[rescue] Re: Re: OH YEA??? [was: Re: Ultra?]
George Adkins
george at webbastard.org
Mon Aug 5 20:30:32 CDT 2002
> > performance out of ss20 servers. I'm just curious if that much external
> > cache makes that much difference-
>
> It depends *entirely* on the kind of code running on the processor.
> A cache will rarely negatively impact performance in any real way, but
> if the code in question is "cache friendly" a larger cache can be a
> big win. It's certainly reasonable that, for a given OS and
> application software mix, that you might not see any benefit from a
> larger cache. There are other attributes of cache design that come
> into play...cache type (write through, write back, etc), line size,
> etc...that can also have a dramatic effect on performance given
> certain memory access patterns.
>
Also very important here are:
eviction algorithm, is it simply FIFO? LRU? Some combination of the two?
Is the cache direct mapped or set associative? If it's Direct Mapped, how
large is the memory space it's mapped to?
Separate or Unified I&D caches at L1? does the incoming data evict
instructions?
Set associativity, 2-way, 4-way, 8-way? and the impact here cascades to...
Cache size and latency, not only a matter of how many cache lines need to be
searched before declaring a miss, but how much longer is it going to take to
get data from main RAM?
How much RAM does the application's dataset occupy?
When it boils down, I've never seen a *Larger* cache adversely affect
performance. But they sure can help...
George
More information about the rescue
mailing list