OneyedK's MSI OCC submission Still finding ways to reach 5 posts |D Can't post url's with results atm... |
fix0red, welcome onboard :hello: |
Thanks! I'll start with a little history, starting at the end of 2004, right after the Barton-days... Good old northwood 2.4: http://users.pandora.be/kastaar/Tone.../cpuz30312.JPG Comes with a worthless SuperPI (not 1.5) result: http://users.pandora.be/kastaar/Tone...perpi30312.JPG Playing with the northwoord 2.8 version: http://users.pandora.be/kastaar/TonesForum/263_14.JPG Everybody hated that Prescott 2.8, got it cheap, external watercooling: http://users.pandora.be/kastaar/HO/270_14_4_5.JPG The AMD-hype got to me, a Winchester 3200+ just had to be tested on a K8V board... After some fiddling: http://users.pandora.be/kastaar/HO/2...sync_HTTx4.JPG After another day: http://users.pandora.be/kastaar/HO/2..._HTTx4_32s.JPG (32 seconds superPI was not bad in those days) It was wintertime so I needed another way to heat my room an i560 came in handy, another prescott in a brand new package, board was a P5AD2E Premium with Corsair ram: http://users.pandora.be/kastaar/HO/i560/spd.JPG Fast stock run: http://users.pandora.be/kastaar/HO/i...0stock_37s.JPG Go for 30 seconds: http://users.pandora.be/kastaar/HO/i...233_18_30s.JPG A little 3Dmark03 (with X800XT): http://service.futuremark.com/result...esultTyp e=10 I was so pleased and warmed with the i560 that I bought an i660 ;-) Little bench: http://users.pandora.be/kastaar/HO/i660_245_18_28s.JPG And SuperPI 1.5 arrived: http://users.pandora.be/kastaar/HO/i...0x14_29578.JPG Too bad I don't have the 3D03 results anymore.. Legendary X800GTO² came to me... So I flashed it to a X850XT PE and pencilmodded it, vGPU and vMEM... 554MHz corespeed and 629MHz memory speed. (zalman aircooled): 3Dmark05: http://service.futuremark.com/result...esultTyp e=12 I had sworn to switch over to AMD when they had the first dual core... And they did, a Manchester 4200+ on the bench: http://users.pandora.be/kastaar/HO/X2/X2_270_10.JPG Combined with good TCCD-ram, I couldn't beat the i660, not with SuperPI and not with 3Dmark03 (wich favours raw cpu power and memory bandwidth): http://service.futuremark.com/result...esultTyp e=10 IRL it was a fast combo, but the DFI with nVidia chipset proved never completely stable. Back to play with an Intel, almost free Celeron D331: http://users.pandora.be/kastaar/HO/C...331/133fsb.JPG http://users.pandora.be/kastaar/HO/C...331/166fsb.JPG http://users.pandora.be/kastaar/HO/C...331/200fsb.JPG full throttle: http://users.pandora.be/kastaar/HO/C...331/38,234.JPG 4329MHz for a 2660MHz Celeron was not bad at all ;-) Bought me a Smithfield D820 : http://users.pandora.be/kastaar/HO/D...0x14_29765.JPG Nice, hot and stable. RoaringEagle came to me to test a D830: Good for a 3Dmark05 run at 4200MHz: http://service.futuremark.com/result...esultTyp e=12 My watercooling needed some work around the time core2 arrived, so I'm aircooling since... E6400: (15.953s) http://users.pandora.be/kastaar/HO/A...00benchmax.JPG E6600: (14.000s) http://users.pandora.be/kastaar/HO/E..._400_14000.JPG Now I've got a Q6600 G0 (alas, not a very good one) on a ASUS P5E (droopin' like hell), Will try to push it for this OCC... My old X1900XTX will be paired with this one... And some OCZ PC2 8000 Titanium Alpha VX2 (not super, but still good) Hope to find some powerfull fans before the 9th, otherwise this list will have to be judged ;-) GL to all contenders! |
Welcome Karl, glad you made it. Now for some recent results ;) |
Quote:
|
up to the 30k :D |
Thanks for the heads up! Ok, step 1 finished... :mad: This is not OC-friendly weather for aircoolers! Roomtemp is over 25°C!!! Maxfsb topped out at 470MHz, luckily for me, 465MHz proved stable enough to bench... I was able to memtest the RAM at 1239MHz, but it's too dangerous to really bench at 2.45Vmem atm. Not sufficiënt cooling and I like to finish this with all hardware still usable. SuperPI @ 465MHzx7 1:1 3Dmark03 @ 465x7 1:1 I'll give the system some cooling down time and then try to better the SuperPI-score. (step 2) Will have to wait for better fans before I heat it all up with 3D03. |
