Navigate: Advice Beginners BIOS Guide CPUs Links Mainboards Memory Network Storage Video/Sound Cards Contact Forum SiteMap Sponsors WebNews Home	.	.
CPU Intel P4 840 D P4 820 D P4 630 P4 640 P4 650 P4 660 P4 670 AMD Athlon64 3500+ 3700+ 3800+ 4000+ X2-3800+ X2-4200+ X2-4400+ X2-4600+ X2-4800+ 1-Way Opteron Opteron 144 Opteron 146 Opteron 148 Opteron 150 Opteron 152 2-Way Opteron Opteron 240 Opteron 242 Opteron 244 Opteron 246 Opteron 248 Opteron 250 Opteron 252 2-Way Dual Core Opteron Opteron 270 Opteron 275 nVidia GF 7800GT GF 6800GT GF 6600GT ATI R X850 XT PE R X850 XT R X800 XT PE R X800 XT R X800 XL Memory Corsair Crucial Kingston Mushkin OCZ What are you shopping for?

Synthetic Benches:

Sisoft Sandra Memory Bandwidth Buff’d

Sandra is somewhat of a bittersweet pill. While the validity of the data is certainly a matter of discussion, it’s a popular and widely used ruler on which to gauge our systems, therefore we include it for comparison. What we see, is the PDP at the higher frequency and latency settings ruling the roost. The bandwidth numbers show in general that using the 3:4 divider and looser timings is able to provide more bandwidth in this test versus using tight timings at the 1:1 ram divider. However, there appears to be a point of diminished returns. If you look at the Crucial Ballistix modules, when the regular PC5300 is run at 280 fsb and 3:4 divider with 4 4:4:12 latencies and compared to it’s Tracer sibling, with a slightly higher 285fsb and 3:4 divider, but then requiring latency settings of 5 5:5:15, the 280fsb and slightly lower timings prevail. It may turn out to be interesting to follow those for comparison later. Also curious is how the Mushkin PC4200 is trailing the PDP even though the same timings are used. Weather this is other SPD optimizations on PDP’s part, or just a bad run for the Mushkin it’s hard to say. In hindsight, I should possibly recollect data on the Mushkin, however this might be considered unfair as each Sandra benchmark was run once and recorded. Any variance in the Mushkin that produced a better score might otherwise unfairly skew data in the presence of other competitors that didn’t have the chance for a “retake”. However, the Mushkin did beat the PDP at the 1:1 CPU:Memory ratio with 3 2:2:8 latencies, so I’d be inclined to wonder about the SPD optimizations and any performance predictors that might come from such changes. This might be interesting to follow as well.

Ed. note: There is a 0.5-1% variance between individual runs of memory benchmarks depending on the boot sequence and applications resident in the background, moreover, the PCB appears to have more impact on the outcome of memory bandwidth benchmarks than what would be expected. In theory, at any given latency settings, all modules should perform identically, in practice, though, there are differences.

Ed. note: In the case of Sandra, we usually find stabilization of the results on the second or third run. Keep in mind though that in most cases all modules follow a similar trend.

next page:    => Everest =>