LOSTCIRCUITS |
|
DDR DIMMs, BIOS and Timing Issues
Dont Do Rambus
|
|
(Review by MS, January 1, 2001)
|
Performance
What is specified here as performance is not going to rely on benchmarks since those are completely independent of any DIMM design or brand but rely mostly on the controller and the BIOS settings. In the case of SDRAM, the difference in DIMM performance is almost imperceptible and, in most cases, differences shown can be accounted for by changes in the operating temperature or relate to other subtle changes in the environment as, e.g., runtime before the benchmark was started, rather than real performance differences of one DIMM versus another.
In the case of DDR, there are quite a few additional parameters that my play into the measured performance, just as example, leakage currents or bleeding of the charges within the primary sense amplifiers can cause shorter page life-times. Thus, additional latencies in form of RAS precharge and RAS-To-CAS Delay can occur in some DIMMs, whereas others still can draw from an open page. This is meant just as an example. Other possibilities include differences in the organization of the DIMMs, any x8 organization will have a minutely different performance profile than a x16 chip organization. On the other hand, these differences are measurable only in memory benchmarks and not reproducible in real world applications, at least, we haven't been able to see them on a system performance level.
As an Entrée here are a few numbers:
SiSoft Sandra Memory Benchmark results at 133 MHz, CAS-2 setting, comparing the Micron 128 MB double-sided DIMMs with the Hyundai chip-based Mushkin DIMMs. There is a minor edge of the Micron DIMM over the Mushkin DIMM, moreover, the Micron DIMM works at the Ultra and Ultra2 setting even at 133 MHz where the Mushkin DIMM does not function. A less than 1% difference in a synthetic memory benchmark is within the margin of error and has absolutely no bearing on real world performance.
Much more interesting here is the compatibility with the settings that define performance or stability. That is, why are the Micron DIMMs running at the Ultra and Ultra2 settings where the Mushkin DIMMs fail?.
next page: => tRAS violation =>
If you enjoyed reading this article and found it useful, please consider making a small donation to LostCircuits.
Thank you!
General disclaimer: This page only reflects the author's personal
opinion and assumes no responsibility whatsoever regarding any of the contents
or any damages that may occur explicitly or implicitly from reading the
contents of this site. All names and trademarks mentioned in this review are the exclusive property of the respective parent companies.
All contents of this site are protected by international copyright laws. Reproduction of the contents
even in parts is not allowed except after written permission by the author and referral to this site.
Copyright 1998 - 2007 LostCircuits
