[Cryptech Tech] sha3 core code problems

[Bernd Paysan]

Am Freitag, 13. März 2015, 09:58:26 schrieb Joachim Strömbergson:
> Aloha!
> 
> Bernd Paysan wrote:
> > That's strange.  What version of Quartus is that?  I use 13.0sp1,
> > because 13.x is the last one to support the Cyclone II of my DE1
> > board.
> 
> I'm using Quartus II 64-Bit version 13.1.2.
> 
> Note that the build does work for C5 and C4E, but not C4GX. Though all
> builds take very long time. I can build the whole trng faster than the
> sha3 core. But for C4GX it never seems to complete. I lat it ran for two
> hours before killing it.

That's really way, way too long. Which part of the build takes so long?  
Keccak uses a lot of routing resources; the project I build sha-3 in has very 
little else (the b16 with debugger is <800 LEs).

On my mobile Core i7 (max 3GHz, still running at 2.8GHz at the end of 
place&route), it takes 48 seconds (39s second time) to analyze+synthesize 
sha3+b16 and 1:33 to do place&route.  Changing the device to Cyclone IV GX, 
makes it synthesize in 43s (second run is always faster, now everything is 
loaded), P&R fails here, because the pin assignments don't work ;-).

Are you already too resource limited to pass place&route?

-- 
Bernd Paysan
"If you want it done right, you have to do it yourself"
http://bernd-paysan.de/
net2o ID: kQusJzA;7*?t=uy at X}1GWr!+0qqp_Cn176t4(dQ*