| *** tpb has joined #yosys | 00:00 | |
| *** miek has quit IRC | 00:15 | |
| *** show has quit IRC | 00:15 | |
| *** marex-cloud has quit IRC | 00:15 | |
| *** guan has quit IRC | 00:15 | |
| *** sandeepkr has quit IRC | 00:15 | |
| *** kuldeep has quit IRC | 00:15 | |
| *** show has joined #yosys | 00:16 | |
| *** sandeepkr has joined #yosys | 00:16 | |
| *** miek has joined #yosys | 00:16 | |
| *** kuldeep has joined #yosys | 00:17 | |
| *** phantomcircuit has quit IRC | 00:17 | |
| *** pointfree has quit IRC | 00:18 | |
| *** pointfree has joined #yosys | 00:19 | |
| *** phantomcircuit has joined #yosys | 00:20 | |
| *** marex-cloud has joined #yosys | 00:31 | |
| *** emeb has quit IRC | 00:35 | |
| *** emeb_mac has joined #yosys | 00:39 | |
| *** citypw has quit IRC | 00:44 | |
| *** gsi__ has joined #yosys | 00:51 | |
| *** SpaceCoaster has joined #yosys | 00:51 | |
| *** gsi_ has quit IRC | 00:54 | |
| *** sorear has quit IRC | 00:58 | |
| *** zignig_bedtime is now known as zignig | 01:19 | |
| *** SpaceCoaster has quit IRC | 01:24 | |
| *** m4ssi has joined #yosys | 01:55 | |
| *** m4ssi has quit IRC | 01:59 | |
| *** sorear has joined #yosys | 02:28 | |
| *** PyroPeter has quit IRC | 02:46 | |
| *** PyroPeter has joined #yosys | 02:59 | |
| *** futarisIRCcloud has quit IRC | 03:28 | |
| *** futarisIRCcloud has joined #yosys | 04:24 | |
| *** rohitksingh_work has joined #yosys | 04:27 | |
| *** rohitksingh has joined #yosys | 04:59 | |
| *** brandonz has quit IRC | 05:14 | |
| *** rohitksingh_work has quit IRC | 05:45 | |
| *** rohitksingh has quit IRC | 05:45 | |
| *** rohitksingh has joined #yosys | 05:46 | |
| *** dys has quit IRC | 05:50 | |
| *** s_frit has quit IRC | 05:53 | |
| *** s_frit has joined #yosys | 05:54 | |
| *** proteusguy has quit IRC | 06:00 | |
| *** proteusguy has joined #yosys | 06:13 | |
| *** kraiskil has joined #yosys | 06:40 | |
| *** ironsteel has joined #yosys | 06:55 | |
| *** emeb_mac has quit IRC | 07:06 | |
| *** jakobwenzel has quit IRC | 07:16 | |
| *** jakobwenzel has joined #yosys | 07:16 | |
| *** adjtm has quit IRC | 08:03 | |
| *** rohitksingh_work has joined #yosys | 08:18 | |
| *** m4ssi has joined #yosys | 08:48 | |
| *** adjtm has joined #yosys | 08:54 | |
| *** rohitksingh has quit IRC | 09:31 | |
| *** rohitksingh_work has quit IRC | 09:41 | |
| *** rohitksingh has joined #yosys | 10:09 | |
| *** kraiskil has quit IRC | 10:41 | |
| pepijndevos_ | ZirconiumX, I wonder if it'd be feasible to synthesize some of these things http://countercomplex.blogspot.com/2011/10/some-deep-analysis-of-one-line-music.html would be more fun than a blinky led. | 10:45 |
|---|---|---|
| tpb | Title: countercomplex: Some deep analysis of one-line music programs. (at countercomplex.blogspot.com) | 10:45 |
| ZirconiumX | pepijndevos_: Probably, as long as you don't invoke something like sin() | 10:47 |
| pepijndevos_ | Lots of 32 bit multiplications get thrown around though, so you'd have to be careful what you select indeed. | 10:50 |
| ZirconiumX | I did look at the 74 series a bit more | 10:51 |
| ZirconiumX | There were multiplier ICs | 10:51 |
| ZirconiumX | The '261, '274 and '275, for example | 10:52 |
| ZirconiumX | Or the 284/285 | 10:52 |
| *** kraiskil has joined #yosys | 10:53 | |
| pepijndevos_ | Are they still being made though? | 10:53 |
| ZirconiumX | I doubt it :P | 10:53 |
| pepijndevos_ | I've been thinking if copy-pasting a booth multiplier from somewhere that uses our adder chips would be better than the xor mess that currently gets generated | 10:54 |
| ZirconiumX | I'm reasonably sure the Booth multiplier is synchronous, not combinational | 10:55 |
| ZirconiumX | You'd want a Wallace/Dadda tree | 10:55 |
