FSB

[mickeywu]

I was just wondering if the fsb of a cpu has any direct effect on the ram. For example, if i have a duron with fsb 200MHZ, can i use it with DDR266? If so, does it take full advantage of it?

 

[juin]

Just think about a chain of production
The 1 move at 200 mghz the others at 266 mghz.Does the chain of production run at 200 mghz or 266 mghz for 5 point

Just next to the lab and the bunker you will find the marketing departement.

 

[lhgpoobaa]

And now for an answer you can understand:

You have two things to consider: Memory bandwidth, and Latency.
The bandwidth is determined by the system bus speed, not the ram speed.
so having a duron at 100 and you ram a 133 isnt very efficient.
But as ram speed increases the latency reduces, reducing access times, so there usually IS a SMALL benefit in running the ram at 133, provided the memory timings are the same.

So in your case
100fsb/100ram cas2 is good
100fsb/133ram cas2 is a bit better
but
100fsb/133ram cas3 will probably not be better.

Usually the ram speed is set to a multiple or fraction of the fsb: FSB + PCI is a common speed for asynchronous operation.



<b>You will die.
If you are lucky you may last ten decades.
If you are especially lucky you will be
remembered for another 5 decades.</b><P ID="edit"><FONT SIZE=-1><EM>Edited by LHGPooBaa on 01/22/03 00:26 AM.</EM></FONT></P>

 

[scamtrOn]

what poo said

<font color=orange><b>as you get older, your hard drive becomes floppy, but don't fear viagra is here. viagra puts the hard back in your drive!!!

 

[mironto]

well, on my MSI 745 with duron 1300 and one 128 PC2100 is 100/100 cas2 a bit faster than 100/133 cas2. Measured in SiSoft Sandra.

 

[Civilized]

tell sandra she is retarded

 

[scamtrOn]

lmao

<font color=orange><b>as you get older, your hard drive becomes floppy, but don't fear viagra is here. viagra puts the hard back in your drive!!!

 

[mironto]

if sandra's retarded, then 3D mark is retarded too, because there I get the same lower results with 100/133 compared to 100/100

 

[lhgpoobaa]

yes. 3dmark is retarded as well. Far too cpu dependent.

And what mem timings are you running? Hopefully the same for 100 as for 133.

Even if they are you have hidden problems with asynchronous ram... namely slighly more lax "hidden timings" and the problems of asychronous operation.

<b>You will die.
If you are lucky you may last ten decades.
If you are especially lucky you will be
remembered for another 5 decades.</b>

 

[juin]

Sandra benchmark are retart use Spec cpu for CPU steam benchmark for memory and HD tach or intel i/o for HD.I personaly better like HD tach as amd lemming will not believe any benchmark come from intel

Just next to the lab and the bunker you will find the marketing departement.

 

[mironto]

maybe sandra and 3d mark are retarded, but I'm not
Of course I use the same setting for 100 and 133MHz, namely DRAM timing conf: turbo; CAS 2T; RAS pulse w: 5T; Row precharge time: 2T; RAS to CAS latency: 2T. Even if I lowered the settings (for both tests) I got higher results in 100/100MHz

 

[mironto]

I realize that those benchmarks are CPU dependent, but hey, when I use my PC I use also CPU, I don't want useless numbers that will tell me that running memory alone at 133MHz is faster than on 100MHz, and I'm not trying to convince you of opposite. I'm just trying to tell you, that on my Duron/MSI745 running memory at 100/100 is faster in cooperation with CPU than on 100/133, that's all to it.

 

[juin]

ok i just tell you to try under a others benchmark

Just next to the lab and the bunker you will find the marketing departement.

 

[mironto]

I tried also STREAM and Mem Tach and guess what? the same result as from Sandra... There must be some slowdown through asynchronous operation of RAM with CPU.

 

[juin]

synch operation will be faster like that a big issue on RDRAM and dual channel

Just next to the lab and the bunker you will find the marketing departement.