Logo Logo The top.

The banner.  Yeah.

Stay informed, man.

Forums

Fuck Scott Walker

Features

Books

Comics

Film

TV

Music

VG

PC

Nintendo

PS2

XBox

Indie

Tech

Politics

Religion

Red Light District

Conspiracy

Facebook Idiot of the Week

Blog Moron of the Week

YouTube Fuckhead of the Week

Myspace Loser of the Week

Livejournal Moron of the Week

Multimedia

Raiderfeed

FAQ Contact

E-Mail Hate Mail!

Best viewed in 1280x1024

The Daily Raider is brought to you by the Project for an Unamerican Century and the Ronnie Gardocki Beard Preservation Society. The Daily Raider accepts donations, but we will only use them for liquor, cocaine and South American prostitutes.

 

RAM Timings

by vyralsurfer

Well Scooby, looks like we got another mystery on our hands

I've always wondered something about DDR memory: is it better to have it running at a fast Front Side Bus speed, or will the best memory timings be achieved at a lower bus speed? Logic says that a faster clock-rate would make the memory more efficiently used, but I think that having less of a pause between operations would be the fastest. The pauses between reading/writing are just a few cycles , but that really adds up when you think about how much your memory is used in just one second.

Get your tools and let's split up, gang!

The software that I decided to use was PerformanceTest by PassMark Software. I liked the simplicity and the layout of this program, and will probably use it in future articles. I ran all available memory tests for each run. The memory in my system is a 256MB module of Corsair ValueSelect PC3200 DDR. The first test I run is with my best timings at 200MHz bus speed. In the BIOS, it's:

  • CAS: 2.5

  • Trp: 3

  • Tras: 6

  • Trcd : 3

There were a few other timings, but they remained the same for both tests: as low as they could go. The second test is run with 166MHz bus speed and the lowest timings I could put in:

  • CAS: 2.0

  • Trp: 2

  • Tras: 6

  • Trcd : 2

Set the trap! Shaggy and Scooby, you're the bait...again

Here's a comparison of the results I got from the tests:

 

It was Mr. Jenkins; the old man that runs the "adult-goods" shop!

Well, as much as I hate being wrong, and I could just pull an Enron on these tests, I do find the results interesting. Being more significant of a change in some areas than others, having RAM at its highest rated bus speed with tweaked timings is the best way to get the most out of your memory. I'll extend on this next week with the same tests run on lower-quality memory.