With 3255Mhz you are faster than this guy: http://hwbot.org/result.do?resultId=724033 who has 3322Mhz, or this one 3.6Ghz and slower than you :) http://hwbot.org/result.do?resultId=732197 |
Quote:
Just finished testing x8, that's part 1 of step 2. Couldn't boot at 465x8, but I could boot at 460x8, 1.4250Vcore set in bios. Cranked the fsb up to 465 with setfsb, no real problems there... Ah yes, I had to strap the chipset to 400MHz, 333MHz was impossible @ 465MHz fsb. --> however, I cannot break the 465MHz-barrier of this combo, thried upping vcore, vNB, vSB and other equalising voltages in tiny steps, couldn't find a sweet spot that would allow me to crank it up. Guess it's the price I pay for overclocking an aircooled quadcore :redface: The results with 465x8: SuperPI (13.594s): 3Dmark03 (23278): 2nd part of step 2 will be playing with the x9 multi. Guess I'll have to do that tonight, a few other things need my attention now... |
|
don't forget your earplugs ;) |
Try to tweak tRD ("performance level" in memset, transaction booster in asus bios) for higher FSB on 333 strap. Could even improve things for quad core clocking. But you really need to tweak gtl voltages in bios if you want the highest fsb's, don't know if your motherboard bios supports it though. Edit : it's tRD, not Trp, messed up the two. |
Quote:
Gtl voltages cannot be set in P5E's bios :no: So I tried multi x9. Reached 445MHz fsb, strapped the northbridge to 333MHz and cranked memory speed up to 1069MHz. A superPI run was possible, working in windows no problem, but 3Dmark03 was too much to ask. Even those great Skythe fans could not remove the heat produced by the quad... If I find some time to spare, I'll try a max 3Dmark03 run with the old X1900XT. (really need a break now...) Last SuperPI result: (12.562) |
Have a look at your memset table : at 333 strap, and probably transaction booster on "auto" (I'm guessing that), performance level is set to 6 (=tRD). With transaction booster disabled (albeit 1, 2 or 3), tRD is increased with 1,2 or 3 cycles. If you go to 400 strap, Trp goes to 10 which destroys your memory bandwidth and latency. Hence much worse results. You have to increase MCH quite a bit of course to get better FSB at 333 strap. Usually, on X38, best results are obtained when using tRD of 6, on 333 strap, and fsb's around 440-475. Higher FSB's have no sense at all as they require higher tRD (or massive overvolting of the NB). No better reading than these on the subject : - Rampage formula review by FCG and - X38 and High FSB's by Raja ;) |
Thanks for the great tips on X38 OC Thorgal ;-) I reached a stable 465MHz on x8 multi, even if I could get higher, it would not be of much use, performance-wise. Problem at x9 multi is that there is not enough cooling to keep up with the total corespeed/voltage. I could try a 500x8 run, but the articles you stated prove that it wouldn't gain me anything beyond 445x8... Problem is that I don't have much time left. (work to do now :-( and beers to drink tonight :devil: ) Transaction booster ("clock twister") was set to "strong", btw... I definately need that stressball :redface: Does the contest end tonight at 12:59:19PM or do we have another 24 hours? |
MSI Overclocking Contests ends on 2008-05-09 23:59:59 http://www.madshrimps.be/?action=contests |
Ok, another day, so maybe a little time to OC some more... |
You can use Rampage BIOS on that P5E. |
Quote:
Leaves me with two problems: 1) time 2) heat --> extra problem: rapidshare :( |
Tried to get higher fsb at x8 multi last night, but to no avail :-( |
All times are GMT +1. The time now is 16:25. |
Powered by vBulletin® - Copyright ©2000 - 2024, Jelsoft Enterprises Ltd.
Content Relevant URLs by vBSEO