ScienceMark thread (with score list)

[Urlyin]

I scored a 1537 ....
175 Dual Core Opteron @2.8
10 x 280
DFI CFX3200
Gskill PC4000 @ 1:1


I'll post a screenie once I get a chance...

 

[Alec§taar]

I scored a 1537 ....
175 Dual Core Opteron @2.8
10 x 280
DFI CFX3200
Gskill PC4000 @ 1:1


I'll post a screenie once I get a chance...

WoW!!!



* 3rd place now, behind yourself & dj dn... both of you running Opteron "DC" CPU's!

APK

P.S.=> BoneFace, maybe you're "pushing too hard" via too much o/c'ing? I don't know...

I just hit 2772.5mhz here as an overclock!

I am now using a 231mhz FSB + 12x Multiplier for that rate... & imo so far @ least? I think I did better @ 2750.1mhz via 250mhz FSB + 11x Multiplier for THAT rate!

(& it doesn't want to run this ScienceMark 2.0 test properly now!)

That tells me I am pushing TOO hard & may have been better off using a lower multiplier 11x by 250mhz FSB - ScienceMark runs on that just fine, whereas this latest o/c won't let it! apk

 

[Urlyin]

I have a little room left .... but I'm surprised you're not getting a little more out of that 4800 X2 .... I'm running little high on the vcore and I'm using a wb ...

 

[Deleted member 24505]

new result at 4285mhz.

you got me by 4 points alecstaar lol

 

[Alec§taar]

new result at 4285mhz.

you got me by 4 points alecstaar lol

You're RIGHT ON MY TAIL!!!

* How the HECK did you get 4285++ mhz?

APK

 

[Alec§taar]

I have a little room left .... but I'm surprised you're not getting a little more out of that 4800 X2 .... I'm running little high on the vcore and I'm using a wb ...

Well, I am STILL "experimenting" here, & it IS very HOT outside (hitting 100F today already), so it may be affecting what my TRUE "top-end" can be, stable...

I did better using:

11x multiplier + 250mhz FSB = 2750.1mhz overclock (196/200mhz DDR-400 RAM speed set DDR-333)

vs.

12x multiplier + 231mhz FSB = 2772.1mhz overclock (184/200mhz DDR-400 RAM speed set DDR-333)

IMO @ least, mainly because the first one yields BETTER ramspeed, vs. the 2nd one, though it has a higher overall mhz gain!

APK

 

[Deleted member 24505]

what about this then-

would'nt run science mark tho' bsod.

 

[Alec§taar]

would'nt run science mark tho' bsod.

No WONDER you could top 4285mhz - you're an "INTEL MAN"!

APK

 

[Alec§taar]

I have a little room left .... but I'm surprised you're not getting a little more out of that 4800 X2 .... I'm running little high on the vcore and I'm using a wb ...

Got a BIT more outta it, now... & FINALLY STABLE too, with better overclock than I have had before (glad I took this test now for sure)!

CPU-z 1.35 CPU (showing ~353mhz SOLID/STABLE overdriven gain over stock speed 2400mhz)

&

CPU-z 1.35 RAM overdriven gain (getting ~197mhz of 200mhz possible on DDR-400 via DDR3 dividers used)

ScienceMark 2.0 score (best yet @ 1401++, & stable):

* There she be, best I can do... even w/ a 100F heatwave ambient temps outside today!

(Finally, I broke 1400 & I am STILL in 3rd place here, lol - wtf!)

I am now 37 points behind dj dn & his Opteron DC cpu @ 1446 points (who trails the 1537 point leader, Urlyin (same cpu type in Opt165 DC also))...

APK

P.S.=> It seems this ScienceMark 2.0 is a damn useful test, because it's also a GOOD test of CPU + RAM stability under duress of overclocking... given this heat, I have not only sped up my system substantially today, but also gotten a STABLE runner outta taking this test!

BONUS!!! apk

 

[Alec§taar]

Again: Achilles in the film "TROY", said it best:

"Is there NO ONE ELSE!?!"

WELL? See above...



* Rinse, Lather, & Repeat 3x... lol!

APK

P.S.=> Test's been a good thing for me, got a stable faster o/c from it, & pretty much have decided that this benchmark's also a decent stability test as well!

It was a challenge to make the "1400++ club" as well...

So, got the machine for it? I know there's more systems here faster than mine, & if you're new? Bring 'em on as well... apk

 

[Deleted member 24505]

i think it is a good stability test,mine failed coz i had the mem at 950,i think it may be too high.i will try it at 800 and summthing at 4.4 and see if it will do the test at that.i'm close to you tho' mate.

i only need another 6 or 8 points and i.ve got ya. lol

 

[Alec§taar]

i think it is a good stability test,

I am fairly convinced of that myself as well!

(I.E.-> Not only is it a decent benchmark, but also a stability tester for overclockers (2 for the price of one)).

mine failed coz i had the mem at 950,i think it may be too high.i will try it at 800 and summthing at 4.4 and see if it will do the test at that.i'm close to you tho' mate.

