should be okay, those timings are going to kill performance tho. amd's take a much bigger hit from timings than p4's do afaik.
Also the 3 or 4 people other than myself with this board running, you guys played any games yet? I get no sound in CoD after the intro for some unexplainable reason:confused:
My combo: A64 3500+ / Asus A8V Deluxe / 2x512MB Mushkin PC4000
I'm currently running this at 9.5x253 @ 1.55V with mem divider at 5:3 @ auto.
It's not Prime95 stable so that sucks. I'm using the stock heatsink/fan.
First thing I'm concerned with is my RAM. It won't run at 250 no matter what the timings/vdimm are.
----------------------
For the ones who have this board and have the CPU overclocked with the stock cooler, have you guys tested stability? If so, with what?
I'm hoping to have the Vmod done shortly. However, I really like the XL stuff. Mine does 228mhz at 2.0-2-2-5 at only 2.8v. 2.0-3-2-5 does 246 at 2.8v. Although I don't anticipate great things with xtra voltage on the XL sticks, I'm sure 250 would be possible at 2.0-2-2-5.Quote:
Originally posted by xgman
Anyone here have Corsair XL and pushed 3v+ in them yet effectively?
Same here. 1:1 is best. Although sandra gets a few extra points with higher HTT, most appsrun faster at 1:1 with the tightest possible timings.Quote:
Originally posted by sierra_bound
I can run 270X9 with 5:3 divider, but bandwidth takes a big hit. I get better SuperPI and Pifast times running 1:1 with the tightest possible timings.
Hopefully tonight:DQuote:
Originally posted by xgman
The regular vdim mod is easy. just clip on and go.
My 2 XL sticks are faring far better than that. But, my results may not be representable.Quote:
Originally posted by xgman
Some guy reported 250 at 2-3-3-5 with 3.2v on a P4. Bigtoe says extra voltage doesn't help OCZ's version of them that much so the jury is still out.
I'm still having random flakiness from the PCI/AGP lock I think. I need to test more. So at higher HTT speeds, it may be the board and not the mem. Just a thought.Quote:
Originally posted by nycxandy
It's not Prime95 stable so that sucks. I'm using the stock heatsink/fan.
First thing I'm concerned with is my RAM. It won't run at 250 no matter what the timings/vdimm are.
I got up to 2630 on the stock cooler no problem. I ran 3DMark2001 SE in a loop while crunching an 8MB string of superPi in the background. Memtest also flew with zero errors. One note...I was surprised at how much voltage it took to be that stable at 2630...it needed 1.725.Quote:
Originally posted by nycxandy
For the ones who have this board and have the CPU overclocked with the stock cooler, have you guys tested stability? If so, with what?
I hope to have more results to share with your guys tonight.
Yeah, right! So is Win XP 64 hehe!:rotf:Quote:
Originally posted by STEvil
if PCI-Express wasnt right around the corner i'd be all over that thing :D
MDZCPA, what HTT you running?
Thanks for this thread Charlie. You kept me form wasting a lot of money. I had just ordered a 3500+ and Asus A8V earlier today. After reading this thread I'd have to be out of my mind to pull my GA-K8NS-Pro and 3700+ for that combination.
Nope, from all the reviews and threads I've read the 939 has no advantage over the 754 (not for the user anyway).
Anyway thanks. Very informative (yet in many ways disappointing) thread!:toast:
:slobber:
I wish I was filth stinking rich... now I can only watch :cry:
LOL, I forgot to plug in my case fan. BAH!
Anyway... new findings.
I'm currently running at 2.4(9.5x253) with mem divider at 5:3(200MHz @ 2-3-3-11).
Idle temp at 45C. Load temp at 56C.
I ran Memtest86 for this memory at 230FSB and it errored. I hope to RMA it soon.
I'm gunning for 2.5GHz. Hopefully I'll be able to obtain this without a third party heatsink.
StormPC,Quote:
Originally posted by StormPC
Thanks for this thread Charlie. You kept me form wasting a lot of money. I had just ordered a 3500+ and Asus A8V earlier today. After reading this thread I'd have to be out of my mind to pull my GA-K8NS-Pro and 3700+ for that combination.
Nope, from all the reviews and threads I've read the 939 has no advantage over the 754 (not for the user anyway).
Anyway thanks. Very informative (yet in many ways disappointing) thread!:toast:
I don't understand, man...
