Tuesday, 2018-07-17

*** tpb has joined #yosys00:00
*** promach_ has joined #yosys00:37
cr1901_modernZipCPU: Path bug == "yosys-config returns the wrong directory on Windows." But I need to test something before actually submitting a report01:26
*** seldridge has joined #yosys01:26
cr1901_modernit's possible I forgot to run "make PREFIX=/path/to/root" while compiling01:26
ZipCPUI just ask because Clifford and I were chasing some path bugs today.01:26
ZipCPUI wasn't using Windows though01:26
* cr1901_modern goes ahead and checks01:27
ZipCPUWe were looking into include paths01:27
ZipCPUI was surprised to discover that yosys allows code in multiple directories, and include paths between them01:27
cr1901_modernWhy wouldn't it?01:28
ZipCPUExactly ... I just hadn't used that feature before.01:28
ZipCPUWell, that and .... there were some bugs when using the feature with Verific ... initially, as I recall, there was no way to create new paths.  Then once we fixed that, everytime you specified an include path it would get reset after a single read command.01:29
ZipCPUBut ... that was just my morning's work ;)01:29
cr1901_moderninclude paths are broken on ISE too if memory serves. I.e. you can only add one01:29
*** [Ristovski] is now known as Ristovski01:29
ZipCPUHmm ... I didn't check more than one01:30
cr1901_modernhttps://github.com/YosysHQ/yosys/blob/master/Makefile#L487 (Why don't ppl use m4 for this?)01:32
tpbTitle: yosys/Makefile at master · YosysHQ/yosys · GitHub (at github.com)01:32
cr1901_modernYea in any case, it's my fault for not running w/ "make PREFIX="01:33
* ZipCPU pulls up the link01:33
ZipCPUWith the PREFIX, does it work?01:34
cr1901_modernyes. Well I haven't tested it, but I know it'll work :P01:34
cr1901_modernWhen I see multiple sed expressions laid out like that, my first instinct is "m4 -DDATDIR=$(DATDIR)" etc01:34
ZipCPUYeah .... but I'm kind of glad it's not using m4 .... that's just one more dependency that would need to be installed to use it.01:36
awyglealso I'd have to learn m4 :-P01:37
cr1901_modern(Tbf, POSIX requires m4 to be present :P.)01:37
awygle(technically m4 would just be a build time dependency, not a runtime)01:38
ZipCPUAll good arguments01:39
*** promach_ has quit IRC01:39
*** pie___ has quit IRC01:39
*** m_w has quit IRC01:49
*** seldridge has quit IRC04:44
*** mjoldfield has quit IRC07:05
*** mjoldfield has joined #yosys07:07
*** jaafar has quit IRC07:39
*** xerpi has joined #yosys07:52
*** maartenBE has quit IRC08:14
*** maartenBE has joined #yosys08:14
*** xerpi has quit IRC08:44
*** fsasm has joined #yosys09:45
*** quigonjinn has quit IRC10:04
*** quigonjinn has joined #yosys10:08
*** knielsen has quit IRC10:17
*** knielsen has joined #yosys10:17
*** knielsen has quit IRC10:44
*** knielsen has joined #yosys10:47
*** keesj has quit IRC11:20
*** cr1901_modern has quit IRC11:38
*** cr1901_modern has joined #yosys11:39
*** fsasm has quit IRC11:59
*** jwhitmore has joined #yosys12:02
*** m_t has joined #yosys12:03
*** jwhitmore has quit IRC12:08
*** ar3itrary has quit IRC12:22
*** ar3itrary has joined #yosys12:25
*** keesj has joined #yosys12:27
*** fsasm has joined #yosys12:39
*** ar3itrary has quit IRC12:42
*** ar3itrary has joined #yosys12:46
*** X-Scale has quit IRC12:47
*** dxld has quit IRC13:08
*** dxld has joined #yosys13:10
*** maikmerten has joined #yosys13:23
*** fsasm has quit IRC13:25
*** fsasm has joined #yosys13:28
*** indy has quit IRC14:01
*** promach_ has joined #yosys14:08
*** pie_ has joined #yosys14:21
*** seldridge has joined #yosys14:21
*** indy has joined #yosys14:24
*** mjoldfield has quit IRC14:32
*** mjoldfie_ has joined #yosys14:32
*** quigonjinn has quit IRC14:35
*** mjoldfie_ has quit IRC14:59
*** mjoldfield has joined #yosys14:59
*** jwhitmore has joined #yosys15:24
*** jwhitmore has quit IRC15:29
*** proteus-guy