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
anticw has quit [Remote host closed the connection]
anticw has joined #yosys
FabM has joined #yosys
FabM has quit [Changing host]
FabM has joined #yosys
famubu[m] has quit [Quit: Idle timeout reached: 172800s]
Wolfvak has quit [Quit: ZNC - https://znc.in]
Wolfvak has joined #yosys
kristianpaul has quit [Ping timeout: 258 seconds]
kristianpaul has joined #yosys
Adrien[m] has joined #yosys
<Adrien[m]> Hi ! I am experimenting routing for Zynq target, using nextpnr-himbeachel instead of nextpnr-xilinx (see https://github.com/YosysHQ/nextpnr/pull/1352 BTW)
<Adrien[m]> But now I'm facing an error which suggests nextpnr-himbeachel for Xilinx arch is no "feature-parity" with nextpnr-xilinx, is that true ?
<Adrien[m]> ERROR: Unable to find legal placement for cell 'user_inst.$abc$4612$auto$blifparse.cc:535:parse_blif$4646.genblk1.genblk1.genblk1.genblk1.genblk1.genblk1.genblk1.genblk1.genblk1.genblk1.mux8' after 53302 attempts, check constraints and utilisation.
flokli has quit [Ping timeout: 252 seconds]
<Adrien[m]> How is support for mux8 ?
<Adrien[m]> Hum weird it's `mux8` and not `muxf8`
<Adrien[m]> Maybe mux8 is a sort of macro generated by yosys and that hibaechel does not know it's a slice
flokli has joined #yosys
FabM has quit [Ping timeout: 265 seconds]
<lofty> Adrien[m]: indeed, the himbächel port is not yet at parity with the original -xilinx. But rather importantly: the himbächel port is in upstream, and -xilinx never made it :p
<Adrien[m]> For now, it seems using -nowidelut in yosys makes the thing placeable by himbaechel
<Adrien[m]> Aud routable !
<Adrien[m]> Except for an assertion after final report of slack histogram
<Adrien[m]> Argh ! its the dreaful IOI SING thingy
<Adrien[m]> Similar issue than for somlo then
<lofty> Adrien[m]: there *is* code to handle MUXF8
<Adrien[m]> Other than that, I'm so happy to see the device database files that are really small, good factorization, thanks authors 👍️ <3
<lofty> Adrien[m]: that's one of the big selling points of himbaechel
<Adrien[m]> lofty yes, but there is not muxf8 directly, rather what I call a macro, `mux8`
<Adrien[m]> which can be expanded into one slice with luts and muxf7+muxf8 indeed at any step in the flow, just it seems that expansion is the missing piece.
<Adrien[m]> drawback of -nowideluts in yosys is that muxf7/8 don't seem to be used at all, there are only luts.
<lofty> that's because -nowidelut disables use of those
<lofty> but like
<lofty> Yosys should be reporting MUXF7/MUXF8 in the output
greeb has left #yosys [WeeChat 4.3.6]
nonchip has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
nonchip has joined #yosys