[Darph]

http://www.tomshardware.com/motherbo...119/index.html

This shows me,that for bartons, cas latency really doesnt seem to matter all that much.

[vidgion]

Wow, that was good reading to be honest.


see, toms isn't that bad :P

[hepp]

Couldn't be bothered to check each number, and I only read the Intel bits....

But in several benchmarks the difference between 2-2-2-5 and 2,5-3-3-7 was bigger than or abut the same as, the difference between the 3.2C and the 3.2EE.
I believe the price difference between 3500C2 and 3200C2 is still smaller than that between the 3.2C and the 3.2EE so I find the conclusion somewhat more suprising than the numbers.

Br
hepp

[Grandpa Dan]

Originally posted by vidgion
Wow, that was good reading to be honest.


see, toms isn't that bad :P

Is the converse: Tom's isn't all that good?

Agreed some good info, but an interesting way to compare AMD and Intel. Why not just state what each cpu does better with? not compare the two...

[vidgion]

yep grandpa, as i was reading through the article i said to myself, this is just tom's way of publishing more benchmarks of the intel platform outperforming the amd platform in a few tests.

[SpaceyWilly]

Originally posted by vidgion
yep grandpa, as i was reading through the article i said to myself, this is just tom's way of publishing more benchmarks of the intel platform outperforming the amd platform in a few tests.

Yeah I thought the same thing, and wtf is with them not testing the amd's at 2-2-2-6? BS they can't do it, that's what I usually run at. Until I read that article, now Im at 2.5-3-3-7

[gamefoo21]

My cl3-3-3-8 kingston(stock) pc3200 get me 91% of the available ram bandwidth. This is dual channel on a nf2 mobo.

[AudiMan]

I don't understand howcome he couldn't run 2-2-2 on an AMD XP system, yet he was able to do so on an Intel system.

[chasingapple]

I can second the results. I get better results when running the memory at a higher BUS Speed then lower with tighter timings. 11-4-4-2.5 is what I have to run at when @ 400Mhz and believe me that is nice and fast

[Eroc]

That is what i found too.

I ran my Corsair 2700C2 ram at like 190 (380DDR) at 2.5-3-3-6 and it smoked it compared to even 160s with 2-3-3-6.

[vidgion]

all this time i was pro timings mean everything, till i read that, but i've been doing memory benchmarks for like 2 yrs but i still didnt quite understand it so i went with what came close to making sense


but now im reevaluating the matter, in all of the synthetic benchmarks i've run, less latency really shows no major performance gain.

and just like the pentium platform, amd still perform well with high latency, of course theres more bandwidth that goes to waste but doenst hurt .

another reason i selected timings over bandwidth is because i was trying new timings and went to like 3-5-5-12 and started up one of my favorite games which is extremely memory/cpu extensive, and it ran sluggish enough for me to realize it in real-time.

so who knows what the correct formula is, one has to exceed the other because if not, we'd be stuck using pc2700.

[Speed_Mechanic2]

I wouldn't put much stock in tests done rather haphazardly.

[Smartiepants]

I found it an interesting read, but as always with Tom's you have to keep a bias in mind. But I found it quite good, just looking at the ram on each thing, not comparing the processors.