flexy Posted May 6, 2005 Posted May 6, 2005 i posted already on xtremesystems, but here a recap....i think this is very odd. I was running my system for a few months OCCT stable (at least an hour or so) at 2.5 4 3 7, 2.9V on the Vdimm and Winnie at 2601 (9x289). SO..recently i got extremely instable at this setting, passing OCCT barely a few minutes. With the SAME settings and same bios (310p). After days of tweaking and even checking out my sticks (heatspreader might have come lose etc...i rechecked and pressed to make sure the spreaders are tight)..... Well...the sticks pass 10 runs of memtest #5, haven't tried more yet. But i still have the same instability in OCCT. So...after tweaking EVERYTHING possible and not seeing any results i am now running at 10x260. Put down my latency and my bandwith is *about* the same as i had before with 289. I was always around 3176MB/S (memtest, Memory bandwith) and now w/ 260 i am at 3068. Happy that it is over 3000 It just really irks m...but i have an SUSPICION that it is not the OCZ memory (as said, need more testing)....but rather the A64 memory controller which, for some reason, got instable in time with settings i had before and whichw ere ok for at least some time OCCT. Who else lost an overclock like i did ? Btw. its NOT an heat-issue...if you dont believe me i can invite y:)ou to Chicago...because it was VERY cold until recently so summer-heat etc. is out of the question. Also, right now i have a fan blowing over the sticks, in addtion to my XP-120. Share this post Link to post Share on other sites More sharing options...
Opponent Posted May 7, 2005 Posted May 7, 2005 I've been noticing too that there are some changes in the behaviour of either the on-die mem controller or the sticks themselves... I posted this already in "official 300+ and plat rev 2" thread but heres a recap what happened to me... As I got this rig I pretty soon found the settings that gave me 300 fsb stable. Could prime for hours, NP with OCCT or any other test or game whatsoever... but going 305 fsb would give me errors in minutes... Used these settings about two weeks pretty much 24/7, was using CPU multi 8 with that so after a while I thought I'd take one more look to see if I could squeeze something more out of CPU with some voltage tuning... So for a couple of days I played the voltage game with overclocking CPU and sometimes gave over 1.7 for quite extended periods and after that when I'm back at my old settings (300 fsb x 8) I find out they don't work anymore... OCCT gave me errors in minutes... So I thought I must damaged the mem controller and now it's degraded or something... But now I could pass OCCT and SuperPi32M with 305 fsb instead! NEVER an error in them while before I would get errors in minutes... But with the same exact settings I couldn't do 300 anymore... So I don't know what exactly has happened but something sure has changed a bit. Share this post Link to post Share on other sites More sharing options...
magg Posted May 7, 2005 Posted May 7, 2005 Happened to me too, couple days ago. Used to run 289x9 for ever, throw any stability you want at it and it'd laugh in it's face. Cold boot it one morning when I wake up, won't go past the windows logo. Had to revert to stock and now that's what I'm running at [iCKY] because I haven't had the time to eff with it. Tempermental board? Or ram/cpu? Who knows... Share this post Link to post Share on other sites More sharing options...
lunatic Posted May 7, 2005 Posted May 7, 2005 hehe kinda like me my sticks wont go on stock voltage or standard timings had to put 3.6v into tccd to get em kickstarted since then i havent been able to run em on my mobo but they work fine on any other motherboard at stock settings :C kinda odd dont ya think? Share this post Link to post Share on other sites More sharing options...
rise4310 Posted May 7, 2005 Posted May 7, 2005 i don't get it opponent. so now you're stable higher and not lower with the same volts and timings? thats bizarre. what are you other voltages if i may ask? i'm running the same 1.425+1.13 with 1.4ldt, 1.6 chip and 2.9 dram. i'm stable at 275x9 but 280x9 crushes me within minutes of occt and prime as well. i don't want to be pumping too many volts thru this thing until i can afford a 3700+ lol Share this post Link to post Share on other sites More sharing options...
flexy Posted May 7, 2005 Posted May 7, 2005 i don't get it opponent. so now you're stable higher and not lower with the same volts and timings? thats bizarre. what are you other voltages if i may ask? i'm running the same 1.425+1.13 with 1.4ldt, 1.6 chip and 2.9 dram. i'm stable at 275x9 but 280x9 crushes me within minutes of occt and prime as well. i don't want to be pumping too many volts thru this thing until i can afford a 3700+ lol well i *know* i am not alone with this problem because i read stories about "losing overclocks" earlier already...i *really* would like to know what is the reason. Also...regarding pumping voltage: It looks that i get MORE instability the more voltage i pump into the winchester....it actually really looks mosyt stable at 1.48, maybe up to 1.53.....but over 1.53 i get MORE instability...it looks like that winchesters dont like voltage AT ALL !!! I can run my 2601mhz at 1.46 or even default voltage fine...but more voltage is definitly NOT doing any good. btw. my winchester is a week 41 Share this post Link to post Share on other sites More sharing options...
