Legit Reviews

I've had a day to play with my new toys ;)



I'm not done yet......I might be following Bio's foot steps and going for a 100% OC

Setup:
E6300 --> 1.425v
2GB G.Skill PC8000 HZ @ 990MHz 4-4-4 --> 2.2v
P5B Deluxe --> MCH 1.65v - FSB volt = max - board unmodded (not even AS5 on chipsets)
Zalman 9500 cooler (need to get a skt 775 bracket for water block)

I can hit 3GHz with stock voltage

nice, just like how I hit 2.6 on stock with the 3800x2 (on air) Get that thing under dry ice already . Then leave it running under water after you've had your fun. Glad to see you back at it ken. That is with new hardware to boot. Though I wasnt' that impressed with my zalman gpu fan, the cpu fan I have I love. those cnps 9k series seem to really do the job well.

good job... you already passed up my E6300!



I'll trade you my B1 revision for your B2 revision ;)

Apoptosis wrote: I'll trade you my B1 revision for your B2 revision ;)

Come and get it

I think 495FSB is my max with this chip! I can't use the 6x divider to test as 6x 500 fails in seconds. I think this is due to the 965 chipiset internal clock speeds.
According to FCG the chipset clockspeed is caculated different than just multi x fsb.
In my case it would be 7/6 X 500 = 583MHz --> which is far too high for 1.65v
Source

I'm also using the 1:1 divider, I know my RAM can do at least 1050MHz 4-4-4 with 2.3v.
I tried 510x7 (3570) but it failed Orthos in 5mins. I have to goto work, but will test more later. I'm actually a little disapointed this was soo easy to do.....I'll have no more tweaking left to do after tonight ;(

I'm sure 3.4GHz will be a nice addition to our folding efforts

3.4 is a pretty darn good OC, that's what I run mine at 24/7........ I did see a guy post on the [H] about one going over 3.78 GHz on air, but as they say, seeing is beleaving............

I've gotten mine to post and run SuperPi at over 3.7, but it will only run Othos within a few minutes..........

does the B1 use less voltage than the B2? cuz Apop's b1 has 1.125v and ken's has 1.4v, couldn't you catch up with 1.4v apop?

dgood wrote:does the B1 use less voltage than the B2? cuz Apop's b1 has 1.125v and ken's has 1.4v, couldn't you catch up with 1.4v apop?

I can hit 3GHz (or more) on default voltage. I just increased to 1.45v as I was unable to get 7x500 stable. I'm proly able to get 3.4 with lower voltage.

I'm trying to learn a new board, new memory and cpu at once and I didn't want to spend hours testing. My aim tonight is to see how far I can push everything, then onto some gfx oc'in to get some hwbot scores ;)

But yes, B2 is generally better than B1 cores! --> this not necessarilly the same for AMD as E4 has a better memory controller than the E6 cores.
The older P4 Northwoods have tons of revisions, but the B0 or M cores normally clocked best. In the end it's all luck......I've seen E6700's that can't do much more than 3GHz and some that can hit 4GHz with only 1.375v

I've seen E6700's that can't do much more than 3GHz and some that can hit 4GHz with only 1.375v

4Ghz........ WOW

sorry guys i'm still new to the concepts of OCing
still thats a crapload of Ghz!!!!

Nice.

Looks like I got some work to do.

EDIT:
Just notice you have a Zalman 9500 too.

Maybe I'll beat you soon.

Nice!

Are you planning to get that thing under water?

Zertz wrote:Nice!

Are you planning to get that thing under water?

For sure! (end of next week hopefully) But I don't think my problem is cooling as I can now run 500x7 but anything higher is no go. I've even increased vcore to 1.525v (61C load (coretemp)).

I think the problem is one of two things, the cpu or the chipset.
I've tested my ram stable @1050MHz 4-4-4-8 2.3v
I ran 5-6-6-12 2.45v when testing 501x7 with all other voltages (chipsets etc) set to max, to focus on the cpu.It fails in 30sec with orthos.
Testing 500x7 is ok for 40mins orthos.

I think the problem is the chipset (again), I'm using the 1:1 (533) divider which is known to make the chipset run very tight internal timings. What I need is the option to change the NB strap to 1066, that or a 680i board.
When I add some TIM to the various heatsinks I might squeeze some more........tomorrow hopefully!

Now trying to do some 3D benchies @3.472GHz ;)

Very nice timings there! So is it completely 24 hour Orthos blend stable at 500mhz FSB with any settings? Linky to mine is in my sig, was using a Scythe Ninja at that time.

