Hipro is right, how about some 'real game' performance and not just 3Dmark?
Printable View
Hipro is right, how about some 'real game' performance and not just 3Dmark?
ok, i tested again with different timings.....man, trust me: timings doesn't bring a lot to performance - look at this (whired) results:
it brings just TINY, TINY more performance (especially @ SuperPI)
9x272,5MHz @ 2.5-4-4-8-1T / stock-GPU
2k1: 24247points /// SuperPI-1M: 34sec /// 4M: 2min 51sec
--------------------------------------------
9x272,5MHz @ 2.5-4-4-7-1T / stock-GPU
2k1: 24211points /// SuperPI-1M: 34sec /// 4M: 2min 51sec
--------------------------------------------
9x272,5MHz @ 2.5-4-3-7-1T / stock-GPU
2k1: 24855points /// SuperPI-1M: 34sec /// 4M: 2min 50sec
--------------------------------------------
9x272,5MHz @ 2.5-3-3-7-1T / stock-GPU
2k1: 24752points /// SuperPI-1M: 34sec /// 4M: 2min 49sec
What are the differences in stability and max clock:
2.5-3-3-7 vs. 2.5-4-4-8 ?
Then run the same tests...
I've just ordered 1G (2*512) of Gskill PC4400 2.5 3 3 7 (280€) and i will test it with my fx53, I hope more than 275 and why not 300 2.5 3 3 7 with more volts!
Ok, Ive got my new DFI 250GB and finally tested my Gskill PC4400 1GB.
First with 1 dimm:
2.7V Vddr.
http://www.xs4all.nl/~famchow/Memory...R600%20004.jpg
http://www.xs4all.nl/~famchow/Memory...400/DDR600.JPG
Just received my pair of G.Skill today and I'm kinda dissapointed with the results. At 2.85v 2-2-2-5, I can run SANDRA up to 210Mhz only and at 2.5-3-3-7, there's loads of memtest errors at 275Mhz......
I feel so like tossing them out of the window right now and go grab some of the OCZ with the new revision of TCCD! LOL
just relax a little
new revision? all TCCD's are the same...OCZ does not make them. (chips)Quote:
Originally Posted by arj
Can you loosen up the timing a bit? How high will it go at cas 2.5?Quote:
Originally Posted by arj
what board are you using? (CPU?) which bank? (1/2 1/3?) Vdimm?Quote:
Originally Posted by p4z1f1st
I cant pass 265Mhz 2.5-3-3-5 1T on the DFI with zero errors in test 5 memtest .. with 2 dimms inserted :confused: And i see you doing SPD speeds :(
I know that Samsung makes the chips. It's just that according to the OCZ guys in this forum, with the new batch of chips that they (OCZ) received, Samsung made some modifications/revisions to them that makes them respond better to higher voltage and oc better.Quote:
Originally Posted by kakaroto
I need to loosen up to 3-4-4-8 to get it memtest stable at 275. Trying to find the memtest stable speed for 2.5-3-3-7 right now.Quote:
Originally Posted by mike.elmes
@ kakaroto: me too :/
i have "a few" errors at each test#5-loop (0-10errors).....and it's only stable @ 266MHz with SPD.....so, i don't know what to think about G.Skill now :(
i'm using dimm1/2 on k8n neo2....
I have the same problem. Could get it down to about 40 Errors each Pass.Quote:
Originally Posted by p4z1f1st
The only stable setting is with voltage set to 'Auto' in BIOS (275Mhz 2-3-3-7 1T) (CoreCenter shows 2.70v VDDR then). But whenever I set the voltage manually, even when it's set to 2.70v I get errors in Memtest #5.
ATM I'm running 280Mhz 2-4-4-8 1T @ auto voltage.
Does any1 know why I get only about 7GB/s bandwith while onepagebook had > 8gb/s ?
See this screenshot
it's normal, the bandwith in sandra depend of the frequency of the cpu!Quote:
Originally Posted by digitalfrost
for example, fsb: 225 2 2 2 5
11*225: 6536/6470
10*225: 6228/6128
9*225: 5853/5778
Its running Dual Channel right? CPU-z says Single...Quote:
Originally Posted by p4z1f1st