Does anyone have an A64 3200 Venice that does not do 2.6 or higher???
I recently upgraded my wife's machine from a 3000 Winchester to a 3200 Venice and let it run stock for a while. This past weekend I finally got around to doing a little overclocking and got a very nice surprise. I started at 250=2.5 with only a small voltage increase and it went into Windows just fine. So I tried 260=2.6 then 270=2.7 with the same 1.40v and it is still folding at 2.7 for 37 hrs @ 43C under an XP-90 and Thermalright fan. Did I get lucky or is this par for the course? I am going to try more soon but want to keep this for a few days to be sure. This is also a testament to those Superflower PSU's I have been pushing on lately too. All rails are dead on!
0
Comments
Also, is your memory running on a divider? I'm really tempted to ditch my winchester, as the memory controller really sucks on it (cant do much more than 240mhz on the memory and retain stability). I've heard that these venice chips are really tolerant when it comes to higher memory overclocks..
If your memory can run upto and beyond 250mhz (PC4000) you can just leave the HTT:Mem ratio at 1:1 (6:6) or DDR400 (200mhz) If your memory is limited you want to run either the DDR333 ratio (HTT:Mem 6:5) or DDR266 ratio (6:4) or even DDR200 ratio (6:3). Even if your memory doesn't clock a hair over 200mhz (PC3200) with the DDR266 6:4 ratio you can reach a HTT300 clock because HTT300 @ (6:4) = 200mhz (PC3200) Memory
SO Set your CPU HTT multiple to 3x your Memory speed to DDR266 (6:4), or DDR333 (6:5) and then increase the HTT (Base clock) from 200 in 5-10mhz increments.
Playing around with CPU Clock multipliers is the advanced class
3000+ ( 9x) 1.8ghz
3200+ (10x) 2.0ghz
3500+ (11x) 2.2ghz
3700+ (11x) 2.2ghz
3800+ X2 (10x) 2.0ghz
3800+ (12x) 2.4ghz
4000+ (12x) 2.4ghz
4200+ X2 (11x) 2.2ghz
4400+ X2 (11x) 2.2ghz
4600+ X2 (12x) 2.4ghz
4800+ X2 (12x) 2.4ghz
http://del.icio.us/nosferatu/amd64
I've read that the 3000 won't match the 3200 and the 3500 is a dead match.
Seems to be the bargain chip of the moment.
can you bench it against the sd using the same vid card?
You want me to let the karma out?
Actually I would like to do my own test to compare the difference the cache makes by clocking each exactly the same in say 3 different increments each w/ only the CPU changed in the same system. I won't be able to do it for a little while though.
I certainly see where a 3500 Venice would be wasted money. Right now I like the idea of the extra L2 on my San Diego but I'm really beginning to wonder if that is wort it. Keep tuned for my direct comparison.
I can't get mine prime 95 stable above 2.4ghz without 1.7V :/.
NOTE:
If a mod would like to edit my poll for me so it included Newcastle core with a yes, over 2.6, not even close and a Winchester core with the same.
Keebs has a brand new Venice - maybe he'll chime in.
Looks like you've got a winner there, mtgoat.
Actually, I'm kind of waiting for EPoX to come out with an X2 capable bios for my 9NDA3+ mobo so I can splurge and get me an X2 3800 to play with. I love the way dually boards run so smooth. If it runs like I think it will, I'll probably make it my main rig and let my P4 machine become a dedicated folding box.
-oh and the poll needs a option for overclocking less than 2.6ghz but more than stock ;p
A bum CPU is a possibility and the Asus isn't as overclocker friendly as the DFI. But I do suspect that you just need a little tweeking to get it on track.
Lets see performace exhaust for my car or a new system.. dang they both will be faster than what I got now..
Settings? i have the HTT multiplier at 4x.. i changed the ram to ddr266, cpu multiplier to 10x, HTT bus to 253, pci express bus at 100mhz, chipset voltage 1.5v , HT voltage 1.2v, and the DDR voltage is at "auto"... its running a 420w Thermaltake psu. what other settings do i need?
-oh btw, i just tried to lower the vcore to 1.45.. it wont boot =/
I am going to guess your rails look something like this;
* 3.3 = 3.2ish
* 5.0 = 4.8ish
* 12v = 11.7ish
+12v = 11.84
+5v = 5.026
+3.3v = 3.376