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

RAM Speed vs Timings

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

doesnotcompute

Registered
Joined
Feb 28, 2003
What would give the best performance? RAM running at 150MHz with slow timings or ram running 125MHz with fast timings?
 
The example that you used was a little extreme. The 150 would definately be much faster, because you are also running your cpu at that speed. But a better comparison would be 150mhz with low timings vs. 140mhz with aggresive timings. With that comparison the 140 with aggresive timings would give the 150 a run for its money. There the faster timings would close the gap.
 
TUK101 said:
The example that you used was a little extreme. The 150 would definately be much faster, because you are also running your cpu at that speed. But a better comparison would be 150mhz with low timings vs. 140mhz with aggresive timings. With that comparison the 140 with aggresive timings would give the 150 a run for its money. There the faster timings would close the gap.

I agree with that statement as well. Just adjusting the timing slightly on my ram (PC2100) got it running faster than PC2700. This is of course w/ a 1:1 ratio on my overclock.
 
Since I don't personally know how timings affect the utilization ratio of bandwidth generated by clock speed, I can't answer that question, but you can benchmark to find out.

Anyway, the faster the clock speed, the higher the THEORETICAL bandwidth. However...

The tighter/faster the timings, the higher the UTILIZATION of that bandwidth. Fast timings = more util / more sends / more recieves. Slower timings = less util / more maintenance.

Personally, I think a simple change like cas2 @ 166 is better than cas 2.5 @ 180, and my benchmarks would bear that out, at least on this machine.
 
i have two systems, one running ddr 333 at slow timings and one running ddr 266 at very aggressive timings, the 266 beats the 333 in SANRDA's mem benchmark.
 
Last edited:
Please correct me if I'm wrong here, but since my mobo runs best in sync, I should consider timings like icing on the cake right? Since I'll probably be limited by my CPU's fsb before my RAM's can I can "use some of the slack" in RAM to increase timings?
 
Ugmore Baggage said:
Please correct me if I'm wrong here, but since my mobo runs best in sync, I should consider timings like icing on the cake right? Since I'll probably be limited by my CPU's fsb before my RAM's can I can "use some of the slack" in RAM to increase timings?

EDIT, I mean decrease timings >> better
 
If I understand you correctly, it is definitely worth testing 133mhz synced @ super fast timings vs 166mhz async with slower timings.
 
ccb056 said:
i have two systems, one running ddr 333 at slow timings and one running ddr 266 at very aggressive timings, the 266 beats the 333 in SANRDA's mem benchmark



Hi :) . Can you please tell me what your system exactly is and the scores you get ? i feel i could do much better whith my system, but i don't know how...
 
the ddr266 system has a xp2100+ w/ 2*256mb modules, 1 single sided and 1 double sided, the timings are 2-2-5-2-1 and the SANRDA scores are

1996 @93%
1907 @89%
 
InThrees said:
If I understand you correctly, it is definitely worth testing 133mhz synced @ super fast timings vs 166mhz async with slower timings.

As I understand it, definitely.
At least worth a test. I understand all this on more of a theoretical level than practice.
 
doesnotcompute said:
What would give the best performance? RAM running at 150MHz with slow timings or ram running 125MHz with fast timings?


The best way to find out is to benchmark at each setting.. I myself have found out that running at 200fsb with timings 5,3,2,2 gives me about 300 more points in 3dmark03 than running at 210fsb 6,3,3,2.5.... Even though the fsb is higher at 210 it is taking extra clock cycles for each command. It's just like running a higher fsb is almost always better than rasing your multiplier.. Like running a cpu at 10x200=2000mhz is better than running that same chip at 12.5x166=2075mhz...


nYmph
 
Re: Re: RAM Speed vs Timings

nymph said:



The best way to find out is to benchmark at each setting.. I myself have found out that running at 200fsb with timings 5,3,2,2 gives me about 300 more points in 3dmark03 than running at 210fsb 6,3,3,2.5.... Even though the fsb is higher at 210 it is taking extra clock cycles for each command. It's just like running a higher fsb is almost always better than rasing your multiplier.. Like running a cpu at 10x200=2000mhz is better than running that same chip at 12.5x166=2075mhz...


nYmph

But if you can run it at 210 6-3-3-2, I think it will beat 200 at 5-3-2-2 or even 6-2-2-2. There is about a 10 MHz (rule of thumb) difference between cas 2 and cas 2.5 for 3D. I always keep things at cas 2.

Many memory spec at cas 2.5 can make it to cas 2.
 
Last edited:
Re: Re: Re: RAM Speed vs Timings

hitechjb1 said:


But if you can run it at 210 6-3-3-2, I think it will beat 200 at 5-3-2-2 or even 6-2-2-2. There is about a 10 MHz (rule of thumb) difference between cas 2 and cas 2.5 for 3D. I always keep things at cas 2.

Many memory spec at cas 2.5 can make it to cas 2.

yes it would be close at those settings you listed and might just beat it, but what I said before was 200@5,3,2,2 would be better than 210@6,3,3,2.5

At least for me it was, (giving me that 200+ more 3dmarks)


nYmph
 
Back