[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]
Re: [f-cpu] use of 1r1w regfile for our 3r2w regfile
On Sun, Jun 15, 2003 at 11:14:27AM +0000, nicolas.boulay@ifrance.com wrote:
[...]
> There is a miss understanding. The worst case the case where it slower. 3r2w
> SRAM memory are ~30% slower than 1r1w SRAM. But if we use 1r1w SRAM bank
> there is some collision that "could" be avoid by the compiler (otherwise you
> loose one cycle).
And what about the extra logic you'll have to put around the 1r1w banks?
Every bank needs its own 2:1 input selection MUX, and you'll also
require three <n>:1 output selector MUXes to connect <n> banks to our
three read ports. Since all of them contribute to the CDP, your ~30%
speed advantage may already be gone.
> I can't say was is worst. It fully depend on compiler.
>
> So big latches array will be oustandly slow.
The point is that it's not big -- only 64x63 (since r0 is hardwired).
There are more flip-flops at the end of the first stage of the integer
multiplier.
--
Michael "Tired" Riepe <Michael.Riepe@stud.uni-hannover.de>
"All I wanna do is have a little fun before I die"
*************************************************************
To unsubscribe, send an e-mail to majordomo@seul.org with
unsubscribe f-cpu in the body. http://f-cpu.seul.org/