5 Attachment(s)
Mushkin XP8500 5-5-4 2.25v 2gb kit
I got 2 kits sent from Mushkin for testing purposes. One kit was quoted to be a "barely" pass kit from their bin profile. 2nd kit was random grabs from "passers". I will not divulge bin profile they setup as I don't know it 100% but even so that's their business. Kits will be referred to as "weak" and "random" from hereon out to define the barely passer or the random grab kit. Brian wanted me to test a barely passing kit to see where a theoretically weaker oc'ing kit would fallout once pushed beyond bin profile parameters. I think this is kind of refreshing actually as for sure there are not cherry picks and in fact would kind of signify at least in the weak kits case what a poor kit would probably do. I haven't gotten 3d benches to show yet b/c one OS is semi hosed and had 3 vid card changes in the testing time which would have made comparisons the poo. The 8800GTX that ended up on the watercooled system will be moved shortly as well and I just have so much chaos in my room that I don't know where things will end up lol. If someone wants proof of 3d stability just lmk what you want me to run, I'm not trying to duck anything it's just I have nothing impressive to post so I figure why bother. I don't have 6800XE on dry ice with super clocked top teir card lol. I low clocked 6300 doesn't exactely make super benches ;)
http://www.xtremesystems.org/forums/...1&d=1164896217
All testing done on either Asus P5b Deluxes(2 diff boards bios 0804) or
Asus M2N deluxe SLI for AM2 testing.
Most screenies will show some form of Orthos testing. As I really don't consider dual super pi 32m a good test for 24/7 stability, whereas orthos within a few hours will tell me relative stability and overnight will usually give me assurance of 24/7 gaming stability.
Quick and dirty take
Random kit on AM2 I went easily to 564mhz with 5-5-4-14 2T timings assigned in bios, all secondary timings left at auto. 2.25v selected in bios/2.25v per Asus software monitor. More voltage didn't help and fiddling with secondary timings left me frustrated to the point I gave up shortly and moved that kit over to my 2nd P5b deluxe. I'm sure in more capable AM2 ocing hands this kit would have gone much much higher. Basically I just wanted to make sure they went did their spec and above on AM2.
http://www.xtremesystems.org/forums/...1&d=1164896312
On Intel side I'm still maxing. Asus P5b deluxe is joy to work with as most of you know. One board and 6300 I had on Watercooling setup, other setup I had either 6300 or 6400 on VapoLS.
Of note, I accidently booted with timings I had set for some other D9 based memory prior to changing anything. Timings were 4-4-4-12 at 2.25v per Asus probe, 2.3v set in bios @1020mhz. I only got orthos SS at 24mins after clocked up to 510mhz bus as I initially booted at 500, hence the difference in CpuZ vs what Orthos is reading, but later I did get longer screenie I'll try to find it or redo another as it was stable and it didn't fallout at that speed.
http://www.xtremesystems.org/forums/...1&d=1164896312
At default timings of 5-5-4 both kits seemed to like 2.35v for 24/7 operation. Higher volts would allow to scale a bit but I saw diminishing returns. To explain the gains I was getting were not consistent with the bumps I was giving and tbh I was a bit leary of giving them high volts for long periods of time as I don't want to nuke them ;). Have plans for running 4x1gb for Vista :toast: Anywho! The weak kit was only weak at 2.25v where it maxed at 1096mhz give or take at 5-5-4 2.25v set in bios. I could run 5-5-4-8-4-4-11-4-11(hope I have those inline right). that seemed to produce best combo of oc'ing and performance, and again I'm sure in other's hands that have more time etc... might have gone a bit higher. However, once the 2.3v barrier was crossed these stix blew past 1100mhz and went to 1160mhz dual super pi stable and 11+hours of orthos stable prior to me stopping test. I really think they could do 1200mhz with more tweaking and me trying different dividers as some funky Asus issues were starting to happen with the 2:3 divider up around 390-400 fsb area. More time may be needed ;).
The random kit was something of an anomoly. It had lesser headroom at 4-4-4-x timings but slightly higher stability at 5-5-4 timings. I will have to pull the pics off of that system, but alot of benches were concentrated at 1:1 on the vapo as it's such a shame to run a low fsb and divider :/. I may have to move sticks around to get new screenies and to max on non VapoLS system. I also want to note that on the Asus watercooled system there was NO cooling on the dimms to represent worse case scenario. No airflow on them at all really as they don't even get the benefit of air from a traditional heatsink/fan combo.
http://www.xtremesystems.org/forums/...1&d=1164896217
http://www.xtremesystems.org/forums/...1&d=1164896309