| *** rohitksingh has quit IRC | 10:56 | |
| pepijndevos_ | Ah, probably true. It's been a while since university made us write out logical algorithms. | 10:57 |
| *** proteusguy has quit IRC | 11:03 | |
| *** Cerpin has joined #yosys | 11:10 | |
| *** adjtm has quit IRC | 11:11 | |
| *** proteusguy has joined #yosys | 11:15 | |
| pepijndevos_ | I have a whole list here, but nothing so far that is interesting and small https://github.com/feilipu/avrfreertos/blob/master/GoldilocksAnalogue/music_formula_collection.txt | 11:16 |
| tpb | Title: avrfreertos/music_formula_collection.txt at master · feilipu/avrfreertos · GitHub (at github.com) | 11:16 |
| *** rohitksingh has joined #yosys | 11:17 | |
| *** anuejn has joined #yosys | 11:17 | |
| *** rrika has quit IRC | 12:33 | |
| *** rrika has joined #yosys | 12:34 | |
| *** pie_ has quit IRC | 13:43 | |
| *** alexhw has quit IRC | 13:59 | |
| *** alexhw has joined #yosys | 14:00 | |
| *** Laksen has joined #yosys | 14:02 | |
| *** kraiskil has quit IRC | 14:10 | |
| *** kraiskil has joined #yosys | 14:22 | |
| pepijndevos_ | ZirconiumX, you know what would be usefull... if we could tell yosys to abuse spare pins for otherwise suboptimal arrangements. I currently have a situation where it makes 7 muxes and one and gate which could have been a mux with a fixed input. But I suppose that's just a distant dream... | 14:34 |
| ZirconiumX | How big are the muxes? | 14:35 |
| pepijndevos_ | Uh, in what sense? | 14:35 |
| ZirconiumX | If I'm truthful here I want to see what happens if you can source '150 16-input muxes | 14:35 |
| ZirconiumX | Input count | 14:35 |
| ZirconiumX | For example, that works with mux4s but not mux2s | 14:36 |
| pepijndevos_ | They are 4x1mux2 | 14:36 |
| ZirconiumX | Then you can't build an AND gate from it | 14:37 |
| ZirconiumX | I think post-Yosys would need a unification pass, to see which muxes could be merged | 14:38 |
| pepijndevos_ | Hrm, then I don't understand what is going on. Never mind then. For one port it has a bunch of muxes and an AND gate connected to the same signals. | 14:39 |
| ZirconiumX | Can you sketch out a diagram? | 14:39 |
| pepijndevos_ | Yosys can :) https://imgur.com/a/bmp4ONT The polarity signal on the left drives these muxes and the and gate which control level on the right | 14:41 |
| tpb | Title: Imgur: The magic of the Internet (at imgur.com) | 14:41 |
| *** SpaceCoaster has joined #yosys | 14:42 | |
| pepijndevos_ | Basically what I'm trying to do is make something that counts up and down efficiently. | 14:43 |
| tnt | pepijndevos_: over how many bits ? | 14:48 |
| *** SpaceCoaster has quit IRC | 14:48 | |
| tnt | pepijndevos_: I mean up/down counter is really just the adder chip with the A input wired to its output (registered) and the B input wired to a signal with LSB fixed to 1 and all the other bits being the counting direction. | 14:50 |
| pepijndevos_ | Right, buuut... ok so what I'm trying to do is make a LED fade in and out. So I have an inner PWM loop and an outer level loop. What I have now works, but I'm tweaking things to reduce chip count. | 14:53 |
| *** vonnieda has joined #yosys | 14:53 | |
| pepijndevos_ | oh this looks promessing | 14:54 |
| pepijndevos_ | 9 chips, so sexy. | 14:55 |
| pepijndevos_ | The sad thing it uses a whole chip for *one* XOR gate, but that seems a fair price for so much awesomeness. | 14:56 |
| *** adjtm has joined #yosys | 14:59 | |
| tnt | does 9 chip also include the actual pwm ? or is it just outputing and N bit value ? (and how much is N) ? | 15:00 |
| *** m4ssi has quit IRC | 15:03 | |
| pepijndevos_ | tnt, it does the actual PWM. So there are 4x4 adders that count up a 16 bit number that is stored in 2x8 flip-flops, then there are 2x4 comparators that compare the lower 7 to the upper 7 bits and an XOR on the MSB that switches the polarity of the PWM every other run | 15:06 |
