DFI better oc

vcv1122

Distinguished
Sep 20, 2003
80
0
18,630
I have an Amd64 system and am looking for ideas on how to oc more, if possible. 267 fsb is as high as I can go without memtest errors. System is on air (Thermaltake x90 heat sink) Temps, cpu @36-41c, chipset @41-45c, system @ 39-44c. Any suggestions are appreciated.

Amd 64 3000 (winnie) @ 2.404 ghz, Multy 9, 1.5v
Patriot xbl, 1 gb, ddr 400 @ ddr 530, 2.5-4-3-7, 2.8v
DFI Lanparty Nforce4 Ultra-D (03/10/2005 bios)
Athena 500W psu
fsb 267, ldt/fsb x 4, ltd bus 16/16
 
Wusy, I took your advise, lowering LTD to 3x and tried to up the HTT but it seems whenever I go higher then 267, I get error messages with memtest. I also tried higher ram voltage and a little tighter timings, but still couldn't pass. It seems that this may be as good as it gets (although 2.4 ghz isn't bad). Any other suggestions? Thanks for your input.
 
You are indeed correct. I should clarify that I attempted looser timings yet still couldn't pass memtest over 267. I tried using a LTD 3x and couldn't gain any advantage at all, including raising FSB or leaving FSB @ 267 and tightening timings a bit. Question..I tried a 3-4-3-7 to no avail. If I loosen timings further and I gain 10 mhz, would this translate to better performance? Or is it better to run tighter timings and slower fsb?
 
scottchen,
I tried those timings, bumped up fsb from 267 to 275. Still lots of errors on memtest. Also tried uping ram voltage (2.9, 3.0, 3.1 and 3.2v) still same results. Also, I tried higher cpu voltage..1.55v..still same results. Seems regardless what I do 267 mhz is a cap, unless there is some other setting I have wrong. But, thanks for your suggestion.
 
I already found this out. Any voltage over 2.8v with this PDP seems to work against me. I have read differing reviews on voltage for Patriot, one said max performance was achieved with voltages over 3.2v and looser timings while another says voltages over 2.8v degrade performance. I, like you, know the latter to be true.