has quit IRC16:03
*** proteus-guy has joined #yosys16:05
*** indy has quit IRC16:18
*** indy has joined #yosys16:21
*** m_t has quit IRC16:24
*** kristianpaul has quit IRC16:31
*** kristianpaul has joined #yosys16:33
*** jwhitmore has joined #yosys16:42
*** digshadow has quit IRC16:57
*** xa0 has quit IRC17:06
*** jaafar has joined #yosys17:11
*** xa0 has joined #yosys17:12
*** pie_ has quit IRC17:17
*** xa0 has quit IRC17:23
*** xa0 has joined #yosys17:31
*** promach_ has quit IRC17:33
*** m_w has joined #yosys17:35
*** jwhitmore has quit IRC17:41
*** emeb has joined #yosys17:46
*** jwhitmore has joined #yosys18:00
*** maikmerten has quit IRC18:00
*** maikmerten has joined #yosys18:10
*** X-Scale has joined #yosys18:21
*** digshadow has joined #yosys18:29
*** fsasm has quit IRC18:41
*** jwhitmore has quit IRC18:56
maikmertenported my design over the the HX8K breakout board, yay! https://github.com/maikmerten/spu32/commit/5e3401f6a67064faa11896b5a037c142c72cd7bd19:02
tpbTitle: initial port of SoC to the Lattice HX8K breakout board · maikmerten/[email protected] · GitHub (at github.com)19:02
maikmertenturns out it helps to actually double-check pin assignments and not misreading the somewhat wacky silk screen on that board19:03
maikmertenin the icestorm toolchain, is there something more than toying around with the seeds to tweak timings? My design somewhat fluctuates between 45 MHz and 56 MHz depending on minor tweaks (pin assignments, minor logic tweaks far away from the critical path) so I guess "placement and routing luck" has a role here19:09
*** jwhitmore has joined #yosys19:29
ZipCPUmaikmerten: Not sure what to say.  The arachne-pnr placer isn't really timing driven, so ... it gets different timing's depending upon the random variables used to generate the placement.19:31
maikmertenwell, I guess that answers the question quite fine, actually ;-)19:37
daveshahYou can pick a seed using -s, or use -r to use a random seed each time19:38
daveshahIf you want to experiment without even modifying the design at all19:38
maikmertenI think -r may be very useful if arachne-pnr and icetime are wrapped into a nice little script with a loop to retry if there's no timing closure19:45
daveshahBecause arachne-pnr is only single-threaded, it would make sense also to run 4-8 times in parallel too19:46
daveshahI've seen a Makefile which did this once19:47
tpbTitle: Makefile - talos-system-fpga - Talos™ II FPGA sources (at git.raptorcs.com)19:47
maikmertenmy Ryzen would sure like that19:47
daveshahI would not say that Makefile is an example to follow19:47
maikmertencan't be worse than mine.19:47
maikmertenoh my.19:48
sorearat the very least use $(seq 1 64)…19:49
maikmertenyeah, -r for arachne-pnr with brute-force threading sure looks like it has very neat potential https://paste.debian.net/1034143/19:53
tpbTitle: debian Pastezone (at paste.debian.net)19:53
*** jwhitmore has quit IRC19:53
*** luismarques has quit IRC19:56
daveshahHaha, you could add two of the most researched and difficult parts of PnR - parallelisation and timing driven - to arachne in 10 lines of bash...19:57
maikmertenof course, that would be parallelisation with no net speed gain ;-)19:58
awyglebut improved QoR which is still a win19:59
maikmertensure :-)20:00
*** X-Scale has quit IRC20:13
*** X-Scale has joined #yosys20:16
*** maikmerten has quit IRC20:20
*** seldridge has quit IRC20:32
*** dxld has quit IRC21:03
*** dxld has joined #yosys21:06
*** seldridge has joined #yosys21:13
*** digshadow has quit IRC21:20
*** digshadow has joined #yosys21:36
*** tinyfpga has quit IRC21:52
*** tinyfpga has joined #yosys21:54
*** digshadow has quit IRC21:58
*** digshadow has joined #yosys22:39
*** knielsen has quit IRC23:17
*** knielsen has joined #yosys23:18

Generated by irclog2html.py 2.13.1 by Marius Gedminas - find it at mg.pov.lt!