AlexDaniel has quit [Read error: Connection reset by peer]
AlexDaniel has joined #yosys
ZipCPU|Laptop has quit [Remote host closed the connection]
<edmoore>
shapr: how exciting to be chums with hillel
ZipCPU|Laptop has joined #yosys
<edmoore>
He seems like a jolly good thing
<shapr>
edmoore: I was chilling with Dan Piponi the night before
<edmoore>
What a happy thing
<nats`>
plop is there a full chain for ultra plus 5 based on yosys (with timing analysis ?)
<shapr>
edmoore: I was telling him about the work you're doing with rockets
<edmoore>
Oh fun
<edmoore>
Another interesting person in tech
<shapr>
edmoore: yeah, I was showing Hillel some contact juggling tricks
<awygle>
nats`: "ultra plus 5"? if you just mean ice40 ultra plus, yes
<awygle>
timing analysis isn't _great_ though
<awygle>
i mean it's fine but it doesn't understand clock domains for example
<nats`>
yep ice40 up5
<awygle>
oh up5k, sure
<nats`>
oky I can't take care of crossing
<nats`>
I have a small design to fit in (pretty fast one anyway) and I was thinking yosys or hell paved way of icecube2
<nats`>
s/can't/can/
<awygle>
as long as you aren't running huge logic in a 10 MHz domain and fast logic in a 100 MHz domain or something like that, you'll be okay
<shapr>
edmoore: from what you've said before, it sounds like you're looking into spare tricks for simulation, much like the graphics sparsity talk Dan gave.
<shapr>
I'm not sure if the grammar for that sentence works correctly
<nats`>
awygle, what do you mean ?
<nats`>
if those two domain are synced by a CC fifo is there a problem ?
<awygle>
nats`: if you have logic which can only run at (say) 11 MHz anywhere on your chip, icetime will just say "hey your fMax is 11 MHz"
<shapr>
edmoore: "sparse simulation" sorry
<awygle>
which means if you have logic with an fMax of 90 MHz someplace that expects to do 100 MHz, you won't see it
<nats`>
it doesn't analyze independant path ?
<awygle>
it does, but i think it only reports like 10? actually you might be okay if you analyze the report yourself
<nats`>
I guess I'll have to go through the hell paved way of icecube2 on linux -_-
<nats`>
ahh I can take care of reading it yes
<edmoore>
shapr: im not completely sure, tho interested as always
<awygle>
gimme 2m and i'll make sure it reports all the paths
<edmoore>
one man's adaptive mesh is another man's sparse sampling, as the old chinese proverb says
<nats`>
good viterbi at you edmoore
<edmoore>
i'll pass on the message, nats`
<nats`>
:D
<awygle>
nats`: so, i can't actually get icetime to print the timing values for the whole netlist. it just shows the critical path, or you can specifically request a given net. might be worth playing with it a bit to see if it gives you what you want
<awygle>
if not, you can always do timing analysis in icecube and everythign else in yosys and arachne or nextpnr
<nats`>
ok I'll try to find information on that :)
<nats`>
you use the github image daily ?
pie_ has quit [Ping timeout: 264 seconds]
ZipCPU|Laptop has quit [Remote host closed the connection]
AlexDaniel has quit [Read error: Connection reset by peer]
AlexDaniel has joined #yosys
AlexDaniel has quit [Read error: Connection reset by peer]
AlexDaniel has joined #yosys
ZipCPU|Laptop has joined #yosys
<nats`>
is it better to stick with arachne-pnr or nextpnr
pie_ has joined #yosys
<awygle>
Nextpnr is supposed to give better results
<awygle>
and yes I run on HEAD when I use the tools