• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

P4 & Dualchannel performance

R

RS3RS

[H]F Junkie
Joined
Dec 10, 2003
Messages
8,197
Well, after hearing everyone saying a P4 running without dual channel is a major no-no, I decided to switch from 768 mb of non-dual channel ram to 512mb of dual channel RAM (traded some RAM around with my dad, he got a gig of RAM out of the deal and I got dual channel). I own a 2.66b processor.

Anyway, I got bored tonight and decided to test out how much dual channel helps, and... Here's what I came up with.

SiSoft Sandra memory bandwidth benchmark:

Dual Channel: 3443/3445
Non - Dual Channel: 2219/2215

No big surprise here.

SiSoft Sandra CPU multimedia benchmark:
Dual Channel: 18526/23141
Non Dual Channel: 18580/23209

Well, I doubt memory affects this all that much, but the non dual channel is very slightly faster.

Unreal Tournament 2004, torlan map, 1024x768 @ 32 bit, default quality settings:
Dual Channel: Average of 50 FPS
Non Dual Channel: Average of 48 FPS

Rainbow Six 3: Raven Shield, 1600 x 1200, default settings:
Dual Channel: 59 fps
Non Dual Channel: 57 fps


Maybe it's just me, but... Am I missing something here? :confused:
 
Dual channel only offers a maximum of 10% increase in performance theoretical.!
So what you see is expected.

enjoy
 
Try a real benchmark; the games you mention probably aren't stressing main memory bandwidth.

I'm thinking something like a gcc bootstrap ought to give main memory a good workout.
 
Well, I guess I was just wrong. From what I'd heared, I always thought with a pentium system the difference was like night and day. I'm sure the dual channel does help with some stuff, but games are what I do the most that stresses the system, and there doesn't seem to be much of a performance increase there.

Maybe the fact that I'm not running a really high FSB processor has something to do with it. My 2.66 B is at 3.2 ghZ, 640 mhZ FSB, sort of between a "B" and a "C" processor.

I dunno, I just expected the difference to be greater. Maybe I'm a noob :p
 
i'd say it's your slower bus speed that you're not seeing the huge jump in bandwidth...and your ram is also a factor.
what kind of speed and latency are you using?
 
it is faster. So why are you complaining.

You have 50% less ram but its still faster.
Dont forget you went from 768 to 512

Also, like someone else said what speed/ratio & timings are you running your ram at.

It would make a bigger difference if your FSB was faster too.
 
Try running Q3demo timedemo, you should see a pretty big difference.
 
I went from dual 512 to single channel 1 gig and saw a huge difference in battlefield 1942 and unreal Tournament and any games all around. I think dual channel sucks. If i had a gig of ram that did dual channel that would be a different story
 
I went from single channel to dual old score was 3218 new score 5386 big difference to me. Everything is so much smoother
 
Based on a i875P Chipset motherboard (200Mhz FSB) and PC3200 DDR 400 SDRAM I think the bandwidth calculation works like this;

Dual Channel (2 x 64 Bit Mem Slots for a total of 128 Bit) DDR (2 x 200MHZ FSB Speed) for 6.4GB Memory Bandwidth (128 / 8 = 16 x 400 = 6.4GB)

Single Channel (1 x 64 Bit Mem Slot for a total of 64 Bit) DDR (2 x 200MHZ FSB Speed) for 3.2GB Memory Bandwidth (64 / 8 = 8 x 400 = 3.2GB)

So if the App your running doesn't saturate 3.2GB/Sec Total Bandwidth there's no advantage whatsoever, if however your running Apps that are memory Bandwidth Hogs your golden :) Also these numbers are theoretical maximums and what you'll actually get may be a bit lower due to a variety of factors. You can modify the numbers to suit your FSB speed etc..

Hope this helps
 
Originally posted by MemoryInAGarden
Sometimes more RAM is better than faster RAM.
Click to expand...

Yeah, but you sure can't go wrong with MORE ram that is FASTER!

But what do I know...I'm still pissed at my geil....should have known better than to get a deal like that on ram... :rolleyes:
 
A Lance Corporal is a E-3 in the United States Marine Corps.
Its my Rank.

You Start a Recruit in boot camp
Then you are a Private E-1
Private First Class (PFC) E-2
Lance Corporal E-3
Corporal E-4
Sergeant E-5 and so on...

:) good ram thread
 
yea i knew lance corporal had todo with rank but i didn't know quite where...thanks for clearing that up

hey RS3RS you never told us your ram speeds or timings....
 
You must log in or register to reply here.
Back
Top