• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

I think my RAM is a bottleneck. What is the best solution?

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

Secret-Ian

Registered
Joined
Jun 29, 2003
Location
GTA
*For those of you who like short posts, here are my questions, otherwise skip this and read the whole post:
1.) Do better RAM timings really make a difference, and are they the only difference between qualityRAM and ValueRAM?
2.) Do you think that running in dual channel mode instead of single channel mode will solve my ut2k3 performance problems? (benchmarks at bottom)
3.) Is it worth it to get 1Gb of RAM instead of 512? Is there a noticeable difference, and if there is, approximately what % could I expect my ut2k3 benchs to increase?




Hey guys. I just built this system, and I was initially happy with the synthetic bench scores, but then when I ran my real world game test (UT2K3, benchmarks at bottom of post) I was very dissapointed with my scores. It seems like a system with a similar setup to mine would be getting like 80 fps in ut2k3 in 1024x768 with noAA+AF. The system which I compared mine to had this configuration:

A7N8X mobo
Barton 3000+
1gb of PC2700 RAM - unknown timings, but prolly better than mine (dual channel)
RADEON 9700 pro

My system has this (full system specs are below, as well as my benchmarks):

A7N8X mobo
Barton 2500+
512mb of PC3200 ValueRAM - ****ty timings, like CAS 3.0 4-4-7 (single channel)
RADEON 9800 pro

Anyways, I realise the biggest difference between my computer and the one I am comparing to is the RAM. Even so that system scores 20 fps higher than me! And I have a better vid card! Also about the timings: I suspected this could be my problem, but going from CAS3.0 4-4-7 to CAS 2.0 2-2-6 makes pretty much no difference, but the one time I actually played a game with those settings, the system crashed (not BSOD, just frozen screen and I had to hard boot). I dunno, should I expect a difference when changing timings or is it normal for there to be very little difference? I have heard that with modern RAM, timings barely make a difference, but maybe I am wrong and my valueRAM just runs like **** no matter what the timings are.

So basically my question is this: Is it possible that my RAM could be bottlenecking my system? And if it is, which of the foollowing choices is the best (going by price vs performance)?

a.) picking up another 512 ValueRAM to run in DC w/ a total of 1Gb RAM($115 CDN) or

b.) Buying something like golden dragon DC512mb QualityRAM ($180 CDN) or

c.) Going all the way and getting golden dragon DC1Gb QualityRAM ($250 CDN)

Thanks a lot for your help guys!

Here are my system specs and benchmark scores:

3DMark2K1:
3DMark Score.................15463
Game1(Low detail)............206.3 fps
Game1(High detail)...........67.0 fps
Game2(Low detail)............302.8 fps
Game2(high detail)...........156.9 fps
Game3(Low detail)............175.6 fps
Game3(High detail)...........76.1 fps
Game4........................130.8 fps
(All display settings set to "High Performance")

3DMark2k3:
3Dmark score.................5530
Game1........................166.2 fps
Game2........................38.2 fps
game3........................33.4 fps
game4........................34.4 fps
CPU Score....................508
CPU test 1...................57.2 fps
CPU test 2...................8.9 fps
(All display settings set to "High Performance")

PCMark2k2:
CPU score....................5706
Memory Score.................5317
HDD Score....................666
(All display settings set to "High Performance")

Unreal Tournament 2k3:
-800x600-
4xAA and 8xAF(FlyBy).........179.4 fps
4xAA and 8xAF(BotMatch)......63.1 fps
noAA and NoAF(FLyBy).........192.3 fps
noAA and NoAF(BotMatch)......63.3 fps
-1024x768-
4xAA and 8xAF(FlyBy).........135.3 fps
4xAA and 8xAF(BotMatch)......62.0 fps
noAA and NoAF(FLyBy).........189.7
noAA and NoAF(BotMatch)......63.3
-1280x960-
4xAA and 8xAF(FlyBy).........95.1 fps
4xAA and 8xAF(BotMatch)......45.7 fps
noAA and NoAF(FLyBy).........168.8 fps
noAA and NoAF(BotMatch)......63.2 fps
-1600x1200-
4xAA and 8xAF(FlyBy).........60.5 fps
4xAA and 8xAF(BotMatch)......31.1 fps
noAA and NoAF(FLyBy).........121.8 fps
noAA and NoAF(BotMatch)......60.8 fps
(All display settings set to "High Quality")

System Info:
Windows 2000 Professional Service Pack 4
Athlon XP 2500+ @ Stock
ASUSTeK Computer INC. A7N8X2.0 REV 2.xx (NForce2 driver ver. 2.45)
BIOS: Phoenix Technologies, LTD ASUS A7N8X2.0 ACPI BIOS Rev 1005
HL-DT-ST CD-ROM GCR-8520B
HL-DT-ST RW/DVD GCC-4480B
MAXTOR 6L060J3(60.04 GB)
512Mb PC3200 ValueRAM, single Channel mode (for now) CAS 3.0 4-4-7
RADEON 9800 PRO @ Stock (Driver ver. 6.14.10.6368)
IBM G78 [Monitor] (15.7"vis, s/n 55-K5150, December 2000)
Creative SB Audigy
 
Crazy.... why you wrote so much? The answer is simple: the system you are comparing is single against double channel. Of course your single channel PC3200 is no match to double channel PC2700. This is bandwidth issue.

Secondly, its the processor you are comparing: Barton 3000+ versus 2500+ ???? you need to OC yours to 3000+ level.

My suggestion: Find another identical 512Mb PC3200 ValueRAM. No need to go 1 GB in single channel setup.
 
Actually I tried overclocking my proc to 2800+ speeds (I dunno, it wouldnt run in 3000+ and I didnt really want to try that hard to get it working that high with my stock cooling). There was an increase, but only about 5 fps...
 
Back