Author: Robert Hyatt
Date: 17:55:39 07/11/98
Go up one level in this thread
On July 11, 1998 at 16:28:55, Tom Kerrigan wrote: >On July 10, 1998 at 20:00:30, Robert Hyatt wrote: > >>>So what do you think the PII/400's L2 cache runs at? >>it runs at 200mhz is what it runs at. But for my program, a PII/200 is >>1/2 the speed of a PII/400, almost exactly. But the PII/400 is not 2x the >>P6/200 because the P6's cache is exactly as fast as the PII/400's cache. > >I'm glad you corrected your mistake (re: cache speed). No apology necessary, of >course. > >Anyway, in an earlier post, you stated the following: >processor speed >PPro/200 1 >PII/300 1.41 >PII/400 nowhere near 2 > >Now if you're arguing that Crafty fits in the 512k L2 cache, and you agree that >the PII/400 part is 33% faster than the PII/300 part, what's up with your >commentary on the PII/400? It seems to me that 1.82 is very near 2. >The only way I can think of to explain the table above is that Crafty is >dependent on memory. > Here is what I said... 1. the PII/300 is 1.41 times faster than the P6/200. I attribute this to the L2 cache on the P6/200 running at 200mhz, while the L2 cache on the PII/300 is running at 150mhx. 2. the PII/400 is twice as fast as a PII/200 running crafty. It is pretty uniform for any PII/XXX I have tried... IE the PII/300 fits exactly between the PII/200 and PII/400. So I'm not sure what the confusion is about... I stated both of those clearly I thought. Now, on to Xeon. I stated that the PII/450 xeon chip was 2.25X faster than the P6/200. That seems to be just what I'd expect. The 2.25 seems to suggest that I am running pretty well in cache, with only the occasional hash probes going to memory... otherwise the PII/450 ought to be even faster, because it is running 100mhz front-side while my P6/200 is running at 66mhz front-side. >>I didn't find "a xeon". I found a *lab full*. Crafty's the latest addition >>to specint98, which gets me lots of benchmark results I can't reveal until >>the chips are released. > >*sigh* I work at Intel. I'm on a Xeon validation project. You're not going to >impress me with these statements. > >-Tom you me either, Tom. I had xeon data 8 months ago, but couldn't say anything about it. I also had 1ghz alpha data 8 months ago, and still do, but can't say anything. You aren't the only one with access to hot chips, neither am I.
This page took 0.01 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.