| pepijndevos_ | So now updating the Kicad generator to handle the XOR chip, and I should be all set to make the most elaborate useless PCB ever. | 15:08 |
| *** maikmerten has joined #yosys | 15:10 | |
| *** kraiskil has quit IRC | 15:14 | |
| *** adjtm has quit IRC | 15:20 | |
| *** emeb has joined #yosys | 15:23 | |
| *** kraiskil has joined #yosys | 15:25 | |
| *** Laksen has quit IRC | 15:34 | |
| *** kraiskil has quit IRC | 15:39 | |
| pepijndevos_ | ZirconiumX, https://github.com/ZirconiumX/74xx-liberty/pull/3#issuecomment-506405866 <3 | 16:06 |
| tpb | Title: Generate KiCad netlists by pepijndevos · Pull Request #3 · ZirconiumX/74xx-liberty · GitHub (at github.com) | 16:06 |
| ZirconiumX | pepijndevos_: That could probably be manually routed, I think :P | 16:07 |
| pepijndevos_ | It could be... you could also design it completely by hand. | 16:08 |
| *** rohitksingh has quit IRC | 16:09 | |
| *** kraiskil has joined #yosys | 16:10 | |
| pepijndevos_ | Honestly, I'm not sure the autorouter is good enough to route a RiscV... | 16:12 |
| pepijndevos_ | There is an issu on Skidl to support Eagle libraries and netlists, but no actual work towards it I think. Let alone Althium or Orcad. | 16:13 |
| ZirconiumX | pepijndevos_: I think PCB of the gEDA suite has a topological router, but I don't know how good it is | 16:14 |
| ZirconiumX | As for manual routing; I think that's perfectly fine | 16:16 |
| *** dys has joined #yosys | 16:27 | |
| *** proteusguy has quit IRC | 16:32 | |
| *** rohitksingh has joined #yosys | 16:42 | |
| *** proteusguy has joined #yosys | 16:51 | |
| *** pie_ has joined #yosys | 16:59 | |
| *** fsasm has joined #yosys | 17:09 | |
| *** adjtm has joined #yosys | 17:34 | |
| *** AlexDaniel has joined #yosys | 18:09 | |
| *** pie_ has quit IRC | 19:04 | |
| *** pie_ has joined #yosys | 19:08 | |
| *** pie_ has quit IRC | 19:09 | |
| *** pie_ has joined #yosys | 19:09 | |
| *** pie_ has joined #yosys | 19:11 | |
| *** rohitksingh has quit IRC | 19:30 | |
| *** maikmerten has quit IRC | 20:25 | |
| *** kraiskil has quit IRC | 20:31 | |
| *** emeb_mac has joined #yosys | 20:55 | |
| *** AlexDaniel has quit IRC | 21:31 | |
| *** AlexDaniel has joined #yosys | 21:41 | |
| *** AlexDaniel has quit IRC | 21:59 | |
| benreynwar | Has any tried using removing irrelevant logic from a design before doing formal testing? I'm having the problem that writing out the VCD trace is the limiting factor. I thought maybe I could run yosys over the design first to strip out all the logic that can't possibly effect any assert or assume statements, and then run the formal testing on the stripped down version. Most of the logic (and state) in the design is in the | 21:59 |
| benreynwar | data path, while what I'm really trying to test with the formal testing is the control logic. Is this a sensible idea, and any suggestions about how I should go about it? | 21:59 |
| ZipCPU | Hello, benreynwar! | 22:03 |
| daveshah | benreynwar: anything sufficiently remote from asserts or assumes should be removed by a call to `opt`, or perhaps even just `opt_clean -purge`, after prep | 22:03 |
| benreynwar | ZipCPU: Hi again! | 22:06 |
| benreynwar | daveshah: Thanks for the pointer. I'm going to go and try to understand what going on behind the symbiyosys scripts and then I expect I'll be back with more questions. | 22:07 |
| *** fsasm has quit IRC | 22:34 | |
| *** vonnieda has quit IRC | 22:35 | |
| *** cr1901 has joined #yosys | 23:19 | |
| *** vonnieda has joined #yosys | 23:36 | |
Generated by irclog2html.py 2.13.1 by Marius Gedminas - find it at mg.pov.lt!