🚀 go-pugleaf

In article <ackerman.5.2DA59B72@skyler.mavd.honeywell.com> ackerman@skyler.mavd.honeywell.com (John Ackerman) writes:
>In article <2o3u6l$ih9@fitz.TC.Cornell.EDU> strider@eos.acm.rpi.edu (Greg Moore) writes:
>
>>>       Thanks for your attempt to help, but myself and the previous
>>>       poster are refering to the external cpu cache, not the ram cache.
>>>
>>        OK, I'm confused.  If the external cpu cache doesn't cache
>>ram, what does it cache?  A cache for the cpu is there to provide
>>quick access to items stored in memory, the memory being RAM.  Or
>>am I missing something?
>
>I think we are confusing hard-disk cacheing (eg smartdrive) with memory
>cacheing.
>
	Who WE paleface?  If you're refering to my post, I know
the difference between the two.

>I was referring to the cache memory, not the memory normally referred to as
>RAM.  Cache memory is the high-speed buffer that the cache controller stuffs
>with data from RAM that it anticipates will be needed by the processor.
>Every time processor has to access regular RAM, with its longer access
>time, things slow down a bit.
>
	No shit.  Read what I wrote.  That's essentially what I wrote.
A CPU cache, L1, or L2 caches main memory.

>It sounds like your 10% improvement with a 128k cache is typical,
>depending on the processor intensity of the application.
>I'm guessing that another 5% tops would be achieved by going to 256k.
>I suspect that claims of extreme degradation without cache are either due to
>some setup problem, or are actually references to disk cacheing problems.
>
	Could be both.  I believe some systems (I don't know
about ZEOS above 16 meg) have trouble caching that memory if you
don't set it up correctly.



--
--
Greg d. Moore           | All that glitters is not gold.
Green Mountain Software |