Sandra Memory Benchmark Question

[Salt]

I recently built a system and put what I believe is high quality RAM in it. After building the box I ran some benchmarks and was surprised to find the results below for my system's memory. I think I should be scring higher than all the reference systems on the list. Does the fact that I have 256MB RAM make the system slower overall? My processor benchmarks, and all other benchmarks are through the roof and beat all of the reference systems on all the lists.

What would make my mem benchmark lower than the other ones on the list?



My System's (Relevant) specs:
Abit BF-6
Pentium III 600MhZ Katmai
256MB PC100 RAM (CAS3)

[This message has been edited by Salt (edited 02-26-2000).]

[Psyduck]

The benchmark from sandra is using CAS2 ram, you ram is CAS 3 which is a little slower. It might be possible to get CAS2 out of your ram however. Try changing it in BIOS. You'll know if it can't as your get errors using Internet Explorer and opening directories, but they will be random.

[Erich]

You're lucky, when I try to benchmark my memory, Sandra crashes and does a (Not Responding) option. It sucks....!

[Hardware911]

try setting the ram in bios to cas2 as said earlier i did it and got scores close to 400 with my 500e at 620 using toshiba 128.

good luck

[Salt]

Okay. I set the CAS latency to 2 and achieved higher performance. I then set the CAS latency back to 3 and tweaked the following bios settings:

SDRAM RAS to CAS Delay: 3 (I left this as it was)
SDRAM RAS Precharge Time: 3 (I left this set as it was)
Cas Latency: 3 (I set this back to 3 to match my RAM)
Precharge Control: ENABLED (This was DISABLED on the previous 2 Benchmarks)

These setting yielded the highest score yet (I have not yet tried to set the CAS Latency to 2 and also enable precharge control.)

Below are my results with these settings.

My next question is what are the optimum settings for the above parameters that will yield maximum performance yet remaining stable. I have not had stability issues with my computer since I put in the new mobo and would not like to introduce any. I am a bit sketchy running my CAS3 RAM @ CAS2, due to the fact that I try to keep things within specs as stability is important to me.

Will running it at CAS2 shorten the life of the RAM, or does it usually cause problems?

Also, does Precharge Control being enabled introduce stability issues? From what I have read, it seems that it would not, as it seems to simply optimize the RAM paging between the rows.




[This message has been edited by Salt (edited 02-26-2000).]

[Hardware911]

im really not sure if it will hurt the ram...i dont thik it will...dont hold me to that i think it will just cause u to have an unstable system at worst...

i did this to mine and i havent had a crsh yet no problems here....with that at least..

i guess just play around with settings and beanchmark see what ya get...

good luck

ohh yea if ya gonna play be willing to pay!!!
if ya cant afford new stuff then leave it alone for now

[This message has been edited by Hardware911 (edited 02-26-2000).]

[bedrock]

hey salt,i had a similar complaint about my sandra memory benchmark. try setting your " in order queue depth" to the opposite of whatever you have it at now. i've got a 600e on an abit be6-2, and bx boards gave me the answer. memory bench went from about 130 to 320 or so. also si-soft sandra 99 is not compatible with the coppermine cpu's according to their webpage. something to do with the new cache on the cumine.
cheers

[NiteOwl]

setting queue depth to 1 cuts your memory benchmark by half. Setting it to 8 speeds things up alot. But one is really stable and allows you to o/c higher. But I wouldn't touch queue depth to 1 with a 10 feet pole. I doubt your queue depth is set to 1, cuz if it was your scores would be under 200 for sure. Your best bet is to try for RAM settings of 2's and see if your system can handle it. Just incase things aren't stable, you should backup your registry before hand.