OC'ed my A8V Deluxe and 3200+ Winchester, weak results

[adeathangel]

Hi, so to start out, here are my specs, then I'll go through with what I did.

Asus A8V Deulxe
Athlon64 3200+ Winchester core
4x256MB Kingmax PC3500
Asus V9999GT Geforce 6800GT

So I read the forums for a few hours, printed out Zebo's guide, and got started. Like he said, I isolated. I put the multiplier down to like, 8x, ram on the loose SPD, and went to work with the Max HTT.. (I had the HTT to 4x, once I got past 250MHz I put it on 3x) I got to the Windows Login screen at 285MHz, but it would crash at 290MHz... so I figured 275MHz was a safe ceiling.

Next, I set the multiplier to 10x, HTT to 3x, and the FSB to 275MHz.. Bad stuff.. I had to clear the CMOS and start over. Next I tried 260MHz with the same settings and got past POST, but Windows wouldn't load. So I went down to 250MHz finally being able to boot into Windows.

Then I loaded up Prime95, and went to the Torture Test. 250MHz and 245MHz failed within 30 seconds. Then the idea of bringing the FSB down to 230MHz and running my ram at DDR400 2:1 ratio in the BIOS struck me. Supposedly my ram is rated to do 234MHz with 10-5-5-3, of course I had 1T enabled, like Zebo said is a must if you want performance. So I bring her down to 230MHz but leave the ram loose, and run Prime95, it runs fine for 5 minutes, so I decide to see if 2:1 will work. Windows wouldn't load, so I had to use the 5:3 divider. The OS was unstable because it crashed after about 20 minutes of just browsing and chatting on AIM. So now I'm at 225MHz HTT with 6x HTT mult and 10x cpu mult. The ram is on the 5:3 multi with 10-5-5-3 and I don't know what to do. These results seem pretty low compared to what other people have achieved. Maybe my ram is to blame?

Also to note, the CPU idles at 38C with my current OC. I have stock cooling.

Here are some screenshots of what I'd done.

This one is when I ran Prime95 for 5 minutes.

This screenshot is my unstable settings with HTT230MHz

And this is my current OC.

I hope this has been enough information for you all to figure out what's been going on and what, perhaps, I've done wrong, or can do. Oh oh.. The VCore I've had at 1.450V and 1.475V and the Vdimm at 1.7V, just trying give as much info as I can think of.

If anyone has any suggestions or help, or anything really, please reply. You guys are really great and helpful, and I'm glad to be a member of these forums. Looking forward to see what we can do with my situation. (By the way, how do you post pictures? <-- n00b hehe)

~A Death Angel

 

[GuitarDaddy]

To get your best overclock, you will most likely need to drop your CPU multi, increase your HTT and use a memory divider

I have the 3200+ on an ASUS A8N-SLI and with the 10x multi and DDR400 the highest I can get is 10 x 245= 2450mhz

But if I drop the CPU multi to 9x and use the 6:5 divider (DDR333 on my board, I think 166 on yours) I can get 9x289= 2601mhz with a ram speed around 240mhz

As you gradually increase HTT, you will have to increase the voltages for your CPU and ram

 

[Thor86]

Drop your HyperTransport settings to 600 (3x) and use Limited Memory Mode and set it to 266 mode (4:3 I think?). This should isolate your memory from running too high, and see what your max HTT (FSB) is on your cpu. I would leave the multiplier of the cpu at 10 to start.

Post back with your results. You may also want to up your vcore to 1.55 to see if your cpu needs more juice to Prime95 test.

