| *** tpb <[email protected]> has joined #yosys | 00:00 | |
| lofty | The `cook` code confuses me; I realise it's for counting the number of unique inputs, but if you keep the LUT inputs in a cut sorted then deduplication is just std::set_union | 00:03 |
|---|---|---|
| *** bjorkintosh <bjorkintosh!~bjork@2600:1700:5400:c80:b0c3:a0a9:f298:8> has joined #yosys | 01:04 | |
| *** notgull <notgull!~notgull@ec2-50-112-148-23.us-west-2.compute.amazonaws.com> has quit IRC (Ping timeout: 246 seconds) | 02:18 | |
| *** notgull <notgull!~notgull@2600:1f14:8f4:3100:e2a2:c920:847d:d23> has joined #yosys | 02:20 | |
| *** FabM <[email protected]> has joined #yosys | 06:16 | |
| povik | you mean because of the 3 of the last 4 testcases? | 06:57 |
| povik | there are other explanations, toymap and abc might not be working on exactly the same aig | 06:58 |
| povik | also there's an edge case where toymap needs to connect a PO to the rest of the network with an inverter, making the depth 1 greater than what the cut mapper comes up with | 07:01 |
| povik | (this is a quirk of the internal representation for now) | 07:01 |
| povik | that can't explain random_control/priority.aig | 07:02 |
| povik | i think i should do proper hashing of the inputs in a cut, instead of adding the pointers :p | 07:03 |
| lofty | povik: I doubt the problem is that you're working on different AIG; I've been frustrated by this issue before and the problem really was a lack of cut diversity | 07:08 |
| povik | ok, the improper cut hashing is what i suspect first | 07:09 |
| lofty | But it's easy enough to test if the problem is that it's working on wrong aig | 07:10 |
| lofty | The input and output netlists must be equivalent, so you can so `equiv_opt toymap` or whatever | 07:11 |
| povik | yeah, that passes last time i checked | 07:12 |
| povik | but there's no guarantee `abc -g aig` gives the same graph as abc uses for lut mapping | 07:12 |
| lofty | whitequark[cis]: beginning to wonder if that command should be renamed `equiv_pass` instead | 07:12 |
| lofty | Or something | 07:13 |
| lofty | povik: Yosys can read aiger files | 07:13 |
| lofty | So use them directly and operate directly on that | 07:13 |
| lofty | I'm pretty sure that the EPFL benchmarks have already been optimised enough that the ABC opts don't do anything | 07:16 |
| lofty | With the point of there being no low hanging fruit | 07:16 |
| povik | not true actually | 07:16 |
| povik | %fraig -x does tremendous cutting down of div.aig | 07:17 |
| povik | and that's one of the few i looked into | 07:17 |
| povik | err, &fraig -x | 07:17 |
| lofty | Okay, but the reference numbers have *no* optimisation | 07:18 |
| lofty | And you still can't match ABC | 07:18 |
| povik | you mean what i use for abc baseline? | 07:18 |
| povik | that's *including* the optimisation | 07:18 |
| lofty | The official numbers given by EPFL. | 07:19 |
| povik | ah | 07:19 |
| povik | haven't checked those | 07:19 |
| lofty | https://github.com/lsils/benchmarks | 07:19 |
| lofty | Those are what you should compare against, IMO | 07:19 |
| lofty | Otherwise you're making things more difficult for yourself | 07:19 |
| povik | i have no issue with that | 07:20 |
| povik | i'm interested in implementing some degree of aig preprocessing into toymap | 07:20 |
| lofty | Like, maybe ABC is discovering structural choices while optimising; those can't cross the ABC/Yosys border | 07:20 |
| lofty | A lot of AIG preprocessing algorithms start with the netlist mapped into cuts anyway | 07:21 |
| povik | 09:20 < lofty> Like, maybe ABC is discovering structural choices while optimising; those can't cross the ABC/Yosys border | 07:22 |
| povik | one think i want to check is whether abc uses *any* structural choices anywhere in the epfl benchmark | 07:22 |
| povik | i find out what i need to look for in the `if -v` output | 07:22 |
| povik | s/i find out/i found out/ | 07:22 |
| lofty | I tend to use &if over if simply because I'm a snob /j | 07:23 |
| povik | so far it seems more like a legend of "abc does structural choices! you can't match that" instead of something that actually makes a difference | 07:23 |
| povik | but maybe i will be proven wrong | 07:24 |
| lofty | What optimisation passes are you using? | 07:24 |
| povik | the default yosys script | 07:24 |
| lofty | `abc` or `abc9`? | 07:25 |
| povik | the blank variety | 07:25 |
| povik | err, by that i mean `abc` | 07:25 |
| lofty | Hmm. There *are* calls to dc2 and dch -f | 07:26 |
| lofty | So there *should* be choices generated | 07:26 |
| povik | i will let you know what i find out | 07:26 |
| lofty | I will be curious to see how you might represent choices, since it's something I've discussed with the others in the past | 07:31 |
| *** ec_ <ec_!~ec@gateway/tor-sasl/ec> has quit IRC (Remote host closed the connection) | 08:21 | |
| *** ec_ <ec_!~ec@gateway/tor-sasl/ec> has joined #yosys | 08:21 | |
| *** nak <nak!~nak@yosys/nak> has joined #yosys | 09:04 | |
| *** indy <[email protected]> has quit IRC (Quit: ZNC 1.8.2 - https://znc.in) | 12:38 | |
| *** lexano <[email protected]> has joined #yosys | 12:38 | |
| *** indy_ <[email protected]> has joined #yosys | 12:42 | |
| *** josuah <[email protected]> has joined #yosys | 13:56 | |
| *** krispaul <[email protected]> has quit IRC (Quit: WeeChat 3.5) | 14:56 | |
| *** kristianpaul <kristianpaul!~paul@user/kristianpaul> has joined #yosys | 14:56 | |
| *** hrberg <[email protected]> has joined #yosys | 15:04 | |
| *** FabM <FabM!~FabM@armadeus/team/FabM> has quit IRC (Ping timeout: 248 seconds) | 15:05 | |
| *** lkcl <[email protected]> has quit IRC (Ping timeout: 245 seconds) | 17:31 | |
| *** knielsen <[email protected]> has left #yosys | 17:50 | |
| *** V <V!~v@ircpuzzles/2022/april/winner/V> has quit IRC (Ping timeout: 246 seconds) | 17:51 | |
| *** indy_ is now known as indy | 17:59 | |
| whitequark[cis] | lofty: no objection to renaming | 18:31 |
| *** lkcl <[email protected]> has joined #yosys | 18:37 | |
| *** V <V!~v@ircpuzzles/2022/april/winner/V> has joined #yosys | 19:25 | |
| *** V <V!~v@ircpuzzles/2022/april/winner/V> has quit IRC (Ping timeout: 240 seconds) | 19:55 | |
| *** V <V!~v@ircpuzzles/2022/april/winner/V> has joined #yosys | 20:04 | |
| *** V <V!~v@ircpuzzles/2022/april/winner/V> has quit IRC (Ping timeout: 244 seconds) | 20:08 | |
| *** V <V!~v@ircpuzzles/2022/april/winner/V> has joined #yosys | 20:23 | |
| *** Guest22 <[email protected]> has quit IRC (Quit: Client closed) | 21:37 | |
| *** lumo_e <[email protected]> has joined #yosys | 22:08 | |
| *** nonchip <[email protected]> has quit IRC (Quit: https://quassel-irc.org - Chat comfortably. Anywhere.) | 22:17 | |
| *** nonchip <[email protected]> has joined #yosys | 22:17 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!