Opponent Posted May 7, 2005 Posted May 7, 2005 i don't get it opponent. so now you're stable higher and not lower with the same volts and timings? Yeah, thats right! :tooth: I'm using 1.4 LDT and 1.7 chipset. Thats what I've been using most of the time, have tried others but didn't see much of a difference... Your guestion made me wanna doublecheck so I'm doing some re-testing as we speak by trying different LDT and chipset voltages just in case I lost track somehow... But I'm sure those were the volts I used and they used to work with 300 and not 305 and now its the other way around My winnie loves so much volts I can't get anywhere without giving it a good jolt... It seems to scale pretty linearly past 1.7... But that's where I start to get worried using that high volts constantly... But if those Venices live up to their hype and I decide to get one I DO wanna see how far it will go temps permitting update: So far checked changing the chipset voltage while LDT volts constant @ 1.4: 1.6 v OCCT passed @ 305, not @ 300; 1.7 v 305 passed, 300 not; 1.8 v 305 not; 300 not... Share this post Link to post Share on other sites More sharing options...
rise4310 Posted May 7, 2005 Posted May 7, 2005 very odd. my winnie, 0508 i think, is the same with the volts. i just tried backing down a bit on the vid and occt passes but prime fails in 10-15 minutes. i am reluctant to go higher because of the range of volts i see people feeding these and getting stability. seems like its all good from 1.475 to 1.7, lol which voltage actually effects the memory controller, the ldt, chip, vid or all the above? Share this post Link to post Share on other sites More sharing options...
PC1X1 Posted May 7, 2005 Posted May 7, 2005 I recomend a good burn in, and no drastic overclocks until a couple hours , volts is like an addiction, once u start you cant stop. be careful not pushing it to hard though guys, although the numbers are pretty, sometimes its not worth killing your winnie friend, for 2 more fps, or 1 minute faster render. Share this post Link to post Share on other sites More sharing options...
flexy Posted June 2, 2005 Posted June 2, 2005 UPDATE: Yes it IS the d*rn OCZ heatspreaders. I checked those a while ago and *thought* everything was ok...well today i checked again...and it was not....d*rn heatspreaders came off on the sides. So i heated the sticks up with memtest, and with careful force and after some work with a small screwdriver (to remove the clips) i removed the spreaders. Back at my old 289MHZ and now even better with a fan blowing over the sticks and no halfa$$ heatspreaders which come off anyway sooner or later. Hint OCZ: I really LIKE your products (you know that maybe), i raved about the sticks on bleedingedge and many more places....but these spreaders have to be redisigned, at least the cliips ! Or sell sticks without spreaders !! These things have NO benefit....push 2.9V in our high-end sticks and let them run a while and i can can guarantee in MAX 2 months they will get lose and you will have a significant decrease in performance like i had. thanks ! Share this post Link to post Share on other sites More sharing options...
AndyOCZ Posted June 2, 2005 Posted June 2, 2005 UPDATE:Hint OCZ: I really LIKE your products (you know that maybe), i raved about the sticks on bleedingedge and many more places....but these spreaders have to be redisigned, at least the cliips ! Or sell sticks without spreaders !! These things have NO benefit....push 2.9V in our high-end sticks and let them run a while and i can can guarantee in MAX 2 months they will get lose and you will have a significant decrease in performance like i had. thanks ! Heya flexy, This is an old debate and while I agree with much you say, for many users heatspreaders serve a function. They keep the chips from getting orange from Cheeto Cheese Puff goo They look pretty in a nice rig. They can help with ESD issues. And lastly they can be used to spread mustard on a sandwich after you take them off. This voids your warranty of course. Share this post Link to post Share on other sites More sharing options...
flexy Posted June 2, 2005 Posted June 2, 2005 Heya flexy, This is an old debate and while I agree with much you say, for many users heatspreaders serve a function. They keep the chips from getting orange from Cheeto Cheese Puff goo They look pretty in a nice rig. They can help with ESD issues. And lastly they can be used to spread mustard on a sandwich after you take them off. This voids your warranty of course. well dont underestimate the value of a good sandwich Anyway, Andy, i have to admit they look pretty cool, and maybe one important factor too is that they make the sticks more "robust"...eg. in handling them, putting them in the slots etc. I was experimenting first using some coathanger metal-clips to tighten the spreaders on the side (i needed more clips, wife has thousand coathangers but i only found ONE with those clips.....or i could have gotten those binder clips....but i didnt have any....so i just remembered a thread on XS where people said they had success in removing the spreaders...so i just jumped on that and removed 'em despites knowing that my warranty will be gone Still glad they work and memtest fine after what i put them thru...including removing the glue remains with alcohol etc.... But as Andy said: it VOIDS your WARRANTY, so dont do it Share this post Link to post Share on other sites More sharing options...
Recommended Posts
Please sign in to comment
You will be able to leave a comment after signing in
Sign In Now