|
Advice Beginners BIOS Guide CPUs Links Mainboards Memory Network Storage Video/Sound Cards Contact Forum SiteMap Sponsors WebNews Home |
. | . |
|
Prices: CPU Intel P4 2.4C-800 P4 2.6C-800 P4 2.8C-800 P4 3.0-800 P4 3.2-800 AMD AthlonXP XP 1700+ XP 2000+ XP 2400+ XP 2500+ XP 2700+ XP 3000+ XP 3200+ Athlon64 Athlon64 3200+ Athlon64 FX-51 Opteron Opteron 240 Opteron 242 Opteron 244 Opteron 246 Memory Corsair Crucial Kingston Mushkin OCZ |
LOSTCIRCUITS |
|
|
| Intel i875 - Canterwood SATA and GbE for PAT | ||
| (Review by MS, April 14, 2003) |
Even though raw memory performance as measured in form of data transfer to the CPU is almost becoming a secondary feature compared to the main dish, which is to provide the necessary overhead for quasi-parallel DMA engines, it is interesting to twain out where the bottlenecks are in the system at hand. From the Granite bay platform, we know that the bus utilization of dual channel memory is way below its single-channel pendant. That is, out of the 4200 MB/sec theoretically possible in dual PC2100 mode on Granite Bay, we only found roughly 3400 MB/sec, equivalent to about 80% which is quite a step-down from the 94-96 % we have gotten used to in single channel memory platforms. Keep in mind, though that the absolute numbers are still overpowering.

Where raw power comes into play, the finesse of only 76% bus utilization does not count anymore. Suffice it to say that the single channel bandwidth of roughly 3.05GB/sec bandwidth is outstanding. GB: Granite Bay; 2:4:4 = CAS-2, tRCD-4 tRP-4, 2:3:3 = ditto.
Overall, the trend is that with increasing total bandwidth, more sacrifices have to be made with respect to bus utilization.

Using the "Burn-in mode" to overclock the system we managed to push past the 5000 MB/sec threshold.
next page: => I/O performance =>