| *** tpb has joined #litex | 00:00 | |
| *** CarlFK has quit IRC | 00:02 | |
| *** acathla has joined #litex | 00:22 | |
| *** tpb` has joined #litex | 00:58 | |
| *** keesj has quit IRC | 01:05 | |
| *** Finde has quit IRC | 01:05 | |
| *** felix_ has quit IRC | 01:05 | |
| *** Dolu has quit IRC | 01:05 | |
| *** acathla has quit IRC | 01:05 | |
| *** tpb has quit IRC | 01:05 | |
| *** futarisIRCcloud has quit IRC | 01:05 | |
| *** daveshah has quit IRC | 01:05 | |
| *** tumbleweed has quit IRC | 01:05 | |
| *** sorear has quit IRC | 01:05 | |
| *** mithro has quit IRC | 01:05 | |
| *** xobs has quit IRC | 01:05 | |
| *** flammit has quit IRC | 01:05 | |
| *** _florent_ has quit IRC | 01:05 | |
| *** nrossi has quit IRC | 01:05 | |
| *** spacekookie has quit IRC | 01:05 | |
| *** key2 has quit IRC | 01:05 | |
| *** tpb` is now known as tpb | 01:05 | |
| *** tpb has joined #litex | 01:12 | |
| *** hitchcock.freenode.net sets mode: +ns | 01:12 | |
| *** hitchcock.freenode.net sets mode: -o tpb | 01:23 | |
| -hitchcock.freenode.net- *** Notice -- TS for #litex changed from 1566435756 to 1550731482 | 01:23 | |
| *** acathla has joined #litex | 01:23 | |
| *** key2 has joined #litex | 01:23 | |
| *** spacekookie has joined #litex | 01:23 | |
| *** felix_ has joined #litex | 01:23 | |
| *** Finde has joined #litex | 01:23 | |
| *** keesj has joined #litex | 01:23 | |
| *** tumbleweed has joined #litex | 01:23 | |
| *** daveshah_ has joined #litex | 01:23 | |
| *** futarisIRCcloud_ has joined #litex | 01:23 | |
| *** nrossi has joined #litex | 01:23 | |
| *** Dolu has joined #litex | 01:23 | |
| *** _florent_ has joined #litex | 01:23 | |
| *** flammit has joined #litex | 01:23 | |
| *** xobs has joined #litex | 01:23 | |
| *** mithro has joined #litex | 01:23 | |
| *** sorear has joined #litex | 01:23 | |
| *** hitchcock.freenode.net changes topic to "LiteX Soc builder and Cores (http://enjoy-digital.fr) / Github : https://github.com/enjoy-digital, https://github.com/litex-hub / Logs: https://logs.timvideos.us/%23litex/latest.log.html" | 01:23 | |
| *** Xark has joined #litex | 01:25 | |
| *** somlo has joined #litex | 01:25 | |
| *** nrossi has quit IRC | 01:26 | |
| *** xobs has quit IRC | 01:26 | |
| *** CarlFK has joined #litex | 01:27 | |
| *** tumbleweed has quit IRC | 01:30 | |
| *** tumbleweed has joined #litex | 01:30 | |
| *** nrossi has joined #litex | 01:33 | |
| *** xobs has joined #litex | 01:35 | |
| *** futarisIRCcloud_ has quit IRC | 02:20 | |
| *** futarisIRCcloud has joined #litex | 02:21 | |
| *** CarlFK1 has joined #litex | 05:00 | |
| *** CarlFK has quit IRC | 05:03 | |
| *** CarlFK1 has quit IRC | 06:51 | |
| *** CarlFK has joined #litex | 07:31 | |
| *** daveshah_ has quit IRC | 07:52 | |
| *** daveshah has joined #litex | 07:53 | |
| somlo | \o/ successfully built litex+rocket on a riscv64 fedora VM (natively) | 10:42 |
|---|---|---|
| somlo | well, timing didn't pass 60MHz, but then I had to re-run it several times on intel to get that to pass, so I'm still declaring victory :) | 10:43 |
| daveshah | FYI, 55MHz is also fine for the TFTP | 10:49 |
| daveshah | and should be much more likely to pass | 10:49 |
| somlo | daveshah: http://www.contrib.andrew.cmu.edu/~somlo/BTCPP/screenlog.0 | 10:55 |
| daveshah | 404? | 10:56 |
| somlo | make that http://www.contrib.andrew.cmu.edu/~somlo/BTCP/screenlog.0 | 10:56 |
| somlo | and it does pass 55 MHz -- I'll try shoving the bitstream at the versa when I make it into the office later this morning :) | 10:57 |
| somlo | looks absolutely no different than on intel | 10:58 |
| somlo | I "cheated" by creating riscv64-unknown-elf-foo scripts calling "/usr/bin/foo $*" to get around the built-in assumption of litex that it's cross-compiling on Intel :) | 10:59 |
| somlo | I'll have to dig around the build script and fix that properly | 11:00 |
| daveshah | Otherwise looks very good | 11:02 |
| daveshah | I guess the emulator is only a couple of times faster than hardware? | 11:02 |
| daveshah | So PnR times might be within feasible territory | 11:02 |
| daveshah | idk what the riscv compiler ecosystem is like, I would have thought you might need a cross compiler for baremetal (even same arch)? | 11:03 |
| somlo | apparently not, command line flags was all that was needed... | 11:06 |
| daveshah | Interesting | 11:06 |
| daveshah | Incidentally, to get the bitstream after a timing failure you will need to run the ecppack part manually | 11:07 |
| somlo | now I'm actively waiting for a batch of trellis boards to try and run Fedora on them :) | 11:07 |
| daveshah | I can always try stuff you send me | 11:07 |
| daveshah | I guess the first step would be to get a network rootfs working | 11:08 |
| somlo | re. ecppack: good to know, beats waiting another 10 hours for a new build -- although I could use a screenlog.0 file that doesn't ostensibly end with a "failure" :) | 11:08 |
| daveshah | You can always modify litex to pass - -timing-ignore-fail to nextpnr | 11:10 |
| daveshah | * --timing-ignore-fail | 11:10 |
| daveshah | The failures here are much smaller than the safety margin (particularly considering the FPGA isn't near max temp) so won't affect function | 11:10 |
| somlo | but then it will still try to run the clock at the requested frequency (rather than what it actually passed) | 11:11 |
| somlo | oh | 11:11 |
| somlo | I'll try to build it at requested 55MHZ (on intel) to see if it works otherwise, then just ask for that on the demo native build | 11:12 |
| somlo | re. network rootfs: that appears to be the next thing on the todo list... and I'll definitely buy you about trying stuff on the trellis board -- thanks! | 11:18 |
| somlo | *bug you about trellis board | 11:18 |
Generated by irclog2html.py 2.13.1 by Marius Gedminas - find it at mg.pov.lt!