You have to step down some it appears from that, this is certain... I went thru the same, & could achieve TREMENOUS overclocks (out past 2828 iirc noted in this thread) & I could post forums & surf all day, download files etc. just fine - stable!

BUT, run my system thru that test all the way? NOT!

Again - it is a stability tester as well as a benchmark (dual bonus)...

i only need another 6 or 8 points and i.ve got ya. lol

Yes, you are almost there...

MISSION - "Get to the 1400++ club stable - pass thru ScienceMark 2.0 to prove it"



* It isn't easy when your system's just about maxed-out... & it's not all about mhz, but also how much of your memory mhz rating you get back too!

I found when the RAM timings are near-optimal "balanced", when you have to use a divider clock (DDR-333 down from DDR-400, but 197/200mhz is used now)?

I get thru this test better, & operate on most things overall, better... trying to balance the speed of the CPU mhz-wise, & RAM mhz as well (trying for optimal default timings on divider use), + depends largely on what I am doing!

I could probably surf the web all day @ 2828++ mhz, but gaming, or a ROUGH program like this one is? Nope, no way... good test all around, for stability AND scoring.

APK

 

[cdawall]

this is at 254x9 ill do the test at 244x10 shortly

 

[cdawall]

the only change is the ram is oc'd higher cpu is at 2.44ghz !!! just short of the leader 1512.03 man if i had ddr400 i would have won :shadedshu i think that puts you at 4th alecstaar and me in 2nd w/ s754 amd a64 3000+ the click on the 2.44ghz and it is only for the 1323.22 ill eventually post the one w/ the ram at 354mhz

 

[_33]

the only change is the ram is oc'd higher cpu is at 2.44ghz !!! just short of the leader 1512.03 man if i had ddr400 i would have won :shadedshu i think that puts you at 4th alecstaar and me in 2nd w/ s754 amd a64 3000+ the click on the 2.44ghz and it is only for the 1323.22 ill eventually post the one w/ the ram at 354mhz

Those scores don't make sense!

You shouldn't do much past 1000 really. Don't ask me how you got this peak score, it's obviously a bug. The last post before does make sense thoe.

 

[dj_dn]

the only change is the ram is oc'd higher cpu is at 2.44ghz !!! just short of the leader 1512.03 man if i had ddr400 i would have won :shadedshu i think that puts you at 4th alecstaar and me in 2nd w/ s754 amd a64 3000+ the click on the 2.44ghz and it is only for the 1323.22 ill eventually post the one w/ the ram at 354mhz

Euhm, look at your primordia score it's 3092.28 vs core duo 2 1596 and FX-62 1357. Somthing aint right.

PS: dang _33 got to it faster then me.

 

[Alec§taar]

i think that puts you at 4th alecstaar and me in 2nd

Man, hate to tell you this, but I have to agree w/ the others here!

Your results don't make sense, if looked @ closely!

I say this, because I've seen it do some OUTRAGEOUS things already, see photo noted below!

It happens if you overclock TOO much (I did mine for stability too, & am running it as noted in my sig & for this test's results as well, stable as we speak & @ those rates below).

Those scores don't make sense!

You shouldn't do muych past 1000 really. Don't ask me how you got this peak score, it's obviously a bug. The last post before does make sense thoe.

Could very well be, his Primordia did the same thing my molecular did (progress bar out of bounds of its groupbox border even)

You can see that on the page of this post 3 pages prior to THIS one (954520.25 or, some outrageous score like that)...

Get ready for a laff, see my 5th post down's photo of my desktop running it & hitting a bug:

http://forums.techpowerup.com/showthread.php?t=14736&page=2

APK

 

[_33]

new result at 4285mhz.

you got me by 4 points alecstaar lol

Did you tweak your memory to achieve that score?

 

[_33]

Could very well be, his Primordia did the same thing my molecular did on the page of this post 3 pages prior to THIS one (954520.25 or, some outrageous score like that)...

Get ready for a laff:

http://forums.techpowerup.com/showthread.php?t=14736&page=2

APK

Congratz dude btw, your machine is quite the performer! On mine, I got value ram, so it's being held down a bit. If I had proper memory say OCZ platinum something, or some GEIL Ultra something, CAS 2, I'd probably hit something liek 1420 - 1450. But, reality is another thing right now.

You know also what's amazing is we get a head to head AMD64 vs P4 here, and it's quite neck to neck. I see on INTEL? usually the mem seems to hold it back. I could be wrong.

 

[cdawall]

http://valid.x86-secret.com/show_oc?id=109170 here would be my cpuz validation w/ ram tweaking sad how a cheap sub $100 system smokes you guys think if it wasn't summer i could have broke 250mhzx10 and really put up some numbers

 

[Alec§taar]

Congratz dude btw, your machine is quite the performer!

Yea, it's doing OK, stable (despite the heatwave, 101F here today)...

Plus, this resulted in my finding the upwardmost possible decent mhz boost on CPU, & moreso imo @ least on the memory hitting 197/200mhz possible for DDR-400 stock type RAM & operating @ 47C under load as well...

