00:02 <somlo> I just upgraded my workstation to F38, and suddenly my old Vivado 2018.2 setup is putting up a fight (I've successfully dragged it along for a good number of years now)

00:03 <somlo> what's a good recent version to use with LiteX (2018.2 worked great for me, it's just that all the system libraries it's looking for are too new on the underlying distro, and that's a pain to deal with)

14:31 <_florent_> somlo: I'm still on v2021.1, but any recent version should work.

14:41 <somlo> _florent_: thanks! I've started wrestling with 2022.2 (their latest). I'm also wrestling f38 (updating software is a right pain in the arse) :D

19:05 <gurki> using fedora for eda tools is quite bold to begin with :D

19:06 <gurki> we have a bunch of separate servers running as-ancient-as-possible-while-still-supported distros *g*

19:12 <somlo> gurki: I've been using fedora since forever; played with lowRISC and then LiteX for the last 5-ish years or so, never had any problems

19:13 <somlo> installed vivado 2018 and was mostly able to rsync it along (as part of my home directory) from f28 all the way to f36 :)

19:14 <somlo> but then Vivado is really the only proprietary thing I've been using natively (cadence and synopsys as part of my ece course work don't count, those are installed on the lab machines and I won't touch anything related to their sysadmin side with a barge pole :D )

19:41 <gurki> somlo: cadence only quite recently dropped solaris

19:41 <gurki> somlo: thats about everything that needs to be said about those two :D

19:41 <gurki> im obv happy when things work for you :)

19:43 <somlo> looking forward to the day when nextpnr will be able to handle 7-series xilinx chips (kintex, virtex) smoothly, then I can ditch vivado and not worry about its linker dependencies anymore either :D

19:44 <gurki> my honest estimate is "it will never properly support 7-series", but i dont mind being wrong

19:44 <gurki> the xilinx stuff has so much details and features and tier-specific stuff that its really hard to build tools for it

19:49 <tnt> yeah, for toy projects maybe, but I don't see that happening for something as complex as 7-series / ultrascale / ... and if I use those chips, there is a reason : I need their perf and features.

19:49 <tnt> When you look that the immensely simpler ice40 and ecp5 are not 100% supported ...

19:54 <somlo> I guess predictions are hard, especially about the future ;)

20:00 <tnt> Yes. I mean, in 50 years someone might do it as a retro computing project :)

20:03 <gurki> theres a bit of traction happening in the foss world, this might trickle down to the foss fpga stuff

20:03 <gurki> foss asic*