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?

Performance Summary

Within the different timing and speed settings, all DIMMs provided identical benchmark results at the 3:3:3 setting. This is not surprising since the chipset actually sets the command parameters and the DIMMs are merely executing (or dumb) parts. In other words, there may be some minor differences but those are within the margin of error and do not hold up under statistical scrutiny.

This also means that simply by changing DIMMs to a higher speed grade without adjusting the chipset parameters / memory bus timing settings, one DIMM will behave exactly like any other. There is a wide-spread misconception regarding this fact but an analogy would be that even with racing tires, it is not possible to find a parking spot any sooner.

The situation changes ever so slightly with moving to more aggressive timing parameters. That still means, that at identical chipset settings, all DIMMs will generate the same benchmark results. However, that also means that if some of the timing parameters are changed from one DIMM to another, the benchmark results will be affected.

How much is the impact on benchmarks? The answer is very simple, it depends on the application. In office applications that are running off the CPU cache, there is very little performance increase even by stepping up from CAS-latency 3 to CAS-latency 2. Along the same lines, reducing tRCD from 3 to 2 cycles results in performance increase that is too small to even notice.

There are differences though and these differences can be monitored using specific memory benchmarks that are geared towards evaluating the impact of performance and bandwidth. I have recently posted such an article with in-depth analysis of the entire spectrum of settings which took 10 pages so I don't really know how to put it into one paragraph here but I encourage everyone to take a look at this Latency vs. Performance article.

Evaluation Summary

All DIMMs included in this roundup by far exceeded not only the performance standards set by the industry but also our own expectations since there is no doubt that the Corsair PC133 CL2, the HSDRAM 150 and the Mushkin PC-133 Rev 2 DIMMs all woulc qualify as PC150 CAS-2 DIMMs if such category were to exist in industry standards.

Alternatively, all three DIMMs could qualify as PC166 CAS-3, that is, HSDRAM 150 even runs at that speed at CAS2.

So what is the verdict or what are the recommendations?

Number 1 recommendation is that spending the extra $20 or $30 on either of these DIMMs is probably one of the best investments one can make since it will pay off in terms of system stability and performance.

Number 2 recommendation is that it appears as if the Corsair DIMMs was going possibly about 2 MHz higher at 2:2:2 settings on the KT chipset than the Mushkin Rev.2 DIMM whereas the situation on the Solano was just the opposite. This, again is within the margin of error and may vary from one board manufacturer to another.

Number 3 recommendation is not a recommendation but a short comment. In a recent roundup of i815E-based boards, Tom's hardware stated that HSDRAM 150 would only run in 3 out of almost a dozen tested boards. Since neither Tom nor Patrick replied to our inquiry about this article (nor updated the article), we have tested 4 of the remaining boards and, further, consulted with Johan from Ace's Hardware who confirmed our finding in at least one case. That is, the incompatibility found by Tom's Hardware could not be replicated in either our, or in their lab. Bottom-line is, there are no inherent problems with HSDRAM 150 and the i815 chipset.

next page:    => more =>