|[UPDATED]|E6600 + P5B Dx & 402XTREME = 4.51GHz / 564MHz FSB --> 1M/16M/32M/3D...

[bachus_anonym]

*** Results have been updated after vcore mod was completed! ***
*** A new chapter in P5B Deluxe torture starts from this post! ***
Hey

I just wanted to share the results I got, since I finally got off my a$$ and hooked up my 1.5yrs old, Chilly1's own "402XTREME" single stage This unit has served me very well so far, from AMD Winchester 3000+, to San Diego 3700+, Opteron 146 and now it landed on Conroe's head So, let see how it did this time.

Intel Core 2 Duo E6600 (L625A481) cooled w/ "402XTREME" by Chilly1
ASUS P5B Deluxe (Intel 965P) vdimm & vmch modded, BIOS 0614 & 0701
Corsair TWIN2X2048-6400C3 2x1024MB
Single EVGA 7800GTX KO @ 490+40/1390, ForceWare 84.21, High Performance driver settings + LOD 3


After insulating my ASUS P5B Deluxe, which involved:
- making a nice mess with dialectric grease in the socket, over it's pins and underside of CPU,

- dumping a gallon of liquid electical tape around the socket, over and underneath nearby caps,


- laying 2 layers of insulation tape, etc.,


... it was time for some benches. It's been unusually warm these last few days here in SF, so I've been forced to bench in 24-26C ambient temps, which does not help at all That's why, at maxed out vcore from BIOS (1.70v and 1.65v with DMM) I was hitting -39C IDLE evap temp and -20C BIOS temp Anyway, end of ranting This is what I got so far:

CPU-Z Validated Suicide --> 555x8=4444Mhz

SPi1M @ 547x8=4377MHz --> 11.672

SPi16M @ 547x8=4375MHz --> 5:25.828

Cinebench - 1 CPU @ 545x8=4358MHz --> 30s

SPi32M @ 544x8=4352MHz --> 11:52.937 ---> evap temp up to -38C LOAD

Pifast @ 544x8=4352MHz --> 21.86

Hwbot's multipi @ 544x8=4352MHz --> 41.156

Cinebench - 2 CPUs @ 538x8=4302MHz --> 16s

Dual SPi32M @ 533x8=4264MHz --> 13:36.641 / 13:34.641 ---> evap temp up to -37.1C LOAD

ORTHOS Blend @ 518x8=4144MHz --> stopped @ 2 hours ---> evap temp up to -34.3C LOAD


3DMark2001 @ 540x8=4320MHz --> 55430

Aquamark3 @ 540x8=4320MHz --> 153,147

3DMark03 @ 538x8=4302MHz --> 22611

3DMark05 @ 530x8=4238MHz --> 9900

3DMark06 @ 530x8=4238MHz --> 5625

So, I have to admit that I'm not overly excited about the results. I was hoping for at at least 4.4GHz 32M-stable and 4.2GHz ORTHOS-stable, but it seems it wasn't happening this time... On top of that, clocks for 3DM05 and 06 kind of left me disappointed, but all in all I think it was successful encounter Considering, this setup is also one of my everyday PCs, it's not that bad at all...

Thanks for dropping by and checking out my results,

[Absolute_0]

Under 12m on the 32m. Awesome stuff bachus, i'm always impressed by your benchmarking sessions.

[DAK1640]

I would be thrilled to no end personally...Great job !!!

[thunderstruck!]

Nice results. Where in SF do you live?

[afireinside]

man 11:52 I need to stick to 3d Nice stuff Bachus!

[Supertim0r]

impressive spi

[mykeos]

me likes the insulation

[epox4life]

Really pushing the FSB and memory on the 32M time very nice for the clocks.
I was running 4476mhz, 497fsb and the 32M time was 11.59mins

You should try a vcore or vdroop/pencil mod.

[sierra_bound]

Nice job, bacchus.

[eva2000]

wow very nice for P5B DX at least there

[bachus_anonym]

Thanks guys

I think that 32M result is good, not awesome, just good Especially, when I compare it to my 12:45 @ 3.7GHz stock cooled. That's about 650Mhz CPU clock advantage (73MHz memory clock less, though) and "only" 53s gain... Kind of sucks, I don't know, I was expecting more, although I don't really think that P5B DX is much slower in Pi - maybe ~15s in some cases. I once saw 11:45 @ 4377MHz but PC crashed as I was trying to take a screenshot - I was pissed!

