_jayjay has joined #tomu15:04
*** Jay_jayjay has quit IRC15:20
*** Jay_jayjay has joined #tomu16:50
*** Jay_jayjay has quit IRC17:01
*** craigo has quit IRC18:41
*** andi- has quit IRC18:59
*** andi- has joined #tomu19:05
*** Jay_jayjay has joined #tomu19:23
*** Jay_jayjay has quit IRC19:24
*** Jay_jayjay has joined #tomu19:54
im-tomu[tcal-x] I rebuilt the VexRiscv using the DSP blocks for a hard multiplier (it failed timing at 44MHz but still works), recompiled the demo with -march=rv32im, and now I have normal donut speed :)20:09
daveshahnice!20:09
daveshahFPGA timings are pretty conservative (-5% Vcc, 85 degrees, worst process) so usually there is plenty of margin20:10
im-tomu[tcal-x] Yep, Tim A had mentioned it to me.   I'm glad the LiteX script goes ahead and flashes the bitstream even when timing fails by a bit...20:11
tntdaveshah: otoh aren't the dsp timing completely ignored atm ?20:22
daveshahoh yeah, unless a register gets folded into them20:23
daveshahI forgot that I never finished that20:23
tntbest way to meet timing is to use a path that's not analyzed :D20:23
*** z3r0g has joined #tomu21:04
*** Jay_jayjay has quit IRC21:35
*** z3r0g has quit IRC21:51
*** Jay_jayjay has joined #tomu22:39
*** Jay_jayjay has quit IRC22:51
*** Jay_jayjay has joined #tomu22:54
whatnickdaveshah: Is there going to be a nextpnr-xray available soon ? I was following the Litex release with Xilinix support, it was not clear to me what that implies for nextpnr23:01
*** Jay_jayjay has quit IRC23:14
whatnickPerhaps better to ask in #litex , but this channel is more familiar to me23:19
*** Jay_jayjay has joined #tomu23:31

Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!