[parisc-linux] Progress

Stan Sieler sieler@allegro.com
Mon, 22 Nov 1999 10:27:25 -0800 (PST)


Re:

> Missed this point because of the strange nop.  The current addib loop is 3
> instructions.  Alignment to a multiple of 16 should be good enough to
> ensure that the loop lies within a cache line.  This would insert a maximum
> of 3 nops before the loop.  This would provide a slightly more deterministic
> result.
> 
> Also, re the BogoMIPS number, I think this should be (loops_per_sec*3)/2000000
...

This should be timed on a PA-RISC 2.0 machine, just in case.  Some of
them are really good at apparently doing things twice as fast as you'd
expect :)    

Our K460 gets twice the MIPs you'd expect ... i.e., it executes a simple 
timing loop at about 360 million instructions per second ... 
on a single CPU machine with a 180 MHz clock!

-- 
Stan Sieler                                           sieler@allegro.com
www.allegro.com/sieler/wanted/index.html          www.allegro.com/sieler