| *** tpb <[email protected]> has joined #symbiflow | 00:00 | |
| *** maartenBE <maartenBE!~maartenBE@freenode/user/maartenBE> has quit IRC (Ping timeout: 120 seconds) | 02:16 | |
| *** maartenBE <[email protected]> has joined #symbiflow | 02:17 | |
| tpb | <sf-slack> <timo.callahan> nextpnr question: Is there a python binding for ctx->getBelLocation(bel) ? Or any other way to get an absolute x/y/z for a bel from the Python API (https://github.com/YosysHQ/nextpnr/blob/master/docs/python.md) ? | 03:25 |
|---|---|---|
| tpb | <sf-slack> <timo.callahan> ...or does parsing the Bel name (e.g. `X15/Y1/lc6` ) for the X and Y give good enough results? | 03:48 |
| tpb | <sf-slack> <timo.callahan> ...hmm I think that's it . | 03:51 |
| tpb | <cr1901> timo.callahan: Cannot speak for what the Python API does, but X and Y are prob fine to parse from the string. | 05:07 |
| tpb | <cr1901> as for, say, the lc6 part, there's no guarantee that z=6 for the bel "X15/Y1/lc6" | 05:08 |
| tpb | <cr1901> https://github.com/cr1901/prjtrellis/blob/facade/libtrellis/src/Chip.cpp#L312-L318 Here's an example where that matters... | 05:09 |
| tpb | <cr1901> DCC6 and DCM6 exist in the tile. As do DCC7 and DCM7. I elect to give DCC6 a z=6, DCC7 a z=7, DCM6 a z=8, and DCM7 a z=9 | 05:10 |
| tpb | <cr1901> I don't actually remember if I use the z coord for anything tho. | 05:12 |
| tpb | <tcal> cr1901: thanks, X/Y does seem to work (I'm ignoring Z for now). Here's a plot that adds some random dithering so they're not all stacked on each other: https://usercontent.irccloud-cdn.com/file/bBYEFYKr/nxsoc2.png | 05:43 |
| tpb | <sf-slack> <unai.martinezcorral> @carlosedp see https://github.com/umarcor/symbiflow-examples/runs/4516424805?check_suite_focus=true | 07:26 |
| tpb | <sf-slack> <unai.martinezcorral> @blox see https://github.com/YosysHQ/oss-cad-suite-build/issues/1 | 07:31 |
| -_whitenotifier-2553- [symbiflow-examples] umarcor opened issue #227: SymbiFlow examples and containers - https://git.io/JDR5h | 07:45 | |
| tpb | <sf-slack> <mkurc> @timo.callahan In nextpnr-nexus Z coordinate of a BEL reflect more-less its index within a tile. | 08:56 |
| *** kraiskil <[email protected]> has joined #symbiflow | 17:36 | |
| -_whitenotifier-2553- [prjxray] mithro opened issue #1803: Publish new spartan7 section in prjxray-db - https://git.io/JDEKl | 18:10 | |
| tpb | <mithro> @mkurc: Any idea what is up with the CMT tile changes in https://storage.googleapis.com/symbiflow-prjxray/artifacts/prod/foss-fpga-tools/prjxray/presubmit/database/zynq7/2326/20211213-024237/database/diff.patch ? | 18:15 |
| tpb | <mithro> acomodi: Can you take a quick look over WIP addition of the spartan 7 database @ https://github.com/SymbiFlow/prjxray-db/compare/master...mithro:master before I push it? | 18:19 |
| tpb | <sf-slack> <acomodi> mithro: I couldn't spot any clearly wrong stuff at a first glance, I think it could be good to check in | 18:27 |
| tpb | <mithro> Pushed! | 18:38 |
| tpb | <gatecat> timo.callahan: the missing getBelLocation binding was an omission; fixed in https://github.com/YosysHQ/nextpnr/pull/872 | 18:48 |
| tpb | <tnt> gatecat: btw, any comment on https://github.com/YosysHQ/nextpnr/issues/864 ? | 19:17 |
| tpb | <gatecat> ugh, the ice40 carry packer, I'll see if there's any easy fix but it's not code I enjoy touching tbh | 19:17 |
| tpb | <tnt> heh, I can attempt a fix, but the code seem written atm to explicitly prevent packing of 'contant' CI carry blocks with a lut and I was wondering if this was an oversight or if there was a reason for that, some bad scenario that can happen or something. | 19:19 |
| tpb | <gatecat> I've got the feeling there was a possibly nasty (possibly an attempted fix to it packing LUTs that didn't actually belong there) but it's a long time ago | 19:22 |
| tpb | <tnt> I mean, I can "imagine" possibly packing the wrong lut if there are severa candidates but ... how bad is it really ... | 19:23 |
| tpb | <tnt> (the more I look at this and another related issue that was solved in yosys with -dlogic, the more I think the LUT and CARRY should be "tagged" during synthesis as belonging together) | 19:24 |
| -_whitenotifier-2553- [prjxray] mithro opened issue #1804: GitHub pages for prjxray-db not publishing - https://git.io/JDESQ | 19:25 | |
| tpb | <gatecat> tnt: I think https://github.com/YosysHQ/nextpnr/pull/873 should at least fix this case with minimal regression | 19:30 |
| tpb | <tnt> gatecat: ack, tx. I'll give that a shot tomorrow or later tonight and report in the PR. | 19:31 |
| tpb | <sf-slack> <unai.martinezcorral> With regard to https://github.com/SymbiFlow/symbiflow-examples/issues/227, • Shall we create hdl/examples as a superset (support any environment setup procedure) of SymbiFlow/symbiflow-examples (conda only)? • Shall the "conda only" content be moved from SymbiFlow to hdl/conda-examples and allow SymbiFlow/symbiflow-examples to include content about multiple "environment setup" procedures? From a | 20:05 |
| tpb | <sf-slack> content organisation point of view, I think that the second option is desirable. However, that is quite disruptive. Even if we used intersphinx, some of the references to the current docs would change in the near future. | 20:05 |
| *** maartenBE <maartenBE!~maartenBE@freenode/user/maartenBE> has quit IRC (Ping timeout: 120 seconds) | 20:25 | |
| *** maartenBE <[email protected]> has joined #symbiflow | 20:26 | |
| *** kraiskil <[email protected]> has quit IRC (Ping timeout: 120 seconds) | 21:04 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!