I'm not quite sure what is holding me at this moment, lack of vcore or cooling... I was stupid and forgot about doing vcore-mod before putting all the insulation and now kind of not feeling like redoing it (although, I did not cover that needed area with LET) . But I'm sure, I could use more than 1.65v (DMM), because the board seems to currently max out @ 558MHz FSB - it would make a nice, almost 4.5GHz

[epox4life]

Here is why you should do the vcore mod right NOW! http://www.xtremesystems.org/forums/...d.php?t=116311

[bachus_anonym]

Here is why you should do the vcore mod right NOW! http://www.xtremesystems.org/forums/...d.php?t=116311

Yep, I've seen your post earlier. More vcore gave you a nice boost

I have been actually getting ready for vcore-mod before but just couldn't really get to it. I will most likely do it today, though... I really want to totally max out this setup on FSB/Mem/CPU, all at the same time - 563x8 would be just awesome!

[ANP !!!]

NB Cooler ?? Stock ??

[ent]

Bachus as usally you are pushing the hardware to the limits!!

Xtreme overclocking with extreme number of benchmarks results Too bad you haven't chance to try X6800

[MaSell]

@Ent

or colder evap

[bachus_anonym]

NB Cooler ?? Stock ??

Yes, sir - SB and NB cooler stock, untouched including TIM. Actually, believe it or not, CPU and its cooling seems to play a huge role when it comes to FSB. On stock heatink, all I saw was stable 517MHz... Now look, 564MHz

Bachus as usally you are pushing the hardware to the limits!!
Xtreme overclocking with extreme number of benchmarks results Too bad you haven't chance to try X6800

Thanks a ton! It's always a pleasure to race with all of you guys, in our 16M Challenge All those vmods, pushing for best results and my will to fight are pretty much inspired by that ongoing, 16M competition

********************
Alrighty, quick vcore mod done (all I could do at this moment, without removing too much of an insulation job )!



A propos, I had one hell of a scare - I made a fundamental mistake *NOT* to make! I preset my pot to 47kOhm and soldered it in. About 15sec after I did it, it came to my mind that I was not sure If I soldered the right leg. But noooo, stupidly I said "screw it" and fired my rig up! Next thing I know, of course, it won't POST. I checked mosfets and they were scorching HOT, although vcore measured on them reported only 1.15v. So, I shut it down, unsolder wire and check resistance - 2kOhm, damn, I was right... WRONG leg! I'm totally surprised it actually still works

Anyway, I pulled a sweet, personal records in 16M, 32M and MAX FSB! Check it out

CPU-Z Validated Suicide --> 564x8=4512Mhz

SPi1M @ 559x8=4473MHz --> 11.422

SPi16M @ 559x8=4473MHz --> 5:17.859

SPi32M @ 558x8=4463MHz --> 11:32.547

Hwbot's multipi @ 558x8=4463MHz --> 40.063

Cinebench - 1 CPU @ 558x8=4467MHz --> 29s / 748 CB-CPU

Cinebench - 2 CPUs @ 558x8=4463MHz --> 16s / 1386 CB-CPU

Sandra 2007 Memory Bandwidth Buffered @ 558x8=4463MHz --> 9425 / 9449

Sandra 2007 Memory Bandwidth UnBuffered @ 558x8=4463MHz --> 5959 / 6015

Everest Cache & Memory Benchmark @ 558x8=4463MHz --> details

Pifast @ 554x8=4435MHz --> 21.44

Dual SPi32M @ 546x8=4367MHz --> 13:17.032 / 13:12.954

I will do 3D and try to push it over 4.5GHz tommorow!


UPDATE, based on posts #30 (09/26) & #36 (09/27)


ORTHOS Blend @ 525x8=4200MHz --> stopped @ 3 hours

RenderBench @ 556x8=4449MHz --> 2m02s


3DMark2001 @ 554x8=4430MHz --> 56237

Aquamark3 @ 554x8=4430MHz --> 154,139

3DMark03 w/out CPU tests @ 550x8=4401MHz --> 22818

3DMark06 @ 544x8=4352MHz --> 5650

3DMark05 w/ CPU tests @ 542x8=4337MHz --> 10199

3DMark03 w/ CPU tests @ 542x8=4337MHz --> 22773

Thanks again for dropping by, guys!

[Gautam]

Rockin the 8x multi...intense.

[afireinside]

Is that sticky insulation tape that doesn't leave a mess or just neoprene sheet? Great 32m times!

[ANP !!!]

Cool.

Forgot to mention, This is By far One of the BEST OCs i have seen in my life.


Congrats n HAVE FUN.


[Edit] Whts the mobo(NB) and Ambient Temp. ??

[rob725]

Thanks for all the detailed info. Really helpful to see fastest possible clocks you could get for different benches.

[epox4life]

You beat my E6600 32M time! The FSB and memory is to much.

[eva2000]

guess vcore mod helped

here's what my E6700 did at 4442mhz at 1.66v vcore on P5W DH

[epox4life]

P5WDH, is good at 32M.

[^don.k's^]

So, as i can see, comparing with my p5w results, 965 is too slow. With 100mhz+ on the ram( 200ddr ), and about 30mhz more on the cpu, P5W still beats P5B times. Of course ram clocks are nice indeed, but i prefer P5W all the way

Very nice done tests man!