PDA

View Full Version : DDR 550 benches - need help



ryanpgroovy
07-12-2003, 05:20 PM
OK boys heres a list , these are random screenies of what I have been playing with

PC-4000 gold 4 x 256
ASUS P4C800
2.4C


These are all 1:1 with sandra unbuffered scores I am pretty certain I am loosing some bandwidth due to the OS / Driver and sandra version combination , the unbuffered all look about 200Mbs shy of what they should be


What I am trying to do here is get the highest possible Unbuffered score , but as you will notice , none of my benches make a bit of sense



I will let the real guys give benches later as this is not really my forte

this is with 4 x 256 btw guys stable.


Herse what I ran at CL 3


Heres 555 at 1:1 3-4-4-6 timings ,

http://www.ocztechnology.com/images/awards/sand5543446.jpg

553 1:1 3-4-4-7 timings

http://www.ocztechnology.com/images/awards/sand5533447.jpg


544 1:1 3-4-4-6

http://www.ocztechnology.com/images/awards/sand5443446.jpg



542 1:1 3-4-4-7

http://www.ocztechnology.com/images/awards/sand5423447.jpg



ok as you see above setting tras to 6 seems to have no positive benefit


I started working on my CL 2.5 benches , heres a few to start you off with

533 1:1 2.5-3-4-5

http://www.ocztechnology.com/images/awards/sand5332345.jpg

533 1:1 2.5-4-4-7

http://www.ocztechnology.com/images/awards/sand53325447.jpg


and my highest Unbuffered ever which really confused the heck out of me

522 1:1 2.5-3-4-7

http://www.ocztechnology.com/images/awards/3600.jpg




Does anyone have any clue , why I need to lower the speed to get a higher unbuffered score.

Amoeba Assassin
07-12-2003, 05:27 PM
Sick numbers!

Anyways, I have that problem too. At 2-3-3-8 I score better in PCMark than with 2-3-3-7. Odd yes, but I guess thats life. Anyways, whens that stuff gonna be on the market?

Holst
07-12-2003, 05:48 PM
Nice scores :D

I cant offer any real advice as im not much of an Intel man (cant afford a decent CPU)

It might be worth trying another bench as well though.

Sandra memory is an innacurate POS IMHO. :P

Xenogias
07-12-2003, 06:00 PM
seems like lowering cas latency had a big benefit for you. dropping ras to cas down to 3 also gave you a big jump even at the cost of mhz speed. Might be something to keep in mind for future Gold series DDR. Mind telling us what voltages you used for these benches?

ryanpgroovy
07-12-2003, 06:09 PM
hmmm

Xe,
I actually got higher benches at 522 2.5 -3-4-7 than at 533 2.5-3-4-5

saaya
07-12-2003, 06:43 PM
cycle legth should be cas+x+x+2 ...how was that rule again?

Any tRAS setting lower tRCD + CAS + 2 cycles will allow the memory controller to close the page “in your face!” over and again and that will cause a performance hit because of a truncated transfer that needs to be repeated. Along with those hassles comes the self-explanatory risk for data corruption. That one is not a real problem as long as the system is kept running but in case it is shut down and the memory content is written back to the hard disk drive, the consequences can be catastrophic. For the drive, that is.

ryanpgroovy
07-12-2003, 07:16 PM
tras is handled a bit difrently in dual channel mode though
from what I recall

ryanpgroovy
07-12-2003, 10:10 PM
anyone else ?

saaya
07-12-2003, 10:29 PM
i dont know why, but thats... normal.
http://www.xtremesystems.org/forums/showthread.php?s=&threadid=15193

check the pics posted in this thread. have you used memtest? maybe you get 1bit errors at 233 so it didnt crash but 22 gave you a better score...

Creative
07-12-2003, 10:29 PM
Nah I commented on it over at the other forum so prob cant really add much more here....:)

I think it comes down to the timings not be so crash hot at the higher fsb, whereas at the lower fsb, the timings are tighter, thus giving better scores?

Not sure :)

ryanpgroovy
07-12-2003, 11:33 PM
the problem is I get higher benches at 522 2-3-4-7 than at 533 2-3-4-5

xigfrid
07-13-2003, 12:11 PM
hope that you are influent in OCZ manufacturing objectives : You are probably the 1st OCZ employee that discover that CAS2 memories (like PC3500EL) are more effective than all the others CH-5 / CAS 3 thing of the latest PC3700 gold.

Btw has OCZ any exchange waranty for those who bought the Gold series ? they have incompatibility with the ABIT IC7 (like every CH-5 in the market)

ryanpgroovy
07-13-2003, 12:25 PM
xigfrid


I might actaully be able to help you fix your problem , so you dont have to warranty the ram , but I will gladly warranty it if needed

drop me an email at oczguy@ocztechnology.com

eva2000
07-13-2003, 01:18 PM
Originally posted by xigfrid
hope that you are influent in OCZ manufacturing objectives : You are probably the 1st OCZ employee that discover that CAS2 memories (like PC3500EL) are more effective than all the others CH-5 / CAS 3 thing of the latest PC3700 gold.

Btw has OCZ any exchange waranty for those who bought the Gold series ? they have incompatibility with the ABIT IC7 (like every CH-5 in the market) same memory and same IC7

@233 2-3-3-7 at 2.8v
http://www.xtremesystems.org/forums/showthread.php?postid=182103#post182103

@250 2.5-4-3-7 at 2.8v
http://www.xtremesystems.org/forums/showthread.php?s=&postid=182742#post182742

or it could be your particular sticks

saaya
07-13-2003, 01:21 PM
read!

http://www.xtremesystems.org/forums...&threadid=15193