Good balance, overall for speed of CPU, memory, & bus - w/out "nuclear meltdown" temperatures or instabilities showing up (stable now for hours).

I am surprised (this benchmark turned out not only THAT, but also a GOOD stability of overclock tester, especially in the RAM area!)...

On mine, I got value ram, so it's being held down a bit. If I had proper memory say OCZ platinum something, or some GEIL Ultra something, CAS 2, I'd probably hit something liek 1420 - 1450. But, reality is another thing right now.

Mine too, my biggest "downfall/detractor" is my RAM, we noted that in the overclocking thread I did, & got a LOT of help on how it is done, & what to note/look for, etc. from others here...



* This benchmark turned out MORE than just a benchmark for me - but, a way to score my system vs. others AND find it's upper limits on overclocks as a stability testing program also: BONUS!

APK

 

[{JNT}Raptor]

http://valid.x86-secret.com/show_oc?id=109170 here would be my cpuz validation w/ ram tweaking sad how a cheap sub $100 system smokes you guys think if it wasn't summer i could have broke 250mhzx10 and really put up some numbers

Your not getting It.....No way your system gets 1500+....Look at the other rigs In this thread and you'll see your rig Is overshadowed by some real beasts.....your test was a Fluke....nothing more.....proving your rig Is more than likely NOT 100% stable In that configuration.........your first posted test score Is more than likely right on the money.

That being said....Heres mine...1334...not to shabby.....and will get better.

 

[_33]

Yea, it's doing OK, stable (despite the heatwave, 101F here today)...

Plus, this resulted in my finding the upwardmost possible decent mhz boost on CPU, & moreso imo @ least on the memory hitting 197/200mhz possible for DDR-400 stock type RAM & operating @ 47C under load as well...

Good balance, overall for speed of CPU, memory, & bus - w/out "nuclear meltdown" temperatures or instabilities showing up (stable now for hours).

I am surprised (this benchmark turned out not only THAT, but also a GOOD stability of overclock tester, especially in the RAM area!)...



Mine too, my biggest "downfall/detractor" is my RAM, we noted that in the overclocking thread I did, & got a LOT of help on how it is done, & what to note/look for, etc. from others here...



* This benchmark turned out MORE than just a benchmark for me - but, a way to score my system vs. others AND find it's upper limits on overclocks as a stability testing program also: BONUS!

APK

Yes totally! Thoe, for a game like, let's say Oblivion, I have to get that 2858 mhz down to around 2817 to be fully stable for hours on end. So at 2817, I'd probably get like a 1345-1350 score. I'm totally surprized by the performance of the Opteron 165! I wanted to switch to Opteron, and that might be it!

It's pretty hot in here BTW, won't do benchmarks today. But my memory is not very stable past 225 mhz CAS 3-3-3-8 with otherwize tight latencies and settings on other parms (DFI oblige).

 

[cdawall]

is there any progream you want me to rnu to prove 100% stable it runs pcmark05 3dmark2001se aquamark3 and superpi so umm.... it could be a glitch but hey i am happy about 2nd place maybe if i tweak the ram some more i can get 1st im not that far away ohh and i think the prob is not the prog i think it is amd s754 setup i am running the htt at over 1000mhz and the ram at 325.3mhz but the primordia score is kinda outta whack so watever ive run it 3x and got sim results so what now we email sciencemark about glitches oh and temp wise (i was watching my temp prog the whole time) was never over 54C

 

[Alec§taar]

Yes totally! Thoe, for a game like, let's say Oblivion, I have to get that 2858 mhz down to around 2817 to be fully stable for hours on end. So at 2817, I'd probably get like a 1345-1350 score.

Now, you're bringing up a point, a VERY solid one & I hit upon it as well earlier:

I noted, that I got BETTER SCORES on this test, with a higher FSB using a lower multiplier...

I had a 231mhz - 250mhz FSB x 12 multiplier, & got WAY higher CPU overclocks hitting into the 2828++ ranges, but instability on this test... & lower scores when it ran as well!

(That is, WHEN it went thru all of its tests fully, clocked up that high on the CPU (sacrificing HUGE on RAM mhz though, was way out of 'sync' @ 184/200mhz only) that is (was VERY "touch & go" getting the test to complete @ all))...

Now, @ the 2828++mhz CPU clock ranges?

Sure, I could run my system just fine & surf the web, do email, etc. but not pass this test, stable... & when I went back to a FASTER FSB, & lower multiplier (figures in my sig)?

I scored BETTER on this test, and it runs consistently & properly all the way thru!

This indicates to me that FSB deals in system communication w/ RAM onboard the mobo & that this test IS "memory bandwidth" intensive & TOUCHY if it is not stable...

I'm totally surprized by the performance of the Opteron 165! I wanted to switch to Opteron, and that might be it!

Yes, they ARE nice, & iirc, this mobo can take one too... they are also DualCore x2 jobbies!



* I will wait out their price break @ some point & possibly replace this one... but, don't you have to get ECC memory too?

APK