 

[adeathangel]

Alright, the results are in. I couldn't get anything to run at a somewhat stable state until I upped the VCore to ~1.6V ... I thought that was a bit high for a Winchester, so I don't know how to explain it. A curious observation I made is this.. when I change the VCore in the BIOS to anywhere from 1.400V to 1.550V, the BIOS as well as CPU-Z report the VCore to be ~1.478V, no matter what setting I have on it. And when I set the VCore to 1.575V or higher, they report the VCore at ~1.64V.. So my question is this the BIOS that's screwing up or what? I can see why I couldn't OC to a higher freq when the VCore was at 1.478V.. at least, that seems like a plausible reason why.. but 1.6V+ seems kind of excessive. So there's the VCore situation, the vdimm is at 2.8V and the Hyper Transport I set to 3x (600MHz) with the HTT at 240MHz. The CPU multiplier is 10x giving me 2400MHz clock speed. I have my ram at the loose SPD timings, 3-5-5-10 1T, and it passes memtest86 no problem with the 5:3 divider setting it at exactly 200MHz. Unfortunately my idle CPU temperature is kind of high at 42C with the stock cooling, and I suspect it's because of the unusually high VCore of 1.648V that's currently being detected, check the screenshot link below. Prime95 ran for 20 minutes no problem before I stopped it, figuring I'll stress it for 12 hours later. Unfortunately, I don't think this is a stable set-up because the OS keeps crashing every 20-30 minutes or so. It also crashes when I reduce the HTT to 230MHz and leave everything else the same. So I want to know a few things,
1) What could be up with my VCore? Edit: I was thinking flashing to 1009 BIOS???
2) Aside from lowering and lowering the OC until it's stable.. is there anything else I can dabble with to attain stability? For now I don't want to buy a different cooling system, I just want to OC to 2.4GHz, maybe 2.5GHz if it's possible.
3) Do you have any suggestions for what else I can be doing, or am perhaps doing wrong, or could do differently?
4) The screenshot also has the memory bandwidth benchmark from Sandra, is that a decent score?

HERE IS MY SCREENSHOT

Edit again: Another thought about the instability is my PC3500 Kingmax ram chips. Although I'm running them at 200MHz, maybe they're just crappy (even though they pass memtest86 at that speed, could that still be a hang-up?) Honestly, I don't really know what to think or look for to figure out what's causing the instability, but I see people OCing HTT to 250MHz with mult 10x no problem on stock cooling and I want to get that too, at least for principle's sake.

 

[adeathangel]

Btw, thanks for posting replies GuitarDaddy and Thor86. (=

 

[Jeff7181]

Doesn't that board use weird PCI/AGP locks... like say from 200-230 it'll use a divider to keep the PCI/AGP buses in near spec, then when you go to 231 to like 260 it uses a different divider?

Also are you using any SATA devices? That board may not lock the SATA controller's frequency either and that could be the cause of instability.

*EDIT* By the way... that doesn't sound much like Zebo's guide... maybe you should re-read it and attempt to follow it a bit more closely and take things a bit slower. It's pretty optimistic to attempt a 750 MHz overclock right away.

 

[adeathangel]

GuitarDaddy, that's pretty good for your ram man. Mind if I ask what kind of ram it is you have? My ram isn't very stable past 200MHz, but I'm not surprised, I bought value Kingmax ram.. (even though it's PC3500, rated for 234MHz it doesn't pass memtest past 220MHz or so..) Right now I'm stress testing 240HTT with 10x multi. 245 wasn't stable, but so far so good on this. I have the Hyper Transport at 4x and the ram divider at 5:3 making it exactly 200MHz. If this ends up being a stable config, next I'm gunna try a 9x multi with high FSB and a 3:2 divider to see what I can get.

 

[GuitarDaddy]

Yeah this ram rocks! I just wish this board(A8N-SLI) would let me run higher than 240, its supposed to do 270-275 with timings like 2.5-3-3-8

I'm currently running it 236mhz 2-3-3-5 w/2.8v

OCZ EL PC3200 platinum rev.2 (the new Samsung TCCD chips) kind of expensive $270-280 for the 1gb kit

 

[Thor86]

What bios version is on the board now and is it a Rev2.0 board? Rev1.0 boards have flakey AGP/PCI locks from what I read on the net. If you have a Rev1.0 board, you need to find an older bios I think 1005 to get proper AGP/PCI locks.

