[rescue] Cleaning out, anyone want some FREE stuff?
Francisco Javier Mesa-Martinez
lefa at ucsc.edu
Fri Jan 14 12:49:25 CST 2005
On Thu, 13 Jan 2005, Jonathan C. Patschke wrote:
> On Thu, 13 Jan 2005, Joshua Boyd wrote:
> >> Clock-for-clock, SPARC is generally faster than MIPS by a pretty big
> >> margin.
> > I believe that was for INT though, not FP. Or am I wrong?
> It's true for both, as of US-II and R10k. I don't have any SPARCv7 and
> R4400 machines to test how far back that's valid.
For most of our codes an R10K @ 200Mhz, performed better than a USII @
300Mhz with twice the cache as the R10K. I'd be interested in finding
out where an US II at the same clock speed performs better than an R10K
(although to be fair an US II should be compared with an R12K, even
though it is pretty much the same thing as an R10K). the out-of-orderness
of the R10K should make it substantially more efficient than the Sparc at
comparable clock speeds.
More information about the rescue