<awygle>
personally i feel like there is no way to validate your suspicion because there's no actual implementation of current VHDL or SystemVerilog to compare against
<awygle>
i also have a bias against languages that compile through other languages (in this case, verilog), but that's, again, a personal problem
<qu1j0t3>
i have a bias toward them :D
<openfpga-github>
[openfpga] rqou pushed 3 new commits to master: https://git.io/v5FrH
<openfpga-github>
openfpga/master 5f622e6 Robert Ou: yosys-netlist-json: Derive Hash where useful
<openfpga-github>
openfpga/master a7b22c0 Robert Ou: xc2bit: Make port names in structure.rs 'static
<openfpga-github>
openfpga/master 59ea442 Robert Ou: yosys-netlist-json: Fix port_directions on cells
azonenberg_mobil has quit [Remote host closed the connection]
azonenberg_mobil has joined ##openfpga
<rqou>
hey azonenberg_mobil: xc2par supports registers now (kinda)
<rqou>
the bare minimum needed to make a clocked AND gate work
<azonenberg_mobil>
lol yay
<azonenberg_mobil>
that's a start
<azonenberg_mobil>
also I'm flying over canada. looking at scattered town lights on the ground under me
<azonenberg_mobil>
it's weirdly comforting to know those lights are peaceful homes and not raging fires like when i fly over Washington this time of year 😛
<rqou>
btw i'm thinking about making a big refactoring of the xc2bit code
<rqou>
replacing the current inline fixed-size array with a not-inline not-fixed-size vector
<rqou>
it turns out that the current "everything inline" data structure is really annoying to construct
<rqou>
i'm also thinking of making the "dummy zia buffers feeding into the AND array" a first-class object
<rqou>
rather than every piece of code having a separate hack for that
<azonenberg_mobil>
yes that makes sense. that's how I'd do it
<azonenberg_mobil>
same as gp_clkbuf etc. makes routing way simpler
GenTooMan has quit [Quit: Leaving]
digshadow has quit [Quit: Leaving.]
digshadow1 has joined ##openfpga
azonenberg_mobil has quit [Remote host closed the connection]
azonenberg_mobil has joined ##openfpga
enriq has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
enriq has joined ##openfpga
digshadow1 has quit [Ping timeout: 248 seconds]
azonenberg_mobil has quit [Remote host closed the connection]
azonenberg_mobil has joined ##openfpga
digshadow has joined ##openfpga
noobineer has quit [Remote host closed the connection]
noobineer has joined ##openfpga
digshadow1 has joined ##openfpga
digshadow has quit [Read error: Connection reset by peer]
digshadow has joined ##openfpga
digshadow1 has quit [Ping timeout: 240 seconds]
enriq has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
enriq has joined ##openfpga
Marex has joined ##openfpga
enriq has quit [Client Quit]
azonenberg_mobil has quit [Remote host closed the connection]
azonenberg_mobil has joined ##openfpga
pie_ has joined ##openfpga
pie_ has quit [Ping timeout: 240 seconds]
<rqou>
azonenberg_mobil: ok, i implemented enough of xc2par that you should in theory be able to use it on "real" designs now
<rqou>
of course, it's *) not useful without LOC and *) probably doesn't work on any actual designs
<openfpga-github>
[openfpga] rqou opened pull request #115: xc2par: What on earth are we going to do about this mess of a PR? (master...xc2par) https://git.io/v5F9X
<rqou>
azonenberg_mobil: why is it so difficult to "just hand xbpar two graphs" and get an answer back? *complain complain*
<azonenberg_mobil>
Feel free to write a better PAR :)
<azonenberg_mobil>
maybe if awygle makes progress on that analytic placer we can make a cpld version
<azonenberg_mobil>
I'd love to try
<rqou>
i want a better PAR interface
<rqou>
that doesn't need this whole subclass of Engine
<rqou>
i also want a plotting tool that doesn't explode when given too many edges
<azonenberg_mobil>
the latter would be great. becausr i need that for my netlist re too
<azonenberg_mobil>
we need a good, scalable graph rendering tool lol. graphviz is waay too slow
<rqou>
apparently most data science types don't deal with graphs that have on the order of a million edges :P
<azonenberg_mobil>
ML people probably do. but they don't render them
<rqou>
yeah
<azonenberg_mobil>
and hmm i was gonna say facebook
Hootch has joined ##openfpga
<azonenberg_mobil>
but they have lots of nodes with relatively low degree
<rqou>
i don't think FB ever renders an entire graph
<rqou>
i think they only locally render parts of things if at all
<azonenberg_mobil>
yes. but they do have some rendering tools
<rqou>
although our graphs are much more structured than a normal graph
<rqou>
it's just that xdot doesn't know that
<azonenberg_mobil>
what i meant is social graphs are high vertex count
<azonenberg_mobil>
but fairly low degree
<rqou>
yeah, we have the opposite
<azonenberg_mobil>
we should try to make a tool that a) is aware of clocks etc and sanely handles them, with lines fanning out in a tree vs a zillion separate directrd edges
<rqou>
also, complaining again: why is LOC not an inherent feature of the common PAR logic?
<azonenberg_mobil>
and b) maybe can take hints about initial placement
<azonenberg_mobil>
good question. one more way I'd do it bettet if starting fresh lol
<rqou>
also, why does the caller have to find bad edges? why can't the common engine do that?
<rqou>
btw, you should at least open my PR
<rqou>
just so you can see the insanity :P
<rqou>
although interestingly all of xc2par is apparently only 4 kLOC
teepee has quit [Ping timeout: 252 seconds]
teepee has joined ##openfpga
azonenberg_mobil has quit [Remote host closed the connection]
azonenberg_mobil has joined ##openfpga
azonenberg_mobil has quit [Remote host closed the connection]
teepee has quit [Ping timeout: 240 seconds]
teepee has joined ##openfpga
pie_ has joined ##openfpga
pie_ has quit [Ping timeout: 240 seconds]
qu1j0t3 has quit [Ping timeout: 264 seconds]
massi has joined ##openfpga
azonenberg_mobil has joined ##openfpga
<azonenberg_mobil>
nearing the coast of england now. still just see water
<azonenberg_mobil>
looooong flight lol
<sn00n>
hii
qu1j0t3 has joined ##openfpga
azonenberg_mobil has quit [Remote host closed the connection]
azonenberg_mobil has joined ##openfpga
teepee has quit [Ping timeout: 258 seconds]
teepee has joined ##openfpga
azonenberg_mobil has quit [Ping timeout: 255 seconds]
massi has quit [Remote host closed the connection]
<awygle>
Should talk to data visualization people about graph rendering. Go to d3.js meet ups or something.
<awygle>
Sleep is hard
teepee has quit [Ping timeout: 248 seconds]
teepee has joined ##openfpga
pie_ has joined ##openfpga
enriq has joined ##openfpga
xdeller has quit [Quit: Leaving]
teepee has quit [Ping timeout: 252 seconds]
teepee has joined ##openfpga
m_t has joined ##openfpga
azonenberg_work has joined ##openfpga
azonenberg_work1 has joined ##openfpga
azonenberg_work has quit [Disconnected by services]
azonenberg_work1 is now known as azonenberg_work
DocScrutinizer05 has quit [Disconnected by services]
DocScrutinizer06 has joined ##openfpga
DocScrutinizer06 is now known as DocScrutinizer05
DocScrutinizer05 has quit [Disconnected by services]
DocScrutinizer05 has joined ##openfpga
juri_ has joined ##openfpga
enriq has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
azonenberg_work has quit [Ping timeout: 255 seconds]
DocScrutinizer05 has quit [Disconnected by services]