|
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 |
LOSTCIRCUITS |
|
| Modding a DIMM for performance / Mushkin Rev.3 This is insane, don't even think of trying it at home | |
| (Review by MS, December 4, 2000) |
Some weeks ago Mushkin released the High Performance Rev.2 256 MB DIMMs. Unable to run these DIMMs at 2:2:2 settings reliably up to 166 MHz, half of the chips were unsoldered from the PCB to reduce overall density of the DIMM to 128 MB. The goal of this experiment was to see whether the chips themselves were capable of running 166 MHz reliable at 2:2:2 settings or whether the high load of 256 MB on the memory clock was the limiting factor. With reduced density, the DIMMs performed flawlessly at 166 MHz memory bus speed and 2:2:2 settings on the ASUS CUSL2 making them the fastest DIMMs currently available. The idea was picked up and resulted in the new Mushkin Rev.3 DIMMs.
A few weeks I posted an article on the, at the time, fastest DIMMs available, each of which was running without difficulties at least 150 MHz at 2:2:2 settings. Short time thereafter, I was approached by Bill Mushkin who had just designed a few new DIMMs at a whopping 256 Mbyte density to run a few stability and frequency tests on the new modules.
The challenge of these circumstances is that, under normal circumstances, a single DIMM will always clock higher than two DIMMs of half its density. That is, a single 256 MB DIMM will run at higher frequency and lower latency settings than two 128 MB DIMMs made of the same chips and using the same PCB. However, the other side of the coin is that even in a single DIMM configuration, the added load, caused by doubling the DIMM's density will have a negative impact on the top frequency and, in particular, the CAS latency. (see footnote)
For this reason, I honestly did not expect too much of a high frequency overhead. However, the big surprise came when running the 256 MB Rev. 2.0 DIMM on the ASUS CUSL2 where the DIMMs worked flawlessly at 2:2:2 settings up to 160 MHz (memory bus frequency). On the CUSL2, there is a hole between the 160 MHz and the 166 MHz bus speed setting and, as not otherwise expected, the 256 MB DIMM did not make it all the way into Win98 (at 166 MHz, 2:2:2-9 timing settings). However, the mere fact that this DIMM even posted at this setting and started loading Windows is quite remarkable, particularly in view of the high density of 256 Mbyte.
Let's take a quick look at the DIMM itself

Mushkin Rev.2 (Stealth) 256 MB DIMM, double sided, 16 128 Mbit Infineon chips, 16M x 8 chip organization.
The individual chips are made by Infineon and mounted on Mushkin PCBs

Close-Up of the Infineon chips used.
At 3:3:3 settings, the DIMM had no problems performing flawlessly at 166 MHz in all applications as ContentCreation Winstone2000 (UDMA disabled), Expendable timedemo etc. The one question that bothered me was whether the failure at 166 MHz and 2:2:2 settings were, indeed, load related, or whether an intrinsic limitation of the CAS strobe of the chips was the primary culprit.
next page: => reducing the density for performance =>
If you enjoyed reading this article and found it useful, please consider making a small donation to LostCircuits.