How memory timings on A64's affect performance - The truth.

This thread ought to be stickied in the Overclock FAQ...

Good results, well thought out and well presented.
 
Deadlierchair said:
This thread ought to be stickied in the Overclock FAQ...

Good results, well thought out and well presented.
Click to expand...

Hmm, maybe I could take it upon myself to do up an Intel version...... ;)

I'd better get back on track and finish the catalyst comparison first.....then I can work on starting a new project. :)
 
well I was just saying that it looks like you're looking at your average fps, rather than what actually happens with your fps. Look at it like this, if your overclocking makes the fps shoot WAY up in some easy parts, but bog down a little bit more, and with timings you get not as high spikes, but less drops, it could easily stagger the average FPS in the favor of the higher Hz looser timings, but the timings give better picture quality because you don't vary as much, which is what human eyes notice more than anything over whatever FPS people decide we can see.
 
I too think this thread should be stickied. It has a lot of helpful information that will help me throughout overclocking. Job well done.
 
Mr. Miyagi said:
I too think this thread should be stickied. It has a lot of helpful information that will help me throughout overclocking. Job well done.
Click to expand...

I'll second that......even though I've already "littered" the OC/cooling forum with 3 stickies....PM a mod and see what they think, you can never have to many informative stickies. :)
 
Stellar said:
Great post, but as has been stated already... the true benefit from those low-latency RAM sticks is they are particularly high-yield... meaning they can be pushed to higher frequencies in an overclocked rig which will DEFINITELY benefit overall real-world performance.

This is exactly why I'm fine with running HyperX PC2700 for the time being.
Click to expand...
This bears repeating... For the OC'er the value of low latency RAM isn't the performance gain of the lower latencies (at stock speeds or otherwise)... But the fact that these sticks were tested under higher scrutiny and thus it's highly probable that they'll reach higher speeds than the RAM which was only tested with higher (looser) timings under the same frequency.
 
So would it be more beneficial to run at say 230mhz @ 8-4-4-3 vs. 220mhz @ 7-3-3-2.5?
 
DaLurker said:
So would it be more beneficial to run at say 230mhz @ 8-4-4-3 vs. 220mhz @ 7-3-3-2.5?
Click to expand...
Well, it depends on what's the speed gain you achieve on the processor by going from 220 to 230 (which depends on your multiplier). There's a bigger gap between 8-4-4-3 and 7-2-2-2.5 than 8-4-4-3 and 7-3-3-2.5 as well for instance.

If you're gaining only like 100 more MHz on an A64 (say your multiplier is 10 and we're using your example) then chances are it'll be a close call, 100 MHz is gonna turn out to be like a 2.5% performance difference and the difference between the loose and tight timings will be around the same. So like I said, it really depends on the range and whatnot.
 
You must log in or register to reply here.
Back
Top