15:24 <litghost> skydive: I'm tracking an OOM failure that occurs on arty100T with some placement seeds. We have been seeing it on CI with some frequency. I believe I know how to fix the issue, and will be working on it this week.

15:28 <skydive> Thanks a lot for the update!

15:30 <litghost> skydive: So as a workaround, you can provide another seed to the placement step, and it "typically" works. To pass a seed to placement set env var VPR_SEED="--seed <int>" before invoking the flow

15:31 <litghost> skydive: Not ideal, but might give you a chance to make progress while we work on a fix

15:36 <litghost> skydive: It is worth noting that I expect the arty100T flow right now to use ~2-6 GB. It is worth noting that not all of that is resident at the same time, but it does imply a minimum RAM requirement for the moment

15:37 <skydive> litghost: thanks for the heads up regarding RAM. I should be fine, I have about 7GB free.

15:38 <skydive> Setting that environment variable seems to break the symbiflow_pack command. "vpr architecture circuit" is telling me --seed is an unexpected argument.

15:38 <litghost> skydive: Huh, let me check, I might have provide a bad argument then

15:38 <skydive> Whoops - i accidently had an equals in there.

15:39 <litghost> skydive: Ya, that'll do it

16:32 <sf-slack> <ryancj14> @mithro: I've studied fpga-tool-perf extensively now and I was wondering if it might be nice to have a wiki for fpga-tool-perf. I've compiled some of the notes I've taken on the repo into an example wiki in ryancj14/fpga-tool-perf: https://github.com/ryancj14/fpga-tool-perf/wiki Do you think something like this would be a good addition to the repo?

16:32 <tpb> Title: Home · ryancj14/fpga-tool-perf Wiki · GitHub (at github.com)

16:35 <mithro> ryancj14: Yes

16:43 <_whitenotifier-b> [fpga-tool-perf] ryancj14 opened issue #189: Adding wiki to fpga-tool-perf - https://git.io/JJ0Ah

16:49 <skydive> litghost: Even after exporting the seed, it still doesn't seem to work reliably. Same problem. Do you have an example seed that functions properly?

16:49 <skydive> I'll be patient - thanks for working on this!

16:50 <litghost> skydive: Huh, can you file an issue and attach the failing log file?

16:51 <skydive> Will do so tonight. What's the optimal way to get a log output? (Command) > file.log 2>&1

16:57 <litghost> skydive: Logs should be in files with extension *.log, so simply tar them all

18:27 <sf-slack> <johnnybui> Is there any specific reason that the icebreaker board is reporting frequency as 81.05 while all the other boards/toolchains have frequencies in the millions (246366100). I assume it's some issue with inconsistency with units (hertz vs megahertz). Here's an example: https://gist.github.com/TypingKoala/80a1144e30eb3d678f30f8f063baea63

18:27 <tpb> Title: arty.json · GitHub (at gist.github.com)

19:50 <litghost> johnnybui: Seems like a bug? Please file an issue with the relevant info so it can be cleaned up, both in the code and in the documentation

20:06 <_whitenotifier-b> [fpga-tool-perf] TypingKoala opened issue #190: Inconsistent units for icebreaker board max frequency - https://git.io/JJEJK

21:07 <_whitenotifier-b> [FPGA-Tool-Performance-Visualization-Library] TypingKoala opened issue #11: Implement enums for AddNormalizedColumn direction parameter - https://git.io/JJETW

21:59 <mithro> hzeller: https://github.com/SymbiFlow/symbiflow-xc7z-automatic-tester

21:59 <tpb> Title: GitHub - SymbiFlow/symbiflow-xc7z-automatic-tester: Tool for automatically testing FPGA designs using a Zynq Series 7 board. (at github.com)

22:19 <_whitenotifier-b> [symbiflow-examples] Skydive opened issue #37: Arty A7 100T genfasm crash (in counter_test) - https://git.io/JJELP

22:20 <skydive> Oh - I was going to notify here. - cool! Bot is automatic

22:21 <skydive> Sorry - I'm on the move for a week. Hard to be timely.

22:33 <litghost> skydive: No problem. Your issue is different than the one I'm investigating, but I'll take a look

22:34 <skydive> Hmm - i couldn't reproduce my earlier logs...

22:34 <skydive> It seems to break in the same section of the Makefile