The A8V/FX-53 combo is every bit as fast (probably faster) than the sk8v/fx53 combo which was ALREADY the world's fastest CPU/Mobo...
Don't bail out already.... this is JUST LIKE when the SK8N/FX51 first came out (8/03) and we were all cursing it...
It seems that MOST A8V DO HAVE a functioning lock, to be implemented via BIOS implementation.
Just wait until the bugs are ironed out and the v-mods are done.... I'm SURE we'll see 12 x 260 = 3080mHz 1:1 @ 260mHz, 2-2-2..... with 3.75v Vdimm on DI and decent phase change...
Those lower and slower 754 CPU's are always gonna be burdened and saddled with that LOW multi...
C
well it got pretty cool here after the thunderstorm so i figured i would try for 2600mhz.
http://www.mechwarrior.tv/pictures/SS/260x10pifast.jpg
(LOL in all my excitement i forgot to switch the 2nd cpu-z box over to memory. it was 216 at 2-2-2-5-7-16)
http://www.mechwarrior.tv/pictures/S...9.5superpi.jpg
(same time at 260x10 but i forgot to same the SS)
I think 3dmark01se would have passed but it rebooted in lobby, due to heat i have to presume since bios read 55c after it rebooted. will try at 270x9.5 to see what happens but i think with a better fan on the retail hsf 2600mhz at that voltage would have 3dmarked. I will get both at 260x10 with the memory shot working in a minites. going to try 3dmark01 at 270x9.5
I agree. Also, not many people with the A8V have benched yet with the newer and faster graphics cards. Once the X800XT and 6800U begin saturating the market, we will see which mobo/CPU combo is the real king. ;)Quote:
Originally posted by charlie
The A8V/FX-53 combo is every bit as fast (probably faster) than the sk8v/fx53 combo which was ALREADY the world's fastest CPU/Mobo...
I know they are faster but not enough to interest me. Maybe when the 90nm stuff ships things will be substantially better, but for now I'm really liking the GA-K8NS and NF3 Pro 250/3700+ combo, especially with the Raptors. The 3700+ cruises at 2620MHz 1:1 100% stable on air. (At that speed it will beat the snot out of most prommied P4s.) The K8NS seems a lot less problematical, even though I really love Asus boards.Quote:
Originally posted by charlie
StormPC,
I don't understand, man...
The A8V/FX-53 combo is every bit as fast (probably faster) than the sk8v/fx53 combo which was ALREADY the world's fastest CPU/Mobo...
Don't bail out already.... this is JUST LIKE when the SK8N/FX51 first came out (8/03) and we were all cursing it...
It seems that MOST A8V DO HAVE a functioning lock, to be implemented via BIOS implementation.
Just wait until the bugs are ironed out and the v-mods are done.... I'm SURE we'll see 12 x 260 = 3080mHz 1:1 @ 260mHz, 2-2-2..... with 3.75v Vdimm on DI and decent phase change...
Those lower and slower 754 CPU's are always gonna be burdened and saddled with that LOW multi...
C
I think 30-31k on air will do until the 90nm stuff comes.:toast:
good SS's of 260x10 superpi/pifast scores.
http://www.mechwarrior.tv/pictures/SS/260x10superpi.jpg
http://www.mechwarrior.tv/pictures/SS/260x10pifast.jpg
Puts me in 41's on the list behind tom holocks P4
Couldnt get 3dmark01se to run at 270x9.5 either, crashed between lobby high and nature so its got to be a heat issue since ive gotta be clipp'n close to 60c:grr: if the temps are reading accuratly.
I think superpi is close to tick'n over to 32s:dammit:
I guess JP's sending me my v5 thursday so next week sometime this sweet little honey gets a skin tight wet suit:up:
Nice...
I couldn't even get Windows to boot at 260x10 @ 1.65V.
What are you temperatures like at 260x10?
asus probe read 52c at 260x10 and bios read 55~55.5 at idle. i dont even want to think about what it hit that far into 3dmark LOL
Mongoose,
So what's the final verdict on locks??? Do yours work, and with which BIOS?
C
they must. i couldnt run at 240x10 asynch with the 1003 bios due to one of the worst ati cards for agp bus ive heard of(about 73mhz agp it :banana: :banana: :banana: :banana:s the bed). Ive been able to 3d bench at 275HTT and could prolly 2d at 280+HTT but it crashed 3dmark there but due to how complex the dividers are on a64's and how infamiliar i am with them i could have had something set inproperly in the boards defence.
