Timings: Performance Differences?

S

shockwave85

Gawd
Joined
Apr 29, 2002
Messages
761
Is there a significant difference between running an A64 system with 2-2-2-5 and 3-4-4-8 timings? In your experience, what performance changes have you seen?

Also, is loosening timings in order to gain more RAM speed really important for overall performance?
 
i personally haven't noticed any difference in normal computer usage between various timings and memory speeds, though some people claim they have. if you benchmark, the difference will show up in the numbers, but you shouldn't lose more than 3-5% in games, at the very most
 
Well that depends on what application you use frequently.

For me, as a graphic designer - I use Photoshop, Illustrator and etc on a 24/7 basis.
Such apps take advantage of the timings for sure. It may not be drastic, but honestly every second counts when you're on a tight deadline.

Gaming wise, I haven't really benchmarked it with different timings but from what I see with my own eyes during gameplay - I can't notice the difference.

Only way you'll definitely notice the difference is when you benchmark it. Other than that, if your system is more stable with slack timings - it shouldn't hurt to use higher.

Going from 2-2-2-5 to 3-4-4-8 is a bit of a jump. Why not 2.5-3-3-6? or of such number.

When people loosen timing to increase speed is done to overclock their overall system generally. So, it ends up increasing your systems performance as a whole.

I'm not really an expert with ram, so correct me if I'm off. ^^
 
shockwave85 said:
Is there a significant difference between running an A64 system with 2-2-2-5 and 3-4-4-8 timings? In your experience, what performance changes have you seen?

Also, is loosening timings in order to gain more RAM speed really important for overall performance?
Click to expand...

It also depends on the quality of the memory modules. Although normally the A64 memory controller is relatively immune to memory bandwidth deficiencies when going from 2-2-2-5 to 3-4-4-8 or even 5-5-5-15, there are a lot of 3-4-4-8 rated memory modules which are just plain cheap (and often are poorly manufactured and/or use crappy IC chips, and are thus substandard compared to the JEDEC reference for DDR400/PC3200 memory), and will noticeably drag down the performance of even a nominally immune A64 system.

Still, it's better to use good-quality brand-name 3-4-4-8 modules than crappy who-knows-which brand 2.5-3-3-7 modules.
 
E4g1e said:
Still, it's better to use good-quality brand-name 3-4-4-8 modules than crappy who-knows-which brand 2.5-3-3-7 modules.
Click to expand...
you're saying that price has an influence on performance given the same timings and mhz? :confused:
 
(cf)Eclipse said:
you're saying that price has an influence on performance given the same timings and mhz? :confused:
Click to expand...

Yes, price does have some influence on quality control. Some of the generics have had quality control issues due to their sloppiness in their assembly (in particular, those modules have never been tested for errors before they were shipped to resellers). As a result, you'll never know what you'll get when you order a "generic" or an unrecognizable brand of memory.

Also, brands of IC chips can vary somewhat in performance, even within the same timings and MHz.
 
E4g1e said:
Yes, price does have some influence on quality control. Some of the generics have had quality control issues due to their sloppiness in their assembly (in particular, those modules have never been tested for errors before they were shipped to resellers). As a result, you'll never know what you'll get when you order a "generic" or an unrecognizable brand of memory.

Also, brands of IC chips can vary somewhat in performance, even within the same timings and MHz.
Click to expand...
first statement is obvious, but you diectly said that an expensive stick of 3-4-4 will perform better than a cheap/generic stick of 2.5-3-3 (putting reliability aside for now..) i can't let that fly. even if you are right that some ic's are a little better at the same timings and mhz (simply because of the secondary timings that are set by the SPD, such as trrd), i'm willing to go as far as promising you that you won't find any cases that proves your statement true :p
 
The point is ALWAYS BUY BRANDED RAM! Like Corsair, Geil, OCZ, Crucial, Mushkin, Kingston.
 
well duh. i just don't buy the fact that there'll be a significant performance difference between any two sticks at the same settings.
 
Per what Eclipse said, it's an obvious comment that you should buy brand name stuff. The same goes for pretty much everything in life. Brand names are brand names for a reason.

Also, just to back up with evidence, my G.Skill Value feels the same in real world performance when it is OC'd to 230mhz or whether its sitting at 200mhz and with tight timings and loose timings. Benchmarking, sure, timings matter. But for gaming and application use only, I still can't figure out the rationale to spend double the money on RAM for an extra 5-8 frames per second.
 
You must log in or register to reply here.
Back
Top