The little lost angel said:
http://www.hexus.net/content/item.php?item=5692&page=3
Well, it doesn't look like smoke and mirrors anymore does it?
No, it looks like a phanboi sight[sic]
The bit I'm curious about is, is there any mistake in the Sciencemark
2.0 latency results? Without an onboard memory controller, the Conroe
E6600 is faster than the FX-62. Strangely enough, the faster E6700
has higher latencies than either of the former.
Yes, these _are_ curious.
For one thing, I never measure anything nearly as fast as 44 ns,
not with my round-trip pgm below. For another, Intel boasts
their Northbridges (MCH) have intelligence and will figure out
memory access patterns, so maybe it figures out the Sciencemark
latency stride and does prefetch.
FWIW, my current results:
Latency CPU@MHz mem.ctl RAM
ns
88 k8@2000 NForce3 DDR400
144 P3@1000 laptop SO-PC133?
148 2*P3@860 Serverworks ??
178 P4@1800 i850 RDRAM
184 K7@1667 SiS735 PC133
185 P3@600 440BX PC100
217 2*Cel@500 440BX PC90
234 P2@350 440BX PC100?
288 P2@333 440BX PC66
I do need to find & test some more modern systems, but I'm
underwhelmed by the slowness of latency improvement.
compile: $ gcc -O2 lat10m.c
run: $ time ./a.out [multiply user time by 100 to give ns]
/* lat10m.c - Measure latency of 10 million fresh memory reads
(C) Copyright 2005 Robert Redelmeier - GPL v2.0 licence granted */
int p[ 1<<21 ] ;
main (void) {
int i, j ;
for ( i=0 ; i < 1<<21 ; i++ ) p
= 0x1FFFFF & (i-5000) ;
for ( j=i=0 ; i < 9600000 ; i++ ) j = p[j] ;
return j ; }
-- Robert