Added a gig of RAM, 3dmark scores lower than before...

[50]

WTF! Switched from 512 to 1.5 gig's of RAM. Why did my score drop 300-500 points? Does adding an Audigy 2 lower your score?

 

[Cosmic_Horror]

are the ram timings in the bios the same? can you benchmark again with out the sound card to see if it has an effect? Or leave the sound card in and benchamar with 512mb, 1Gb and 1.5gb of ram and see if there is a progressive differeence?

Also do game and other realworld benchmark suffer this kind of reduction?
if not, i wouldn't even worry about it.. unless all you do with your computer is run 3dmark!

 

[bjc112]

I would think the timings have changed.

Check those for sure, did you add the same speed memory? If not, it slows all the memory down to the slowest stick.

 

[alexrocks]

From your sig it looks like you have an Intel system, so I think the problem might be that dual-channel mode only works with the same amount of ram in each channel. If you have two 512mb sticks in 1 and 3, and one 512 mb in 2 or 4, dual-channel mode won't be enabled.

 

[AgaBoogaBoo]

Originally posted by: alexrocks
From your sig it looks like you have an Intel system, so I think the problem might be that dual-channel mode only works with the same amount of ram in each channel. If you have two 512mb sticks in 1 and 3, and one 512 mb in 2 or 4, dual-channel mode won't be enabled.

That, is if he has a dual channel motherboard.

Also, its called PAT

 

[AgaBoogaBoo]

Originally posted by: bjc112
I would think the timings have changed.

Check those for sure, did you add the same speed memory? If not, it slows all the memory down to the slowest stick.

Thats what I was thinking. Also, is that 300-500 in 03 or 01SE? IF its in 03 there is something else, if its just 01SE, it doesn't really matter but its probably the latencies

 

[Corporate Thug]

the ram timings must have changed.

 

[50]

What does the ram timings changed mean? How do I change them to normal? Oh I forgot to mention a big issue that might be causing my lower scores. At first I had one GEIL ddr400 chip, but I bought 2 more kingston chips at ddr333 (I know the chips all go down to the lowest speed). I just wanted to see how my performance would be if I kept the GEIL and whether it was worth keeping. My mobo is not dual channel btw and also it is a DDR333 board that unofficially supports DDR400. Thanks for all the help guys

 

[Ionizer86]

Go into bios and lower latencies such as the Cas latency.

 

[50]

Okay I checked out Kingston.com and found out my CAS latency is 2.5. What if the GEIL ram has a different CAS latency? It seems to be auto detected in BIOS.

 

[bjc112]

Originally posted by: Cartman2003
Okay I checked out Kingston.com and found out my CAS latency is 2.5. What if the GEIL ram has a different CAS latency? It seems to be auto detected in BIOS.

Well if the Hyper X is 2.5, and it's not the GEIL "ultra dragon" edition ( or something to that effect ) it is probably 2.5 also..

You can try each stick out and see if they can run @ cas 2

Bump your ram voltage for added stability..

(2.8v)

Fits the bill.

 

[50]

Well I did set the RAM to CL 2.5. I just did some tests and my 3dmark03 score went up about 100 points. However my pcmark2002 went down on my CPU score (nothing big), my RAM went up a lot though and my HD went up a little.
Original Benchmark
CPU: 5828 Mem: 5123 HDD: 753
New Benchmark
CPU: 5811 Mem: 5780 HDD: 851

...I guess everything is alright???

 

[dmw16]

check your timings. I went from 512 to a gig and my scores went up.
-doug

 

[bjc112]

Originally posted by: Cartman2003
Well I did set the RAM to CL 2.5. I just did some tests and my 3dmark03 score went up about 100 points. However my pcmark2002 went down on my CPU score (nothing big), my RAM went up a lot though and my HD went up a little.
Original Benchmark
CPU: 5828 Mem: 5123 HDD: 753
New Benchmark
CPU: 5811 Mem: 5780 HDD: 851

...I guess everything is alright???

Could just be the test.. Not sure how consistent things are.. Sandra for one is almost ALWAYS bouncing around.

I would consider everything to be ok.

 

[chocoruacal]

You guys SERIOUSLY worry about 100 points in 3dMark?!?! I guarantee you that if I run the test 10 times, I'll get 10 different scores. Which is why you want to take an AVERAGE.

And 1gig of RAM isn't going to make a a noticable difference in 3dMark/games/surfing the net either way, so don't even waste your time benching it like that

 

[stevewm]

Originally posted by: alexrocks
From your sig it looks like you have an Intel system, so I think the problem might be that dual-channel mode only works with the same amount of ram in each channel. If you have two 512mb sticks in 1 and 3, and one 512 mb in 2 or 4, dual-channel mode won't be enabled.

Actually dual channel mode works just fine with different sized sticks. Many people seem to think otherwise though. You can have a 512MB and a 64MB stick, 512 and 1GB, etc... it really does not matter. Dual channel mode is not to be confused with IDE RAID

 

[REVNU]

Originally posted by: 50
WTF! Switched from 512 to 1.5 gig's of RAM. Why did my score drop 300-500 points? Does adding an Audigy 2 lower your score?

If you have a Dual Channel Mobo you need to take one of those sticks out. Running dual-channel will help free up some memory bandwidth...

The key is your RAM Timings!

I was getting scores in the 3200's with my Radeon 9700Pro before I decided to check my RAM timings. I am running 1 - 512MB stick of PC3200 Corsair XMS on an ASUS P4P800-D. The timings were at 2-3-3-6 and after I changed the timing to 2-2-2-6... My score immediately jumped to 4762! I know that it could be faster since I'm bottlenecking my system by not running another stick of RAM in Dual Channel Mode, but the score is pretty decent for a stock system and 1 stick of RAM. I'll be getting that extra 512MB stick soon... I'll let you know if that helps out the score anymore.

Still hard to believe that something that simple could make such a big difference....

 

[aka1nas]

Originally posted by: stevewm

Originally posted by: alexrocks
From your sig it looks like you have an Intel system, so I think the problem might be that dual-channel mode only works with the same amount of ram in each channel. If you have two 512mb sticks in 1 and 3, and one 512 mb in 2 or 4, dual-channel mode won't be enabled.

Actually dual channel mode works just fine with different sized sticks. Many people seem to think otherwise though. You can have a 512MB and a 64MB stick, 512 and 1GB, etc... it really does not matter. Dual channel mode is not to be confused with IDE RAID

I thought that P4s in dual channel are "real" dual channel and do need to be matched while Athlon Nforce platforms are actually "dual memory controller" and thus can have mismatched amounts of RAM in the 2 "channels"?

 

[Auric]

Originally posted by: chocoruacal
You guys SERIOUSLY worry about 100 points in 3dMark?!?! I guarantee you that if I run the test 10 times, I'll get 10 different scores. Which is why you want to take an AVERAGE.

Indeed. Plus, it ain't points that are important but percentages. 100 in 3DM03 prolly reflects a difference of 2% in your system which, as said, is statistically unimportant when considering normal variation.

And 1gig of RAM isn't going to make a a noticable difference in 3dMark/games/surfing the net either way, so don't even waste your time benching it like that

I got the 1GB religion with the DesertCombat and Forgotten Hope mods for BF1942 since all the extra stuff pushes the Commit Charge Peak near 800MB.