I tried a 100% OC on air and would have made it, but my RAM gave out just over 1,007mhz. (Tested that by setting a 2.33 or 2.5 mem multiplier and lowering the FSB to reach 1030mhz on the RAM) I'm happy enough though as the Corsair 6400 2gb kit uses ProMOS chips... a tRAS of 7 seems a bit silly for DDR2, loosening mine to 15 made all the difference for staying stable so you might try that. ;)

1.4-1.45v seems to be mandatory for 100% stability around 3.5ghz, even if the chip posts and runs well at stock vcore around 3.2ghz, which mine doesn't.

Kougar wrote:Very nice timings there! So is it completely 24 hour Orthos blend stable at 500mhz FSB with any settings? Linky to mine is in my sig, was using a Scythe Ninja at that time.

I tried a 100% OC on air and would have made it, but my RAM gave out just over 1,007mhz. (Tested that by setting a 2.33 or 2.5 mem multiplier and lowering the FSB to reach 1030mhz on the RAM) I'm happy enough though as the Corsair 6400 2gb kit uses ProMOS chips... a tRAS of 7 seems a bit silly for DDR2, loosening mine to 15 made all the difference for staying stable so you might try that. ;)

1.4-1.45v seems to be mandatory for 100% stability around 3.5ghz, even if the chip posts and runs well at stock vcore around 3.2ghz, which mine doesn't.

500x7 is stable for 45mins Orthos, after that I don't know. I do know 501 fails in seconds. I've tried 510 / 515 & 520 and all fail quickly too, so I've given up overclocking the cpu.
The GPU is another matter

3DMark '06
Old best with X2 @ 2.6 & 2GB @ DDR520


New Setup:

that's a pretty good jump in scores from the 3800x2 to the core 2 duo on 3dmark06. It's more than I get on either of those ones though lol. I get 5180 with my fx-60 at stock and my x1800xt. (hence the 1900xtx being better )

Looking good! A little over 3,000 is about what I get for my own CPU scores on 3dmark06...

What sort of temps is your E6300 getting at your highest stable OC, both regular and coretemps?? With the Scythe Ninja it got as high as 68c on the cores during Orthos runs, but I kept the fan around 1,600rpm. Unless hitting the volts if it rose to ~70-75c on the core then it would pretty quickly generate errors even at previously stable settings. I switched to watercooling and it didn't make a lick of difference on the OC, so I'm sure the RAM is my problem at least.

Kougar wrote:Looking good! A little over 3,000 is about what I get for my own CPU scores on 3dmark06...

What sort of temps is your E6300 getting at your highest stable OC, both regular and coretemps?? With the Scythe Ninja it got as high as 68c on the cores during Orthos runs, but I kept the fan around 1,600rpm. Unless hitting the volts if it rose to ~70-75c on the core then it would pretty quickly generate errors even at previously stable settings. I switched to watercooling and it didn't make a lick of difference on the OC, so I'm sure the RAM is my problem at least.

60C running Orthos large FFT's priority 10 with core temp, asus probe says 58C. I'm also running 1.45v now....I'm still trying to find the lowest I can run........my guess will be about 1.4v. This is @ 3.4GHz.
I'll be switching back to water cooling just to save on noise! I already have my setup, just need a new bracket for the cpu block.
I still have to test the ram......currently I'm @ 4-4-4 and tras @ 4. Trfc is 25 and most other available options are near their lowest. I know tras doesn't do anything, but I know I can't do 500+ fsb, so why not?

Well, as long as ya know it's not what is affecting yer OC is what I meant regarding RAM timings!

For my own 24 hour Blend run in my sig I stuck with 1.475v. Interestingly 1.5v would cause instability, but that may have been the extra heat it was making too.

Ya can correct me if I am wrong but C2D's seem to come with 1.25vCore even for many X6800's... My E6300 is set for 1.325v though, I'm sure it wasn't one of the better chips since it needed the higher stock voltage at such a slow speed. I'm really curious about this new rumored "low power" Kentsfield stepping INQ mentioned that's supposedly timed to launch with their Q3 price slashing... A $266 Q6600 with the new stepping should be the next jewel overclocker, assuming it is true and Intel is crazy enough to follow through with their planned Q3 price cuts!

So is the ram underclocked just a tad to get tighter timings?

elkmanwsu wrote:So is the ram underclocked just a tad to get tighter timings?

The RAM is underclocked for now. I'm not done yet, Currently the RAM is running 1:1 with the fsb, I still have to try setting the ram to use the 667MHz divider, but I'm not using 2.4v 24/7 and or Cas 5 on my rig! C2D chips run best with the RAM @ 1:1 anyway ;)

Nice overclock im at 3.5ghz with an e6700 ? mind u i havent pushed it further. If u can nearly reach this with an e6300 i might try and go higher