cr1901 has quit [Read error: Connection reset by peer]
cr1901 has joined #yosys
lexano_ has joined #yosys
emeb_mac has joined #yosys
X-Scale` has joined #yosys
X-Scale has quit [Ping timeout: 256 seconds]
X-Scale` is now known as X-Scale
<shorne>
lkcl: I have run into that, we have MMU when we program mappings the 'way' address encoding if specified wrong will fail to update the MMU mapping
<shorne>
I have just added assumptions to clarify what is a valid way address
<shorne>
But formal testing is only as good as my definition of what is "valid"
<shorne>
But it is very quick to uncover what I have done wrong, speaking of which, I do a lot of my development over a ssh connection. But when I do digital design I always have to be at my workstation to view waveforms with gtkwave
<shorne>
does anyone have a good command line alternative for waveform viewing? I have seen a few but nothing that I can use yet
emeb_mac has quit [Quit: Leaving.]
<lkcl>
claire wolfe wrote one of the first formal correctness proofs for RISC-V cores, and it turned out to be extraordinarily valuable in finding horribly-obcscure bugs
<lkcl>
my favourite one i heard about - scuse me this was several years ago already so may be slightly mis-remembered - was an INT divide instruction affecting the results of FP
<lkcl>
something like that, and it turned out to be that because the DIV was done as a FSM, a register was left active, which then bled information back into certain instructions
<lkcl>
no unit test would ever find that!
<lkcl>
if you ever find a good command-line waveform viewer do let me know
evantandersen has joined #yosys
<evantandersen>
hey, I heard there might be someone here that could improve the performance of yosys/nextpnr? we're a commercial entity currently targeting the ice40 LP1K, icecube currently manages to fit our design into 1000 LUTs, open source is at 1230 (yosys -abc9/nextpnr)
<evantandersen>
thats a ~20% overhead
<evantandersen>
we can offer some money if that's the limiting factor
<evantandersen>
add me on discord if you're interested! charliehorse55#2689
evantandersen has quit [Ping timeout: 256 seconds]