My Asus was a Rev2.0 with Bios 1008 already on it, and the AGP/PCI(SATA) locks were working no problems - using the VIA SATA controller.

 

[adeathangel]

I have a rev2 that came with 1008.003 already on it. CPU-Z said 1.xx and I got nervous so I looked on the actual motherboard near the PCI slots, and it said Rev2, so phew =P But I flashed it to 1009.007 and it seems to be working alright.

 

[Zebo]

Next, I set the multiplier to 10x, HTT to 3x, and the FSB to 275MHz.. Bad stuff.. I had to clear the CMOS and start over. Next I tried 260MHz with the same settings and got past POST, but Windows wouldn't load. So I went down to 250MHz finally being able to boot into Windows.

You did not isolate..
Sure you now know your boards max HTT, but so what? You don't start there when testing CPU it's just there for refernce so you know where to stop..

Nest step should have been, starting with 200 raise in about 5Mhz increments. .

- Set multiplier to it's maxium
-Set vcore to 1.6 -1.65 for 130nm chips, 1.50-1.55 for 90nm chips
-Set mem to SPD and 100, this will take memory out of equation
-Raise HTT up to boards max found earlier or until you start failing PRIME95 which will probably happen first..only give it 5 minutes max, that's good enough for now.
-At the point it starts failing within that 5 min time period, back down 5%, (CPU MHZ * 0.95), thats your max CPU speed, now test prime again at 95% for 12 hour for stable.
Prime 95 Link http://www.mersenne.org/freesoft.htm


I did just edit post cause I can see ambiguity caused by what I said...

 

[adeathangel]

Thanks for replying. So far what I've been able to accomplish is OCing the HTT to 230MHz with the 10x multiplier and having it somewhat stable. I haven't run Prime95 on it for 12 hours yet, but it ran 3 hours without errors. Seems my CPU doesn't like going past 2.3GHz which kinda sucks, but oh well. I also tried 9x multi with 260MHz HTT and it passed Prime95 for a good 3-4 hours too. But as soon as I ran 3DMark it froze, so blegh to that. I've been trying to keep the ram at around 200MHz until now. The computer doesn't crash or freeze at all if I bump down the ram to ~150MHz but that's slow and crappy. Could the value ram I bought that's rated for 234MHz at 3-5-5-10 really be so crappy as to be bad at ~200MHz? What's weird though, is when I run memtest86, there aren't any errors until I exceed 220MHz on the ram. But Windows really doesn't like it when I go past 200MHz, so I'm not sure. But yeah Zebo, I'm definitely going to take your advice and start over again taking it slow, and step by step. But it's going to be extremely time consuming testing 5MHz increments for such a long period of time. But that's the way it's done and I'll try it out.

I'd probably have an easier time OCing if I got better ram modules, right? I was thinking about selling these and buying two 512MB PC4000's, any suggestions?

 

[Thor86]

What ram ratios have you used? If you use a ram ratio, you should be able to overclock you cpu higher regardless of the speed of your ram.

 

[adeathangel]

I've used 5:3 and 3:2 mostly. Even if I go down to 1:1 where ram would never be an "issue" that really defeats the purpose of what I'm trying to do. I mean, what's the point of OCing my CPU is the ram runs slow?

 

[GuitarDaddy]

Originally posted by: adeathangel
I've used 5:3 and 3:2 mostly. Even if I go down to 1:1 where ram would never be an "issue" that really defeats the purpose of what I'm trying to do. I mean, what's the point of OCing my CPU is the ram runs slow?

At 1:1 is when you ram is most an issue.

Using a divider like 3:2 removes ram from being an issue.

And although ram speed affects performance, CPU speed is definately more important.

2500mhz with ram @ 185 will easily beat 2200mhz with ram @ 220 in every benchmark except Sandra memory bandwith. Bottom line is on A64 memory bandwidth(speed) is not nearly as important as it is for Intel, or even older AMD chips, because of the onboard memory controller.

