ScienceMark thread (with score list)

[Urlyin]

there that should make you happy made my primordia drop happy now i changed some of the settings and ran it at 2.44ghz and all my other settings what ever the prob was fixed when i changed the element thingy in primordia

Changed what element thingy? I thought the test was ran at it's default settings ... any changes would skew the test and result in exactly what happened... an unbelievable score... but none the less, that is a good score for a 3000+ cpu on your last run posted

 

[cdawall]

no the last one when i got a 1023.22 i changed the element that is the one that was non default all the others ran the full time and all the same test you guys did even on that i am going to change my dram timings to get a higher score and run again at default and we will see if i can top my near 1600 pt. score

 

[cdawall]

ok i found a big glitch in primordia and set the bar in primordia 8k haha

 

[Alec§taar]

ok i found a big glitch in primordia and set the bar in primordia 8k haha

I don't know about you guys, but I ran ALL the tests @ their default set (no user-defined stuff etc. on ANY test)... first of all.

Secondly, man... you really OUGHT to write the program's developer on your Primordia results... seriously!

AND, do point him to my "Super-Result" on the Molecular test too, the photo of it... cannot hurt!

(We may end up w/ a better version in 3.x because of it!)

APK

 

[cdawall]

yeah prob should but hey watever i redownloaded the prog and it seems to be having less funny results ill post new scores later

 

[Deleted member 24505]

hello alecstaar.

my ram is this-http://www.patriotmem.com/products/groupdetailp.jsp?prodgroupid=63&prodline=5&group=PC2-7200&catid=2

i have had it to 1000mhz at 3.75ghz,at the moment it is at its recomended setiings of 450/900 at 5.5.5.15 at 2.3volts.compared to ddr1 on an amd the timings are really high,i hated it at first but i know ddr2 has higher timings but more bandwidth.on sandra it is about 8000mb/sec.and i know games like high bandwidth.

i too ran all the tests at default settings.

 

[Alec§taar]

hello alecstaar.

my ram is this-http://www.patriotmem.com/products/groupdetailp.jsp?prodgroupid=63&prodline=5&group=PC2-7200&catid=2

i have had it to 1000mhz at 3.75ghz,at the moment it is at its recomended setiings of 450/900 at 5.5.5.15 at 2.3volts.compared to ddr1 on an amd the timings are really high,i hated it at first but i know ddr2 has higher timings but more bandwidth.on sandra it is about 8000mb/sec.and i know games like high bandwidth.

i too ran all the tests at default settings.

Thanks man... I have 2 types from 2 folks so far that look REALLY good, as far as more "overclock ready" types of RAM, vs. the stock/oem DDR-400 Corsair stuff I use now!

APK

 

[Alec§taar]

Yes, yet again? Achilles from the film "TROY" said it best:

"Is there NO ONE ELSE!?!"

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):

APK

 

[cdawall]

i have done testing and realized i could oc up to 2.47ghz but had to drop my htt from ~976mhz (w/ my 2.44ghz setting and 800mhz htt) to ~741mhz @2.47ghz is the extra 30mhz worth the loss of 235mhz of htt??? i know this is pro the wrong spot but i trust you guys to tell me right so there it is????

 

[Alec§taar]

i have done testing and realized i could oc up to 2.47ghz but had to drop my htt from ~976mhz (w/ my 2.44ghz setting and 800mhz htt) to ~741mhz @2.47ghz is the extra 30mhz worth the loss of 235mhz of htt??? i know this is pro the wrong spot but i trust you guys to tell me right so there it is????

It's funny you ask that, because I am wondering THAT myself, especially during the ScienceMark 2.0 competition here this week!

IMO?

I did a "BETTER" o/c using 11x multipliers (even better using "fractional" ones my BIOS offers in 11.5x) as far as regaining RAM-mhz (memory tab section of CPU-z 1.35) than using 12x multipliers, especially concerning DDR-400 -> DDR-333 "hit" I took using "dividers" as the guys here call this technique!

12x multipliers (2773mhz CPU though) only allow me a 185/200mhz possible on RAM speeds...

11x multipliers ("only" 2750mhz CPU here) however, get me 196/200mhz (avg. of 11x-11.5x ratings I saw in CPU-z)

* Which IS better, indeed, is the question to ask!

APK

P.S.=> Imo, in theory @ least?

For SOME things, higher RAM-mhz return rates is going to benefit you more than a few mhz gained on CPU-clock rating... memory-intensive ops WILL or should, gain more by it (faster FSB = faster CPU->Memory->& back circuit speeds)...

ON THE FLIP-SIDE?

