| *** tpb has joined #yosys | 00:00 | |
| *** cr1901 has quit IRC | 00:23 | |
| *** cr1901 has joined #yosys | 00:24 | |
| *** emeb_mac has joined #yosys | 00:46 | |
| *** emeb_mac has quit IRC | 02:19 | |
| *** AlexDaniel has quit IRC | 02:27 | |
| *** PyroPeter has quit IRC | 02:30 | |
| *** PyroPeter has joined #yosys | 02:43 | |
| *** emeb has quit IRC | 03:07 | |
| *** emeb_mac has joined #yosys | 03:23 | |
| *** s_frit has quit IRC | 03:38 | |
| *** s_frit has joined #yosys | 03:39 | |
| *** citypw has joined #yosys | 03:40 | |
| *** attie has quit IRC | 04:14 | |
| *** citypw has quit IRC | 04:15 | |
| *** citypw has joined #yosys | 04:15 | |
| *** TFKyle has joined #yosys | 04:17 | |
| *** attie has joined #yosys | 04:23 | |
| *** s_frit_ has joined #yosys | 05:27 | |
| *** s_frit has quit IRC | 05:29 | |
| *** fevv8[m] has left #yosys | 06:33 | |
| *** dys has joined #yosys | 06:34 | |
| *** Jybz has joined #yosys | 06:47 | |
| *** indy has joined #yosys | 07:05 | |
| *** emeb_mac has quit IRC | 07:14 | |
| *** s_frit_ has quit IRC | 07:40 | |
| *** _whitelogger has quit IRC | 08:45 | |
| *** _whitelogger has joined #yosys | 08:47 | |
| *** proteusguy has quit IRC | 09:05 | |
| *** proteusguy has joined #yosys | 09:07 | |
| pepijndevos_ | Warning: Yosys has only limited support for tri-state logic at the moment. (../benchmarks/MCPU.v:61) | 10:20 |
|---|---|---|
| pepijndevos_ | Trying to synthesize https://github.com/cpldcpu/MCPU/blob/master/verilog/MCPU_0.1a.v and failing | 10:20 |
| tpb | Title: MCPU/MCPU_0.1a.v at master · cpldcpu/MCPU · GitHub (at github.com) | 10:20 |
| tnt | pepijndevos_: yeah. "assign data = states!=3'b001 ? 8'bZZZZZZZZ : accumulator[7:0]; " needs some change. | 10:26 |
| tnt | Also, is that supposed to be a top level ? I mean tristate really only makes sense for IOs. No modern fpga has internal tristates. | 10:27 |
| corecode | yea too bad that there is no easy way to communicate an output enable other than explicit second line | 10:27 |
| corecode | but i guess it makes it more explicit | 10:27 |
| pepijndevos_ | tnt, not totaly sure tbh, just some core I snatched from the web. | 10:29 |
| corecode | tnt: do you think it would make sense to automatically convert tristates to (hidden) enable signals and then connect to buffer OE? | 10:32 |
| corecode | seems messy and too implicit | 10:32 |
| corecode | very behavioral, not RTL | 10:32 |
| *** adjtm has joined #yosys | 10:35 | |
| *** _whitelogger has quit IRC | 10:54 | |
| *** _whitelogger has joined #yosys | 10:56 | |
| *** Jybz has quit IRC | 11:13 | |
| *** Jybz has joined #yosys | 11:17 | |
| *** dys has quit IRC | 11:25 | |
| *** adjtm has quit IRC | 11:58 | |
| *** maikmerten has joined #yosys | 11:59 | |
| maikmerten | with latest yosys master LUT usage is back to normal (nextpnr can pack things together nicely again) | 12:16 |
| maikmerten | so thanks a lot! :-) | 12:16 |
| maikmerten | with relut being at work and nextpnr being able to pack things again, LUT-usage overall is now lower than ever | 12:24 |
| *** rrika has quit IRC | 12:33 | |
| maikmerten | f_max is also back to normal, which means ~35 MHz (which I'm used to, caused by a carry chain) vs. ~45 MHz (with the original relut branch merged, with the carry chain no longer being the critial path). I wonder how the pre-fix yosys/nextpnr flow managed that. | 12:33 |
| *** rrika has joined #yosys | 12:33 | |
| *** cr1901 has quit IRC | 13:17 | |
| *** cr1901 has joined #yosys | 13:17 | |
| *** dys has joined #yosys | 13:22 | |
| *** adjtm has joined #yosys | 13:36 | |
| *** Jybz has quit IRC | 13:41 | |
| pepijndevos_ | Huh, I'm confused. Looking at rtlil.h I swear I saw the actual implementations of add* functions, but I can't find them. I guess I'm being thick. | 15:03 |
| pepijndevos_ | git grep addAdd does not seem to find any actual implementation. | 15:03 |
| pepijndevos_ | oh, addAdff is there alright... but where are the others... | 15:05 |
