<TommyT>
It's Sunday and I figured I'd try out yosys and the iCE40 stick. It works brilliantly. However ...
<TommyT>
... I played around with the rot example and got a very currious result with the timing analyser. A simpler circuit times in at 169.89 MHz whereas a circuit with a two-bit longer adders reaches 245.30 MHz
<ZipCPU>
Wow, an iCE40 can do more than 150MHz at all?
<ZipCPU>
mithro: I've never used iverilog
<mithro>
TommyT: Take a look at the timing output network
<mithro>
ZipCPU: You use verilator?
<ZipCPU>
mithro: A little googling suggest that dumpfile isn't sufficient, you'll need dumpvars as well./
<ZipCPU>
mithro: Yes, I'm a very heavy verilator user.
pie__ has quit [Ping timeout: 264 seconds]
<ZipCPU>
Although ... now that I've gotten involved in Formal, I don't do quite so much with Verilator as I used to. ;)
<TommyT>
@mithro: fast version "Resolvable net names on path: 0.640 ns .. 1.229 ns divider[0] ... 3.270 ns .. 3.859 ns $auto$alumacc.cc:474:replace_alu$13.C[9]"
<TommyT>
Total number of logic levels: 11 Total path delay: 4.08 ns (245.30 MHz)
<TommyT>
The simpler/slower circuit: "Resolvable net names on path: 0.640 ns .. 0.640 ns divider[8] 3.964 ns .. 3.964 ns divider[8]$2 4.413 ns .. 5.886 ns $abc$204$n6"
<TommyT>
Total number of logic levels: 2 Total path delay: 5.89 ns (169.89 MHz)
m_t has quit [Quit: Leaving]
<TommyT>
What's the right way to dig deeper?
<mithro>
TommyT: What icetime are you running?
<TommyT>
This is freshly cloned from github an hour ago
<mithro>
TommyT: I meant what arguments are you giving it?
<TommyT>
I see, the -v doesn't mix with the other options
<mithro>
TommyT: Should give you a pretty detailed information on how it calculated the timing
<TommyT>
:) Thanks, I'd have to read up to understand that, but AFAICT, the routing is randomly worse for the simpler case. I guess that surprised me (I've never see such an extreme case with X or A)
<mithro>
TommyT: I guess with large cells, the difference between one and two tiles make make a big effect?
<TommyT>
TIL: giving arachne-pnr a -r can make a big difference. Now the simpler design hits 275 MHz. I guess I should also play with the seed
<mithro>
Clifford was complaining at me regarding that -- with arachne-pnr not being timing driven the seed really matters apparently
<TommyT>
Dramatically so apparently. Seeds 1..10 gets results all over [162;274] MHz!
promach_ has quit [Quit: WeeChat 2.1]
lutsabound has quit [Quit: Connection closed for inactivity]
lutsabound has joined #yosys
promach has quit [Quit: WeeChat 2.1-dev]
seldridge has quit [Ping timeout: 256 seconds]
TommyT has quit [Ping timeout: 252 seconds]
zino has quit [Ping timeout: 240 seconds]
promach has joined #yosys
lutsabound has quit [Quit: Connection closed for inactivity]
mwk has quit [Ping timeout: 264 seconds]
mwk has joined #yosys
emeb_mac has quit [Ping timeout: 264 seconds]
promach has quit [Quit: WeeChat 2.1-dev]
samayra has quit [Ping timeout: 245 seconds]
swick has quit [Ping timeout: 240 seconds]
marbler has quit [Ping timeout: 245 seconds]
fevv8[m] has quit [Ping timeout: 245 seconds]
eduardo_ has joined #yosys
promach has joined #yosys
fevv8[m] has joined #yosys
marbler has joined #yosys
samayra has joined #yosys
swick has joined #yosys
promach has quit [Quit: WeeChat 2.1-dev]
promach has joined #yosys
promach has quit [Quit: WeeChat 2.1-dev]
promach has joined #yosys
fsasm has joined #yosys
jwhitmore has joined #yosys
pie__ has joined #yosys
ar3itrary has quit [Quit: Be vigilant]
ar3itrary has joined #yosys
m_t has joined #yosys
promach has quit [Quit: WeeChat 2.1-dev]
AlexDani` has joined #yosys
AlexDaniel has quit [Ping timeout: 256 seconds]
promach has joined #yosys
lutsabound has joined #yosys
luismarques has joined #yosys
seldridge has joined #yosys
promach has quit [Quit: WeeChat 2.1-dev]
promach has joined #yosys
seldridge has quit [Ping timeout: 255 seconds]
emeb has joined #yosys
seldridge has joined #yosys
seldridge has quit [Ping timeout: 255 seconds]
promach_ has joined #yosys
jwhitmore has quit [Ping timeout: 240 seconds]
zino has joined #yosys
pie__ has quit [Read error: Connection reset by peer]
pie_ has joined #yosys
develonepi3 has quit [Remote host closed the connection]
jwhitmore has joined #yosys
pie_ has quit [Remote host closed the connection]
pie_ has joined #yosys
m_t has quit [Quit: Leaving]
X-Scale has quit [Ping timeout: 248 seconds]
X-Scale has joined #yosys
[X-Scale] has joined #yosys
X-Scale has quit [Ping timeout: 244 seconds]
[X-Scale] is now known as X-Scale
lutsabound has quit [Quit: Connection closed for inactivity]
m_w has joined #yosys
promach_ has quit [Ping timeout: 240 seconds]
luismarques has quit [Quit: luismarques]
pie__ has joined #yosys
pie_ has quit [Ping timeout: 244 seconds]
luismarques has joined #yosys
lutsabound has joined #yosys
m_w has quit [Remote host closed the connection]
fsasm has quit [Ping timeout: 255 seconds]
AlexDani` is now known as AlexDaniel
digshadow has quit [Ping timeout: 260 seconds]
pie___ has joined #yosys
pie__ has quit [Ping timeout: 268 seconds]
luismarques has quit [Quit: luismarques]
luismarques has joined #yosys
digshadow has joined #yosys
luismarques has quit [Quit: luismarques]
dys has joined #yosys
luismarques has joined #yosys
pie___ has quit [Ping timeout: 240 seconds]
lutsabound has quit [Quit: Connection closed for inactivity]
proteus-guy has joined #yosys
luismarques has quit [Quit: luismarques]
jwhitmore has quit [Ping timeout: 260 seconds]
pie___ has joined #yosys
AlexDaniel has quit [Ping timeout: 244 seconds]
m_t has joined #yosys
seldridge has joined #yosys
lutsabound has joined #yosys
luismarques has joined #yosys
digshadow has quit [Ping timeout: 240 seconds]
digshadow has joined #yosys
mazzoo_ has joined #yosys
mazzoo has quit [Read error: Connection reset by peer]
mazzoo_ has quit [Read error: Connection reset by peer]