IMO, Higher mhz on the CPU will gain in CPU intensive tasks, albeit you will probably lose SOME mhz on the RAM-mhz speed regain (if you used "dividers" as I had to, going DDR-400 down to DDR-333 rates, & THEN, doing the "ram tightening of memory timing-chain numbers" as I call it to TRY get more of RAM-mhz back)...

I don't THINK there is a "perfect answer", hence WHY I am exploring 12x multipliers @ ALL in the first place (heck, I wanted 2.6ghz, that's all... I get WELL OVER 2700++ ranges now because of all of this - might as well investigate ALL of the possible avenues imo & do tests, then lastly, making a decision!)

"B A L A N C E - balance" quoting Sean Connery as Ramirez in "The Highlander" movie...

IMO @ least? Heck - it's all about that, best optimal return you can get doing BITS of both (high cpu mhz, vs. high RAM-mhz return)...apk

 

[cdawall]

you know what is funny though with my htt down the cpu runs cooler so is this a goo trade off??? near identicle performance identicle in super pi so what should i do cooler is better right??? it sucks i thought i had the thing up to 2.5 and the dam thing give me some f***ing blue screen in xp stupid pos had to clock it down and i mean i tried everything dropping ram down to 100 spec (they make that???) and losening timings etc. i got a screen that said shuting down win. to prevent damage whats up w/ that???

 

[Alec§taar]

you know what is funny though with my htt down the cpu runs cooler so is this a goo trade off??? near identicle performance identicle in super pi so what should i do cooler is better right???

Absolutely!

By the way:

33 gave me a REALLY nice tool for monitoring temps, light, unobtrusive, & FULLY functional down to the last temps details (& more, would be a GREAT addition to CPU-z code imo)...

Heck, it even runs in the tasktray w/ a tooltip for monitoring temps "live"...

IN fact, I want to write the coder of it, & give him the "secret" to LOW->NORMAL->HIGH->REALTIME cpu usage code!

All so his app can run @ LOW while minimized (on CPU usage) & normal when made fullscreen/normalsized... very easy to do via the Win32 API!

it sucks i thought i had the thing up to 2.5 and the dam thing give me some f***ing blue screen in xp stupid pos had to clock it down and i mean i tried everything dropping ram down to 100 spec (they make that???) and losening timings etc. i got a screen that said shuting down win. to prevent damage whats up w/ that???

Oh, lol, man... I must have "BSOD'd" 20x already during my overclocking!

(Especially initially, before the guys here "turned me on/tuned me in" to HOW to do the memory tighten cycle. Not doing it FOR me, but explaining the reasoning & WHY's/HOW's of it (ALL OF IT, especially POGE & Ketxxx), i.e.-> giving me a pole & teaching me how to fish, not fishing for me!)

BSOD's are not 'bad' man, they are there to PROTECT YOU, against you!

* Try think of them THAT way when o/c'ing...

APK

P.S.=> I have also had the CPU itself shut the system down, overheating is why (33's tool turned me onto THAT fact, when I was blowing WAY over 60c (danger temp iirc, for AMD cpu's))... apk

 

[cdawall]

lol bsod sucks but as you can see my ram settings cant be tightened any more though considerin how much nicer they are on xp than 98se (i got blue screen of death a lot on 98se) bte when you upgrade to some real ram like this stuff on sale at tiger http://www.tigerdirect.com/applications/SearchTools/item-details.asp?EdpNo=1820448&CatId=558
(there is a 1gb stick on sale to would you mind sending me the ddr 400 you got ????????? it would be very nice to have some real ram in my system oh and umm most temp monitors dont work on my machspeed board i basically have core temp (from _33) and speedfan and thats all that work

 

[Alec§taar]

would you mind sending me the ddr 400 you got ????????? it would be very nice to have some real ram in my system oh and umm most temp monitors dont work on my machspeed board i basically have core temp (from _33) and speedfan and thats all that work

LOL, the RAM I use? It BLOWS by comparison to what others here use!

(I know it's what's holding me back from using 1:1 ratio (no divider usage, ala DDR-400 speeds step downs to DDR-333 & having to pull the "memory loosen/tighten" cycle))

Corsair "matched pair" DDR-400, 2x256mb sticks - this is NOT "overclocker ready" RAM by any means... trust me on that account!

* I.E.-> I can assure you, that it won't even go "over volt" (2.6v std.), not even to 2.65v or 2.7v... which is WHAT you can do to try "get more" from your RAM during O/C, just like overvolting the CPU or chipset!

