View Full Version : Memory question
VulkanBros
12-11-2007, 01:00 PM
I have bought the sticks (4x1) very cheap, but....what are the best BIOS settings for these sticks :
5-5-5-15 at 1.8v or 3-3-3-10 at 2.3v
Kingston KHX6400D2ULK2/2G 2GB DDR2 800MHz Non-ECC
Ultra Low Latency CL3 (kit of 2 - 1GB)
The datasheet says:
Timing Reference: 800MHz 5-5-5-15 at +1.8V / 800MHz 3-3-3-10 at +2.3-2.35V
is 5-5-5-15 better than 3-3-3-10 or the other way around......or is it only better/worse under special situations ?
(like overclocking / benchmarking)
Hawk1
12-11-2007, 01:34 PM
3-3-3-10 will be better at the same clock speed (800MHz), but make sure there is airflow going over the sticks. At 2.3-2.35v, I would imagine they would get quite warm and need some good airflow if you want them to last at that voltage.
You would use 5-5-5 if you wanted to overclock the memory, say to 1066 or 1200MHz (if they can actually do those speeds).
VulkanBros
12-11-2007, 01:54 PM
thanx Hawk1....
I just did a bench with 3DMark06 first with
3-3-3-10 at 2.35v
and then with
5-5-5-15 at 1.8v
...and......I got 11 more point with 3-3-3-10
So I guess it isīnt worth running with 3-3-3-10 ... thinking about
the lifetime/health of the memory-sticks....
Hawk1
12-11-2007, 02:03 PM
thanx Hawk1....
I just did a bench with 3DMark06 first with
3-3-3-10 at 2.35v
and then with
5-5-5-15 at 1.8v
...and......I got 11 more point with 3-3-3-10
So I guess it isīnt worth running with 3-3-3-10 ... thinking about
the lifetime/health of the memory-sticks....
Right, at the same speeds, your not going to get much of a difference. If you took the memory to 1100 or 1200MHz, at 5-5-5 you may see better results, but again at higher voltages. I run my memory at 800MHz 3-3-3 2.2v simply because I cant get it to run at 1066 or 1100 without dropping my FSB so significantly that it drops total system performance. That, and my memory has lifetime warranty and I have a backup set(not as good) to use in case I have to RMA.
VulkanBros
12-11-2007, 02:32 PM
Just did another test :
3-3-3-10 at 2.35v with just 2 x 1 gig (instead of 4x1 gig)
still with 3DMark06....and....now I am confused:confused:......I get 98 more points
why that ? system faster with 2x1 than 4x1 .... can you explain that ??
I will test this with 5-5-5-15.....odd....
Ketxxx
12-11-2007, 06:12 PM
Unless you want to kill those sticks dont run 3-3-3 @ 2.35v without a fan blowing over them. see if you can run 4-4-4 @ 2.1v. Nice middle ground.
Ketxxx
12-11-2007, 06:15 PM
Just did another test :
3-3-3-10 at 2.35v with just 2 x 1 gig (instead of 4x1 gig)
still with 3DMark06....and....now I am confused:confused:......I get 98 more points
why that ? system faster with 2x1 than 4x1 .... can you explain that ??
I will test this with 5-5-5-15.....odd....
To save going into any complex details the long and short of it is the system is probably marginally more unstable with all 4 sticks and needs a smidge more juice. Of course, it could also just be that 98pts is within the margin of error, supposedly up to 200pts is within 3DMs margin of error, though I have never experienced a margin of error that big, only around 20pts.
VulkanBros
12-11-2007, 10:04 PM
right Ketxxx......I have testet my bud out .....
There are diff. between every test...one time 11, next 23 then 34 ...but I cant get the 98 point diff. anymore...mayby I made a mistake...OR Windows is not what is seems to be ...
Enough....this isīnt worth spending time on....barely 10 to 50 points gain.....
Thanx for your inputs guys and girls.....
vBulletin® v3.7.0, Copyright ©2000-2008, Jelsoft Enterprises Ltd.