Well I found myself up late last nite into the early morning tweaking and turning up the sister of my system of a Co-worker that..i helped build awhile back...i amazed myself as some good results..i got out of after the final steps where done but...i was really was upset to find that the system was Full of DUST..I hate a dusty system what sense doe's it make to build a overclocking system and you don't keep it dust free, Please Air compress or Vacuum your systems out once a week if it runs constantly to keep the dust build up down and to keep it cool inside we all most know this...Prime example my co-worker system was so full of dust the heatsink Fan seiged up and won't turn at all(cm Hyper 212) http://www.frostytech.com/articleview.cfm?articleID=2206 i had to replace it with cm 120mm led..i had in my antec 300 until picks 1 up at the micro center.. 2009-08-08 his specs Biostar I45 Q9400 2.66ghz 1.362v max oc'd to 3.60ghz idle at 16c on the board by the tail,tail (led) reading G.Skill 2x2 f2-8500CL5D-4GBPK,DDR2 1066 (PC2 8500) Evga GeForce 9800 GT model: 512-P3-N973-TR Wd 250 model:Wd2500YS Cache 16mb http://valid.canardpc.com/show_oc.php?id=642846 Stardard intel burn test High intel burn test here's 7 hours and 41 mins of Orthos running after I went to bed and got up to check it intel burn running Intel burn finished running it's test i'm happy with the out come i had to do a bio's update to P45AA512.BSP http://www.biostar.com.tw/app/en/t-power/bios.php?S_ID=365 reason cause when you upgrade to a quad core this is needed this update was out after I got my board last year and he got this board this year b4 may there all rev 5.x it was dated 2009-05-12 with the quads on these boards they stay throttled even if you have the C1E & Eist disabled in the bio's it only kicks in when your about to strees the cpu it's idle function in the board..i did'nt have that problem with my E8400 EO core 2 it was locked on x9.0 with my Q9550 it was locked in on x6.0 and x8.5(stressed) until i did the update and it stayed locked at x8.5 idled and stressed..problem solved Enjoy Building Ya'll
I'm not sure this is the right place for this question, but... I'm a bit confused about my 3DMark 06 score. My record is 19038, Q9550@ 4,2GHz and my 4870@ 815/1030. But when I raise the FSB from 495 to 499 the score drops to ~18300 points :S. I'm just wandering why this is happening.
This is a weird anomaly that happens with 3DMark. I'm not sure if it signifies a real drop in speed or if 3DMark just happens to not like your clocks or what. I've seen drops of over 1000 points just raising the FSB a few MHz, but real-world performance figures say otherwise. For the record my score is ~17800 in Crossfire and ~12500 with a single card. Upping my speed to 3.8GHz leaves my single card score untouched but drops my Crossfire score to ~17000. Anyway yeah I'm sure if you ask there are quite a few people who will report the same thing you saw.
The short answer to that is that 3dmark is crap. It basically only tests your CPU nowadays, the score is highly variable, and it decays the more you test with the same hardware. Run a real game benchmark
Really 3D Mark is seriously dated and heavily Nvidia/Intel biased so you can't really expect accurate results for an ATi card anyway. It was only a good test back when the 8800 series first came out. I mean Kokki99 is beating my score with a single card only because he has an Intel quad.
Well my X2 7750BE @ 3.2GHz and my 9800GTX+ score a pretty even 13000 vs my AMD quad and a single 4870 at 12800. So you tell me...
Ouch, only 12800? I was pretty substantially over that with my core 2 I think, something like 15000 with one 4870.
I got around 12000 last time I ran it and that was with my cpu at 3.1 and 4830 at 685 core / 1000 mem. all goes to show how wrong the results are. 3dmark is all about bragging rights and not about true performance though...
Well put it this way. I KNOW how my PCs perform and 3DMark is basically 100% wrong and/or biased. Ok, so I don't use it to benchmark my PC then, only as a good example of how synthetic benchmarks are swayed To really test your OCing you need like a fairly recent video game that is somewhat CPU dependent. I use CoD4 to test my CPU OCs because it scales nicely up and down with the clock speed. I know I'm definitely getting everything out of my 3.7GHz. Though I still need to mess with the HTT and NB speeds.
@Russ messing around tonite with the Old E6750 and the old corsair xms2 800 and gotta to 3.6 stable this is the Old parts..i 'm trying to sell my Friend sandra your favor http://valid.canardpc.com/show_oc.php?id=667115 trying to lower the voltage 1.456v will not pass locked on 1.468v in the bios temps idle at 38,32 cores
Will, I guess it all in the luck of the draw. The replacement E6750 I got was apparently much better than yours as 3.8 was achieved in the low 1.42 range. This was lower than the voltage it took to get to 3.55GHz with the old one on the same P35-DS3R MB! The last few days I had it at 1.502v before I RMA'd it. I had been having to add a bit of voltage every month or so for a while before it finally started giving corrupted files and other odd-ball errors! Oh, I personally prefer the older Sandra. I don't much care for the new one myself! Personally I think Scisoft likes to keep everyone guessing! LOL!! Best Regards, Russ
now see Russ you got me wanting to go for 3.8 now to luck your record in you realize this is 800 memory..I got in here to get 3.6 (Corsair) I'm gonna slap the G.Skill in to see what happens to see if..I can get the Old girl Stable under intelburntest at 3.8...i'm not gonna hold it there i'll slap the 800 back in and back it down to 3.5 keep it there with lowest voltage i can... now you suggest to me the new sandra now you telling me you don't like it ? Is because of the mips( readings) are not the same as the old one? i was wondering that also Russ
lmao, old girl, a 3.6Ghz E6750.... Technology does sometimes make us lose our sense of perspective...
yea this Old Girl will put out to pasture after i get the funs to build my centurion back up again with the new parts..i have in plan @Russ no luck my friend with the G.Skill in she will boot at 3.8 = 475x 8.0 but that's at maxs volts of 1.5v for this 65nm chip..I'm not goin to push it on this board p35-ds3r rev 2.0...I'll believe if you can get a extra Ghz out of a chip with low voltage your doin fine..I never had this chip in my I45 board maybe it would act different or not any she's gonna be retiring or up for sale anyway after i get the CM590 back together...i think i was luck enough to get the corsair xms2/ 2x1 to do 450mhz never ran it that high b4 at 3.6ghz temp are at 38c,32c on the cores...i don't have a good heatsink on this board at the time bein just Cooler Master rr-lee-l911-GP 92mm Sleeve X Dream 4... http://www.newegg.com/Product/Product.aspx?Item=N82E16835103039 i had laying around in it's package in my drawer of parts and i think a tall sink will not fit in this modifyed case..I have to measure it to see but the cpu is fine running at 1.468v in the bios everest reads 1.42v that's fine thanks again for all your help Russ enjoy building Ya'll
Will, put one stick of memory in it and it should do even better. A 512 would be pretty ideal to see if you can get 3.7 with the same volts! Best Regards, Russ
i'll try that Russ and see what happens but sense..I believe..i don't have a good enough cooler on this board to use..I'm not gonna push the heat issue..I stressed at 3.6ghz with intelburntest and Passed it but it got up to 82c kinda scared me that's warm buddy they say heat kills cpu's along with voltage and voltage produces heat that's a battle in 1 to get a good Oc..i wanna share a Quote with you below i read you may have seen this site b4 it's below A beginner's guide to overclocking i'm not saying i'm a beginner naw i'm just trying to learn more as..i go with voltage and such to get that good oc with a lower voltage and less temperature build up so ..i can teach and share with others in hand
Two sticks are better than one for overclocking. Four are obviously worse, but non-dual channel in my experience harms performance.