Maximize you CPU and forget about memory speed!

 

[Duvie]

Does anybody how 4 sticks of ram works in this board?? I have heard of i tbeing a limiting factor in mem speeds and cas timings allowed. May have to juack the vdimm up as high as it goes since all of the slots are populated by ram...

 

[GuitarDaddy]

Yes they will work, but I beleive you will have to use a 2T command rate which is a big performance hit on A64's. I have read one post from a person claiming to be running 4 sticks at 1T, most posts I've read with people using 4 sticks have to use 2T.

I just yesterday tested the affect of 1T and 2T on my board.
I benchmark the following two settings and the #1 setting beats the second one in all the benchmarks by 2-5% even the Sandra Memory Bandwidth

1. 283x9=2557mhz DDR333 ram=233mhz 2-3-3-5 1T
2. 287x9=2583mhz DDR400 ram=287mhz 3-4-3-8 2T

 

[Duvie]

Originally posted by: GuitarDaddy
Yes they will work, but I beleive you will have to use a 2T command rate which is a big performance hit on A64's. I have read one post from a person claiming to be running 4 sticks at 1T, most posts I've read with people using 4 sticks have to use 2T.

I just yesterday tested the affect of 1T and 2T on my board.
I benchmark the following two settings and the #1 setting beats the second one in all the benchmarks by 2-5% even the Sandra Memory Bandwidth

1. 283x9=2557mhz DDR333 ram=233mhz 2-3-3-5 1T
2. 287x9=2583mhz DDR400 ram=287mhz 3-4-3-8 2T

That is what I was thinking....There need to be some degradation somewhere to get it to work properly with all 4 slots populated....

I thinkon the neo2 board it automatically reverts to 166 divider...someone confirm that!!!

 

[adeathangel]

Criminy, I have four sticks of ram, and not only that, crappy value ram. So with all my dimms being populated, that would probably explain why things are so unstable past 200MHz. If I have it at 3:2 and leave it well below 200MHz I can run it at 1T, and I guess since memory frequency isn't that important, I'll try OCing my CPU once again. Jeez, I wish I had gotten two 512MB chips of something decent to OC my memory a bit too. I guess this explains my weak results, I was trying to keep the memory around 200MHz, maybe even 204MHz or 208MHz... Well, now I'll leave it down and work on CPU. Thanks for all the posts btw guys! I'll reply with new results, although it may take some time finding something stable (=.

 

[adeathangel]

Originally posted by: GuitarDaddy

At 1:1 is when you ram is most an issue.

On my board there's a 1:1 and a 2:1. I think the 2:1 is 1:1 for everyone else, but my 1:1 runs at 100MHz if the HTT is 200MHz, so I don't think that's quite right for it "being an issue" (=.. or maybe it is?

 

[GuitarDaddy]

Wow that sounds backwards? 1:1 ratio should mean HTT=DDR speed. 2:1 should give 100mhz w/ HTT 200

 

[adeathangel]

I know, right? This A8V Deluxe board is wacky..

 

[loafbred]

>>On my board there's a 1:1 and a 2:1. I think the 2:1 is 1:1 for everyone else, but my 1:1 runs at 100MHz if the HTT is 200MHz, so I don't think that's quite right for it "being an issue" (=.. or maybe it is?<<

That's the way it works on my Asus K8V SE Deluxe (at least the 2:1 part). When I select 2:1, it shows "400 MHz" in parenthesis.

I'd say you've done very good with 4 dimms if you got it to run DDR 400 at 2T with 2.8v. Have you tried dropping the mem ratio low enough to keep the memory at or below DDR 333 (I'm not familiar with what's available on that board) and lowering latencies to something like 2.5-4-4 with 1T. You might be able to keep the dimm voltage down and get what you want.

 

[BigEm]

-Set mem to SPD and 100, this will take memory out of equation

Sorry but i don't really get what you mean by setting it to speed and 100. Does that mean to stock speed?

 

[BigEm]