whitequark changed the topic of #yosys to: Yosys Open SYnthesis Suite: https://github.com/YosysHQ/yosys/ | Channel logs: https://libera.irclog.whitequark.org/yosys/ | Bridged to #yosys:matrix.org
tpb has quit [Remote host closed the connection]
tpb has joined #yosys
Stary has quit [Quit: ZNC - http://znc.in]
Stary has joined #yosys
kristianpaul has quit [Read error: Connection reset by peer]
kristianpaul has joined #yosys
FabM has joined #yosys
FabM has joined #yosys
<Adrien[m]> Yes it does. when run without -wideluts.
<Adrien[m]> This `mux8` error is weird, I will look at the netlist.
<lofty> well, the error you posted was that it couldn't place a cell named "mux8"
kristianpaul has quit [Read error: Connection reset by peer]
kristianpaul has joined #yosys
<Adrien[m]> Debugging a bit, the cell type is `SELMUX2_1`
<Adrien[m]> I'm not sure what behaviour to expect from nextpnr, is it supposed to decide to place this in a LUT or in a MUXF location ?
<lofty> Adrien[m]: uuuh, I think this is where you have to show your source verilog, because that's not a valid cell
<Adrien[m]> Yes. Need to reduce the design to something reproducer-friendly then.
<Adrien[m]> Interestingly, nextpnr-xilinx does handle this design.
<lofty> okay, it apparently *is* a valid bel type, but hmm
<Adrien[m]> And I find no occurrence of selmux in input json !
<Adrien[m]> From the nextpnr-hibaechel log :... (full message at <https://catircservices.org/_matrix/media/v3/download/catircservices.org/TPDvTLLAzyRNgUAKnfhsVShr>)
<Adrien[m]> I'll reduce the design to something smaller
<Adrien[m]> maybe just a mux 8->1 is enough
<Adrien[m]> Had to do a mux 31->1 to ensure yosys generates a muxf7.... (full message at <https://catircservices.org/_matrix/media/v3/download/catircservices.org/aoqVEbTJfBVlsQzvUZlynAxq>)
<Adrien[m]> yosys -p "synth_xilinx -arch xc7 -top zedboard_top_mux; write_json zedboard_top_mux.json" zedboard_top_mux.v
<Adrien[m]> nextpnr-himbaechel --device xc7z020clg484-1 -o xdc=zedboard.xdc --json zedboard_top_mux.json -o fasm=zedboard_top_mux.fasm --router router1
<Adrien[m]> If tracking of this is pertinent given development status, I can open an issue, please just say.
<Adrien[m]> BTW I think this is also a Yosys sub-optimality
<Adrien[m]> For a mux 8->1 it generates 3 LUTs, instead of 2 LUT6 + 1 muxf7
<Adrien[m]> Does this deserve to open an issue for yosys ?
<lofty> Adrien[m]: what if you synth with -abc9
<lofty> actually
<lofty> you want -widemux 8 to infer that
Lord_Nightmare has quit [Quit: ZNC - http://znc.in]
Lord_Nightmare has joined #yosys
<Adrien[m]> lofty: most impactful option is indeed -widemux. not -abc9 for this topic.
<lofty> Because -widemux does exactly what you ask for here, but it does come with headaches
<Adrien[m]> ok good to know
skipwich_ is now known as skipwich
<agg> bit of a long shot, but i'm using an ecp5 MULT18X18D with input and pipeline registers enabled and both SIGNEDIA/SIGNEDIB inputs hardwired to 1, and it seems to not sign-extend the first multiplication after a reset
<agg> (with latest yosys/nextpnr-ecp5)
<agg> anyone have... any ideas
<agg> i haven't tried this on diamond so not entirely sure if it's specifically yosys/nextpnr or some cursed hardware thing though
<whitequark[cis]> that's really cursed
<agg> yea :( you can imagine it took a while it discover this was precisely what's going wrong too
<agg> if i turn off the MULT18X18D's input registers and replace them with my own input registers instead, everything seems to work fine...
ec has joined #yosys
<whitequark[cis]> i wonder if there's a secret SIGNEDA reset register
skipwich has quit [Quit: DISCONNECT]
skipwich has joined #yosys
skipwich_ has joined #yosys
skipwich has quit [Ping timeout: 252 seconds]
skipwich_ is now known as skipwich
skipwich_ has joined #yosys
FabM has quit [Ping timeout: 265 seconds]
skipwich has quit [Ping timeout: 252 seconds]
skipwich_ is now known as skipwich
skipwich_ has joined #yosys
skipwich has quit [Ping timeout: 264 seconds]
skipwich_ is now known as skipwich
skipwich_ has joined #yosys
skipwich has quit [Ping timeout: 248 seconds]
skipwich has joined #yosys
skipwich_ has quit [Ping timeout: 276 seconds]
skipwich_ has joined #yosys
skipwich has quit [Ping timeout: 258 seconds]
skipwich has joined #yosys
skipwich_ has quit [Ping timeout: 252 seconds]
kristianpaul has quit [Ping timeout: 245 seconds]
kristianpaul has joined #yosys
skipwich_ has joined #yosys
skipwich has quit [Ping timeout: 255 seconds]
skipwich has joined #yosys
kristianpaul has quit [Ping timeout: 248 seconds]
skipwich_ has quit [Ping timeout: 252 seconds]
kristianpaul has joined #yosys
skipwich_ has joined #yosys
skipwich has quit [Ping timeout: 252 seconds]
skipwich_ is now known as skipwich
kristianpaul has quit [Read error: Connection reset by peer]
kristianpaul has joined #yosys
skipwich_ has joined #yosys
skipwich has quit [Ping timeout: 272 seconds]
skipwich_ is now known as skipwich
nonchip has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
nonchip has joined #yosys
kristianpaul has quit [Quit: WeeChat 3.5]
kristianpaul has joined #yosys