whitequark changed the topic of #glasgow to: glasgow debug tool · code https://github.com/whitequark/Glasgow · logs https://freenode.irclog.whitequark.org/glasgow
emily has joined #glasgow
emily is now known as Guest22571
Guest22571 has quit [Client Quit]
emily__ has joined #glasgow
emily__ has quit [Client Quit]
emily has joined #glasgow
emily is now known as Guest35520
Guest35520 has quit [Client Quit]
emily__ has joined #glasgow
emily__ has quit [Client Quit]
emily has joined #glasgow
sorear has joined #glasgow
yorick has joined #glasgow
zkms has joined #glasgow
zkms has quit [Client Quit]
zkms has joined #glasgow
zkms has quit [Client Quit]
zkms has joined #glasgow
<whitequark> gruetzkopf: so, the next step would be, replacing all the flags currently used with PF and a configured threshold
<whitequark> with the expected no functionality change
_whitenotifier-c has quit [Remote host closed the connection]
<Hellsenberg> building nextpnr with 8 GB of RAM... fail
<Hellsenberg> I run out of memory and I don't have swap on this machine
pke has quit [Ping timeout: 245 seconds]
<tnt> Hellsenberg: for which architecture ?
<tnt> ah well, ice40 I guess given the channel :p I thought this was ##openfpga ...
<tnt> 8G should be plenty ... assuming you don't have like a 6G chrome running in // of course.
<tnt> make sure you don't use parallel build.
pke has joined #glasgow
<gruetzkopf> 12G is definity enough for -j4
pke has quit [Ping timeout: 246 seconds]
pke has joined #glasgow
<Hellsenberg> tnt: noted, thanks
<Hellsenberg> I was using a -j8 on 8G of RAM, with the OS using much less
<tnt> yeah, one of the build step is making the 'bba' binary database and these are a couple of Gig each so building them in // won't work.
<tnt> I wonder if there is a way to tell cmake that those targets shouldn not be built in //.
<Hellsenberg> I think I am building on a tmpfs too ._.
<Hellsenberg> yes, in /tmp
<Hellsenberg> if not I can build on poweredgy with 64GB of RAM
pke has quit [Ping timeout: 240 seconds]
pke has joined #glasgow
pke has quit [Ping timeout: 250 seconds]
pke has joined #glasgow
pke has quit [Read error: Connection reset by peer]
<Hellsenberg> okay, only building nextpnr for ice40 and not ecp5 helps
<Hellsenberg> now to try to make it work
<Hellsenberg> I should print a pinout for the board
pke has joined #glasgow
pke has quit [Ping timeout: 240 seconds]
pke has joined #glasgow
_whitelogger has quit [Remote host closed the connection]
_whitelogger_ has joined #glasgow
pke has quit [Remote host closed the connection]
pke has joined #glasgow
pke has quit [Ping timeout: 250 seconds]
_whitelogger has joined #glasgow
<gruetzkopf> whitequark: as far as i could guesstimate on the device i had at hand, timings looked identical
<whitequark> gruetzkopf: thanks
<whitequark> please proceed with the next step above
<gruetzkopf> i'll try moving all to PF and report again
<Yatekii> question: do you folks focus more on the HW or the SW part (considering the IP core rather to be SW) or both alike? I mean is it important to you to have a nice UI or do you rather want to have nice HW which you then for example plug to sigrok?
<whitequark> both hardware and software are essential to glasgow
<whitequark> sigrok is very limited
<whitequark> e.g. it can't display multibit signals as buses
<whitequark> having a nice UI is important, and I consider the existing CLI fairly nice
<whitequark> there ought to be a GUI on some point but for now I focus on basics
<whitequark> still, the internals anticipate having a GUI
<Yatekii> ok, cool! thx
<Yatekii> really looking forward to this :) I'll definitely get one