I also tried my 2x256 xms3500c2 sticks at speed and was able to 3dmark at 230 1:1 also that would give a 77mhz agp bus if the agp locks were not working that i know for a fact my 9800xt couldnt handle. unfortunatly im runn'n IDE drives and no pci cards so i cant confirm or deny pci locks. Unfortunatly my limited understanding of this chipset kind of holds me back in being able to test confirmation of this sort:am: i am very confident that the agp locks at the very least work.
I wonder what's holding me back from that 260x10 overclock. My RAM probably?Quote:
Originally posted by Mongoose420
asus probe read 52c at 260x10 and bios read 55~55.5 at idle. i dont even want to think about what it hit that far into 3dmark LOL
probably CPU cooling...2600 on air isn't easy....Quote:
Originally posted by nycxandy
I wonder what's holding me back from that 260x10 overclock. My RAM probably?
Just eleiminate the ram by running a divider... if it runs, then it's the ram...if it STILL won't go then it's probably the CPU/cooling.
C
So far I'm right where I was with the sk8v/fx53 at 228x11.5 1-1., at 2-3-2-5-1. All stock voltages max and 020 bios showing locks. Temps seem higher than sk8v/fx53 . Just burning in a bit. Here Super Pi is 32. Anyway, I hooked up the same vmod as the sk8v and something must have been oiff, because a burning smell started and I kept seeing the voltages clim way high. So I took it off and surprisingly everythiung seems fine. What resistor are we supossed to use for the vmod in the lower corner pin 6?
MDZCPA, any luck on your vdim mod? What did you use?
Maybe your clip shorted two legs together?
well i decided to try my 3rd set of bh-5. 2 basicly unused other than slight testings 256 meg sticks of corsair xms3500c2pt. the result is
http://www.mechwarrior.tv/pictures/SS/240x10.5sync.jpg
only about 20 marks faster than 270x9 async tho even with the 100mhz cpu clock advantage.
Hi there...Quote:
Originally posted by Mongoose420
wierd. according to sisoft sandra im gett'n slightly better memory bandwidth at 270 5:3 than i get at 220 1:1 with only a 1mhz ddr mhz diffrence.
6467/6397 at 270 5:3 221mhz ddr 2-2-2-5-7-16
6403/6334 at 220 1:1 220mhz ddr 2-2-2-5-7-16
...I´d say that´s just a margin of error.
And in addition, that´s what I expected and what absolutely makes sense if you look at the ram speeds of A64.
A64 has no "async" memory speeds, it´s always generated by an integer divider "CPU/xx".
So just "slower" clock-speeds, but still synchronous.
And therefore I´m a bit puzzled when I hear something of a "big performance hit", because there should be no additional latencies. At 220 1:1, the CPU/11 divider for DDR400 (with CPU multi of x11) is used, at 270 5:3 the CPU/11 divider for DDR333 (with CPU multi of x9) is used.
Same RAM speed, same dividers...so any explanation for a performance hit? :confused:
Not from me...makes no sense in my eyes. I´d expect same performance with both settings.
Greetz, Flox
here's a spread of 240x10.5 sync 2-3-2-5-7-16 results
http://www.mechwarrior.tv/pictures/S...syncpifast.jpg
http://www.mechwarrior.tv/pictures/S...syncsandra.jpg
http://www.mechwarrior.tv/pictures/S...yncsuperpi.jpg
Also my take on it. but that was a responce to a post about running a divider killing there bandwidth.Quote:
Hi there...
...I´d say that´s just a margin of error.
And in addition, that´s what I expected and what absolutely makes sense if you look at the ram speeds of A64.
A64 has no "async" memory speeds, it´s always generated by an integer divider "CPU/xx".
So just "slower" clock-speeds, but still synchronous.
And therefore I´m a bit puzzled when I hear something of a "big performance hit", because there should be no additional latencies. At 220 1:1, the CPU/11 divider for DDR400 (with CPU multi of x11) is used, at 270 5:3 the CPU/11 divider for DDR333 (with CPU multi of x9) is used.
Same RAM speed, same dividers...so any explanation for a performance hit?
Not from me...makes no sense in my eyes. I´d expect same performance with both settings