AMD Overclockers Club

[Athlon2K15]

NEW AQUAMARK FOR ME AT 2.88Ghz on x2 3800,and the 7600gt at 750 core and 920 memory

 

[JC316]

Quite honestly, I don't know what the deal is with Aquamark. My X1800XT SHOULD be scoring at least 18-20K, yet it only gets 10K. It was the exact same way with my X1800GTO as well.

SCmark dosen't like my system either, yet I am fairly dominate in 3dMark06 and Sandra.

 

[Athlon2K15]

yeah that is very odd,have you tried using the directcpll fix for AM3? dont know if that will help

 

[JC316]

I think so, before it wouldn't run, so I had to use some fix on it.
EDIT:

Confirmed, I have used the fix.

 

[Wile E]

NEW AQUAMARK FOR ME AT 2.88Ghz on x2 3800,and the 7600gt at 750 core and 920 memory

That's a damn good overclock on a 939 3800X2. What kind of voltage did you have to run to get that?

 

[Tatty_Two]

lol just noticed this, my single core besting all other single cores and all but 1 dual core, an the margin in that could be considered well within the margin of error

Told you! Sceincemark is very memory biased and as you have DDR2 it does make a big difference which is why your CPU does not compete with single cores in other tests, just look how much you are in front on SM2 and then compare it with the 3D 2006 CPU test....I rest my case your honour!

 

[Judas]

new SM score ...needs adding

 

[cdawall]

******N-BENCH******

****DUAL CORES****
JUDAS.................3475
GREEK.................3199
WILE E................3128
STEVOROB..............3105
517LINKIN.............2972
ATHLONX2..............2935
JLEWIS................2887
EASYRHINO.............2868
PYROINC...............2619

****SINGLE CORES****

TATTYONE...........3543
DRDNA.................3273
BLACKTRUCKRYDER......3111
JC316..................3104
JLEWIS................2887
EASYRHINO..........2868
CDAWALL..............2745
OILY17................2643
PYROINC..............2619
ATIonion..............2194

only tests im not in last place

 

[stevorob]

Updated guys. Sorry, wasn't on at all yesterday and last night.

 

[Athlon2K15]

SLACKER......

 

[Judas]

Too many beers i reckon...

 

[PyroInc]

Too many beers i reckon...

yup

 

[stevorob]

Too much crown royal

 

[Athlon2K15]

me thinks stevo is a lightweight.....

 

[stevorob]

I'll drink you under the table...

 

[Ketxxx]

FEAR is very video based.

Wouldn't do us any good.

I mention it because unlike all the tests so far, its not synthetic.

 

[Ketxxx]

Told you! Sceincemark is very memory biased and as you have DDR2 it does make a big difference which is why your CPU does not compete with single cores in other tests, just look how much you are in front on SM2 and then compare it with the 3D 2006 CPU test....I rest my case your honour!

SCM2 isnt overtly biased. Most its tests are CPU architectural efficiancy tests (molecular dynamics, primordia, cryptography, etc) which have very little to do with available physical memory bandwith, but internally available Lv.1 & 2 CPU cache bandwidth. The only specific physical system bandwidth test is the memory benchmark, if this had as a significant impact on the final result as some believe, all them P4s would be slaughtering us with the ludicrous bandwidth them things can get

In summary, SCM2 is more sensative to overall CPU frequency (regardless of single or dual core), the CPUs architectural efficiancy and core tweaks \ optimisations, than it is to available overall system bandwidth. A good example to give would be to compare a 939 system running BH5 @ 2-2-2-6 DDR520 or so vs. AM2 with DDR2 @ around 1200MHz 4-4-4 and to run SCM2s memory bandwidth test, you will notice the scores won't differ much, in fact i'd wager the BH5 beating out the DDR2

 

[Tatty_Two]

OK I get that but how then would you explain 2 Athlons with say 450Mhz speed difference getting such different scores, surely there cannot be such a huge difference in chip efficiency between an AM2 running at 2750 and a Sandy running at 3200 and at those speeds my memory was fairly tight, like 2.5-3-2-5 @1T @ 480Mhz! Just look at the highest Sandy score against mine on the SM2 table, that amount of difference you would suggest isnt just memory ???

And if you look at this clubs SuperPI/NBench/2005CPU scores I am wiping the floor with those same AM2 chips that can arsekick me in SM2.

 

[Greek]

couldnt agree more with tatty, its definatelly the ram making those high scores!

 

[pt]

tatty and ketxxx, pls post your sm2 score here and we will see that

 

[Tatty_Two]

it already is somewhere in here, I have not kept a copy but I think mine was 1521 and his was like 1622 and my CPU was running at 3.1Gig, if you remember back then PT 3.1Gig @480Mhz on the ram gave me a higher score than 3.25gig at only 450 on the ram which would suggest that memory was doing the business.

 

[JC316]

I would agree that memoy is a HUGE factor in SCmark. I went from 2.6 to 2.8 and my SCmark went DOWN, yet everything else went up. At 2.6 my memory was running 440MHZ, at 2.8 my memory was running 370MHZ. On top of that, tighter timings improved my SCmark as well.

 

[pt]

it already is somewhere in here, I have not kept a copy but I think mine was 1521 and his was like 1622 and my CPU was running at 3.1Gig, if you remember back then PT 3.1Gig @480Mhz on the ram gave me a higher score than 3.25gig at only 450 on the ram which would suggest that memory was doing the business.

i don't care about the score, i want to compare the results in primordia, and the others
it's probabily on the sm2 thread but i don't know where
can you find yours?

 

[Tatty_Two]

i don't care about the score, i want to compare the results in primordia, and the others
it's probabily on the sm2 thread but i don't know where
can you find yours?

Ahhhh right........here ya go:

http://www.techpowerup.com/forums/attachment.php?attachmentid=4714&d=1165014608

 

[pt]

Ahhhh right........here ya go:

http://www.techpowerup.com/forums/attachment.php?attachmentid=4714&d=1165014608

thanks
now we just need ketxxx score