APK

 

[cdawall]

lol you know considering i am running ddr2100! anything is an improvement and i know what overvolting is my ram now runs at 2.65v over my stupid f'd up boards 2.55v? but hey i am hitting near ddr2700 w/ it (329.4) w/ the timings 2.5-2-2-6-9 2T which should run relative to 2700 spec. plus im poor white boy in houston (go figure) so there is honestly no money to upgrade the ram and if i did have the money it would go to a high end hsf vs the one that came in the box (lol did you know that crappy little fan can hit 4700rpm)

 

[Alec§taar]

lol you know considering i am running ddr2100! anything is an improvement

You DO have a point there, lol!

and i know what overvolting is my ram now runs at 2.65v over my stupid f'd up boards 2.55v? but hey i am hitting near ddr2700 w/ it (329.4) w/ the timings 2.5-2-2-6-9 2T which should run relative to 2700 spec.

Your system's pretty impressive imo, & IF you scores are indeed "legit" (which is WHY I asked you write the author of ScienceMark 2.0, so he can verify it, one way or the other, & so that this "err" you & I have BOTH seen, is error trapped for! It obviously isn't @ this point imo @ least!)

If your scores are "straight" though? I wouldn't sell that rig... I'd upgrade it to its max & keep it as a server in years to come...

plus im poor white boy in houston (go figure) so there is honestly no money to upgrade the ram and if i did have the money it would go to a high end hsf vs the one that came in the box

No rich man here by any stretch-of-the-imagination/means @ all, too!

(lol did you know that crappy little fan can hit 4700rpm)

Nope! But, I guess, now I do...



* Not a wasted day, live & learn!

APK

 

[cdawall]

err had to change ram sett ing 2.5-3-3

 

[Alec§taar]

Well, that said on your end? Heh, gotta do it (It's "tradition" in this thread)

"Is there NO ONE ELSE!?!"

* All comers are welcome!

IMO @ least so far? Heck, this is a decent benchmarking program AND stability test in 1 box (2 for the price of 1 FREE program no less)...

APK

 

[cdawall]

well looks like stability at over 2.44 is to much to ask w/ stock cooling

 

[{JNT}Raptor]

well looks like stability at over 2.44 is to much to ask w/ stock cooling

The 754 platform has a Roof Limit of 800Mhz HTT...not 1000Mhz HTT like the 939 platform....more than likely...thats what was causing the Extreme Primordea scores.

Makes sense as well with getting a good temp drop....you were way over the 800 Limit.

Congrats on getting It sorted.

 

[Deleted member 24505]

i might have a dabble today and try and see why it fails at the magic 300fsb.it is a bit cooler here today so i might take the chance and try.

has any body any ideas?.more vcore?.more ram volts?.

some guy had his to 4.8ghz at 1.57v so i am pretty sure some more is there.i just need to find the sweet spot at 4.4ghz.

maybe i need to run the mem lower at that speed.

 

[_33]

The 754 platform has a Roof Limit of 800Mhz HTT...not 1000Mhz HTT like the 939 platform....more than likely...thats what was causing the Extreme Primordea scores.

Makes sense as well with getting a good temp drop....you were way over the 800 Limit.

Congrats on getting It sorted.

Actually cdawall has admitted a couple posts earlier that he personalized the test to obtain high scores and so he cheated. higher HTT usually doesn't give much (if any) performance improvements.

 

[Tatty_Two]

i might have a dabble today and try and see why it fails at the magic 300fsb.it is a bit cooler here today so i might take the chance and try.

has any body any ideas?.more vcore?.more ram volts?.

some guy had his to 4.8ghz at 1.57v so i am pretty sure some more is there.i just need to find the sweet spot at 4.4ghz.

maybe i need to run the mem lower at that speed.

memory will be a limiting factor, if the overclock is pushing your memory speeds at or exceeding its limit you will need to loosten your timings....if you are on 1T....just start by dropping to 2T before you ramp up to 300Mhz rather than changing everything at once. Voltage of course will also be an issue and if you have the cooling you may just want to start by raising her by .25 in any case, usually you will find that a speed that is attained by a lower multiplier and therfore higher FSB requires more volts than the same speed attained by a higher multiplier and therefore lower FSB.....make sense? Of course this in itself will have an impact on memory speed so you really need to try all permitations to see what sits best with your rig. Take things easy and be patient! don't get too over enthusiastic and try it all at once or you may be in need of a new CPU!!

 

[Alec§taar]

Actually cdawall has admitted a couple posts earlier that he personalized the test to obtain high scores and so he cheated.

That's NOT acceptable...



APK

 

[{JNT}Raptor]

Actually cdawall has admitted a couple posts earlier that he personalized the test to obtain high scores and so he cheated. higher HTT usually doesn't give much (if any) performance improvements.

I realize higher HTT value over stock doesn't give much If any performance boost.....But It can give Instability In benchmarks and app execution.

If In fact he did Indeed admit to cheating...he should be kicked because he sat In this thread and bragged his A** off about "Smoking" the higher end rigs In this thread.

Cheating only shows the true nature of an Individual..