| pepijndevos_ | ah... DEF_METHOD | 15:06 |
| pepijndevos_ | Is $mux always two-way? Because there is a $_MUX4_ but I've never seen $mux4, yet in rtlil.h it seems only two-way. | 15:11 |
| daveshah | Larger muxes would end up either as $pmux or $shiftx depending on coding style | 15:12 |
| daveshah | Or perhaps even a tree of $mux if coded using a tree of ?: | 15:13 |
| pepijndevos_ | ah ok | 15:14 |
| pepijndevos_ | I was looking at this code https://github.com/tgingold/ghdlsynth-beta/blob/db6d9f374de1eb1c074c2b9828bc6d99055b3624/ghdl/ghdl.cc#L340 | 15:14 |
| tpb | Title: ghdlsynth-beta/ghdl.cc at db6d9f374de1eb1c074c2b9828bc6d99055b3624 · tgingold/ghdlsynth-beta · GitHub (at github.com) | 15:14 |
| pepijndevos_ | For a mux4 it indeed creates a tree of muxes, and was curious if that's the correct way to do it. | 15:14 |
| daveshah | It's as good as any | 15:15 |
| pepijndevos_ | Cool | 15:15 |
| daveshah | I'm not sure why ghdl has a mux4 cell in the first place... | 15:15 |
| pepijndevos_ | I'm going to try to add module instatiations. We'll see if just adding it as a cell and connecting wires does the job. | 15:16 |
| daveshah | Yes, modulo parameterisation | 15:16 |
| pepijndevos_ | I think that's handled on the ghdl side, but not totally sure. Tristan mentioned he implemented modules in synthization but not on the yosys side, so I thought I'd try doing something useful. | 15:18 |
| daveshah | For anything that isn't a blackbox leaf cell, you'll probably want to strip parameters from the instance and give each parameterised variant a unique name | 15:19 |
| pepijndevos_ | right | 15:19 |
| maikmerten | is there a list of nextpnr python console commands? | 15:28 |
| daveshah | No, unfortunately not | 15:31 |
| maikmerten | ah, okay :-) | 15:33 |
| daveshah | For querying the Arch database you can follow https://github.com/YosysHQ/nextpnr/blob/master/docs/archapi.md | 15:33 |
| tpb | Title: nextpnr/archapi.md at master · YosysHQ/nextpnr · GitHub (at github.com) | 15:33 |
| maikmerten | just wondering if there is a way to highlight critical paths in the graphical overview | 15:34 |
| maikmerten | but perhaps stuff like that is not what the python console is actually meant for | 15:34 |
| maikmerten | (I imagine it's more for orchestrating processing steps) | 15:34 |
| daveshah | No, unfortunately there's no Python or C++ api to get critical paths yet | 15:35 |
| daveshah | The timing stuff isn't really integrated very well | 15:35 |
| maikmerten | it's still plenty neat :-) | 15:36 |
| maikmerten | help() at the time being is a halt-and-catch-fire alias | 15:38 |
| maikmerten | I guess that's because it's doing interactive stuff with stdin and stdout presumably | 15:39 |
| daveshah | Tab completion should work | 15:40 |
| maikmerten | it does! | 15:40 |
| *** emeb has joined #yosys | 16:07 | |
| *** citypw has quit IRC | 16:29 | |
| pepijndevos_ | What's a blackbox module supposed to look like in `dump`? Currently ghdl just makes a module \name end | 16:40 |
| *** adjtm has quit IRC | 16:40 | |
| daveshah | It also needs ports and a blackbox attribute set on it | 16:40 |
| pepijndevos_ | I thought as much... some work to do then. | 16:42 |
| *** adjtm has joined #yosys | 18:16 | |
| *** maikmerten has quit IRC | 18:26 | |
| *** emeb_mac has joined #yosys | 18:53 | |
| *** pie_ has quit IRC | 19:05 | |
| *** cr1901 has quit IRC | 19:12 | |
| *** cr1901 has joined #yosys | 19:12 | |
| *** Thorn has quit IRC | 20:05 | |
| *** pie_ has joined #yosys | 21:32 | |
| *** pie_ has joined #yosys | 21:33 | |
| *** pie_ has joined #yosys | 21:35 | |
| *** emeb_mac has quit IRC | 21:35 | |
| *** Thorn has joined #yosys | 21:54 | |
| *** cr1901 has quit IRC | 22:02 | |
| *** cr1901 has joined #yosys | 22:02 | |
| *** _whitelogger has quit IRC | 23:48 | |
| *** _whitelogger has joined #yosys | 23:50 | |
Generated by irclog2html.py 2.13.1 by Marius Gedminas - find it at mg.pov.lt!