SuperPi 1M Benchmark

timmylol · **Posted:** Sat Mar 06, 2010 8:32 pm

Cecil wrote:

And the reason you are ahead timmy, is because you have a native quad. The X2 unlocks arent as fast as true quads for some reason. I think its because not all the cache is usable, even though its reported in CPU-Z. Check any review with unlocks, and it shows the real quads being a little bit faster.

It was the same thing when I was running my 550X2, Zorch had to run higher clocks just to beat the score for my lower clocked one.

Nagoshi · **Posted:** Sat Mar 06, 2010 8:39 pm

(entered)

There. SS up!

Phenom II X2 550 BlkEd (Deneb [C2]) 211x19 @ 4009MHz [ASUS M4A79XTD EVO / Crucial Ballistix Tracers DDR3-1688 @ 8-8-8-24 1T]

Sorry again, I do the SS too fast before it crashes and I forgot yet again to put up the memory tab... goddamnit :lol:

NB speed was running on the 2600 multi too, so that makes it run at 2743 MHz.

-ENTERED-

Cecil · **Joined:** Sat Jan 23, 2010 1:50 am **Posts:** 963

timmylol wrote:

Cecil wrote:

And the reason you are ahead timmy, is because you have a native quad. The X2 unlocks arent as fast as true quads for some reason. I think its because not all the cache is usable, even though its reported in CPU-Z. Check any review with unlocks, and it shows the real quads being a little bit faster.

It was the same thing when I was running my 550X2, Zorch had to run higher clocks just to beat the score for my lower clocked one.

Probably sub-timing related then. Different boards set the sub timings differently. This is the main cause for some boards to be faster clock for clock then other boards of the same chipset.

timmylol · **Posted:** Sat Mar 06, 2010 9:06 pm

Cecil wrote:

timmylol wrote:

Cecil wrote:

And the reason you are ahead timmy, is because you have a native quad. The X2 unlocks arent as fast as true quads for some reason. I think its because not all the cache is usable, even though its reported in CPU-Z. Check any review with unlocks, and it shows the real quads being a little bit faster.

It was the same thing when I was running my 550X2, Zorch had to run higher clocks just to beat the score for my lower clocked one.

Probably sub-timing related then. Different boards set the sub timings differently. This is the main cause for some boards to be faster clock for clock then other boards of the same chipset.

The 550X2 was running on a ECS GF8200A v1.0

ZorchThatCPU · **Posted:** Sat Mar 06, 2010 11:10 pm

Cecil wrote:

1t also does nothing to help performance, at least with i7 anyway. A lot of people say 2t is faster.

I agree, and I am one of those people.

I actually stopped running my i7 memory at 1T after messing around with a little benching. The runs with 2T were consistently faster.

Obviously 1T makes a HUGE difference with original DDR, and to some extent with DDR2 if you happen to have memory that can do it. But not with DDR3, or at least not with mine.

Cecil wrote:

And the reason you are ahead timmy, is because you have a native quad. The X2 unlocks arent as fast as true quads for some reason. I think its because not all the cache is usable, even though its reported in CPU-Z. Check any review with unlocks, and it shows the real quads being a little bit faster.

Makes sense, Cec. I was certainly smoking his sorry ass as far as overclocks! :lol:

Benching almost 300MHz faster and coming in a full second slower in SP1M. But at least now I know why, and Timmy can F-off and stop being so insufferably smug. 8-)

The day of reckoning is coming, Timmy, and it's coming for you... a dark rider on a pale horse by the name of AMD_Fan. He's got a true quad for you, and your name is on his list! :twisted:

Anyway, good info, Cecil, and right as always!

Cecil FTW! :num1:

EDITED TO ADD:
Sorry, Goshi, but HWBot screenshot rules do apply. You can plead your case to the Black Pumpkin, since he's the bench record mod, but I doubt it's gonna fly.

Nagoshi · **Posted:** Sun Mar 07, 2010 12:57 am

ZorchThatCPU wrote:

EDITED TO ADD:
Sorry, Goshi, but HWBot screenshot rules do apply. You can plead your case to the Black Pumpkin, since he's the bench record mod, but I doubt it's gonna fly.

Shoo. And to think such screenshots made their way to HWBot and are still unreported (not that one, but other 754 runs :lol:

)

I wont cry if it doesn't count, but I know the RAM was running on the 1600 multi at 8-8-8-24-36 1T, and NB on the 2600 multi too. If need be I'll work an overclock and see how high can I get once I get in my room :banana:

ZorchThatCPU · **Posted:** Sun Mar 07, 2010 7:05 am

Nagoshi wrote:

I wont cry if it doesn't count

Aw, where's the fun if you don't cry? :roll:

Guess we'll have to count it. :lol:

Next time get your damn screenie straight, though!

Nagoshi · **Posted:** Sun Mar 07, 2010 11:50 am

ZorchThatCPU wrote:

Nagoshi wrote:

I wont cry if it doesn't count

Aw, where's the fun if you don't cry? :roll:

Guess we'll have to count it. :lol:

Next time get your damn screenie straight, though!

Yes, sir! I will, sir!

:lol:

ZorchThatCPU · **Posted:** Sun Mar 07, 2010 4:53 pm

@ Nagoshi:
(entered)

ZorchThatCPU · **Posted:** Thu Mar 11, 2010 12:56 pm

Here's a run with each of my 555 rigs. The first one with my MA790X-UD4P and 555 #1, and the second one is with my new ASUS M4A89GTD Pro and 555 #2. I spent a while on the first one the other night, but the second I only had about an hour since I just built it!

I'll probably wait until after I get water on this thing (this weekend) to do wPrime runs, since the multi-threaded stuff runs hotter. Using the trusty Big Typhoon until then. :mrgreen:

But here we go for now, updating 555 #1 and debuting 555 #2 with the new 890GX chipset.

:num1:

-ENTERED-

Overclockt.com

Register

Login

Forum rules

SuperPi 1M Benchmark

Print view

Previous topic | Next topic

Who is online

	Standard HWbot format for screenshots. Please note info included in current listings and add to your submissions.