a8v deluxe problems

[magx]

i've got a Asus A8v Deluxe Mobo and when i set my ram voltage to 2.8v my system locks up in windows no matter what i do. Any ideas? Also, i've got a amd 64 3500+ winchester @ 2.2 ghz right now...all i wanna know is HOW DO I GET THIS SUCKER AT 2.4 GHZ with my memory staying at 200 mhz (400 mhz) with the hypertransport bus at 1000 mhz? is there a way i can download a modded bios to allow me to adjust the hypertransport bus by 10 mhz each or something? Mine is set like:200 mhz 400 mhz 600 mhz 800 mhz then 1000 mhz...

I'm also confused about the memory dividers too...

here's the 3 main things in my system that i'm using to overclock:

OCZ 520W Powerstream
OCZ EL Platinum Rev 1 2-3-2-5 1t (2x 512 MB)
AMD 64 3500+ Winchester @ 2.2 GHZ with 0 stepping.

 

[risk_reversal]

If you raise the fsb to 240 and drop your multi to 10 (since lower multis are unlocked), your cpu's clock speed will be 2400Mhz. If you then change the cpu to memory dividers to 5:3 ie 166 your memory will then work at DDR400 speed ie 2400/12=200

see the thread below for relavant tables

http://www.ocforums.com/showthread.php?t=295328

In respect of the HTT you will have to drop the value to 800 in the bios. The actual calc is 200x5=1000 and in the case of 240, then 240x4=960.

http://www.ocforums.com/showthread.php?t=355689&highlight=a64+lower+multis+unlocked

Hope this helps

 

[magx]

k, that problem is solved...i managed to hit 2509 mhz stable so far at 1.525v 228 mhz htt (default multi) and memory divider was used 5:3 or something and is running at around 210 mhz /w 2.5 3-2-7 1t timings. I also got chipset voltage set to 2.6v and ram at 2.7v BUT if i switch it to 2.8v then my pc will randomly lock up as it has with OCZ's Rev 2 chips. I mean a total lock up where the mouse pointer freezes and that's it...i'm forced to reset my pc. My power supply rails are more than capable of handling everything i got in my system...i'm sure of it so what's causing the freezing? i've tryied other ram and nothing will run at 2.8v NO MATTER WHAT

 

[magx]

Stable Voltages on each rail

Toasty Temps (125 F max)

Prime95 torture test (large)

 

[magx]

2509 Mhz OC (don't ask me why the voltage is at 1.275 v)

 

[risk_reversal]

magx, there is something wrong. You raised your fsb/htt to 228 and left your multi unchanged at 11 to give a clock speed of 2509Mhz. You say you used the 5:3 memory divider but have a memory speed of 210.

From what I can see the numbers do not work. The 5:3 divider should give a memory speed of 179 ie 2509/14=179.

The memory dividers on a number of bioses from 1011 to 1014 (I think) were faulty. What bios are you running?

Have you read the thread below highligting this problem

http://www.ocforums.com/showthread.php?t=408517&highlight=a8v+owners+read

In my opinion running your board 228 htt/fsb, 11 multi, ram 210 is not really optimal. Better altogether would be 240 htt/fsb, 10 multi, ram will fall in at 200 with a working 5:3 divider.

There are many posts which confirm that the most (95%+) performance gains to a system come from raising the fsb and not from increasing the speed of the memory.

Also for stability purposes, you need to run Prime95 for 24hrs.

Good luck

 

[magx]

there is a problem....

when i set my mem ratio to 5:3 (ddr 333) in bios (0017.002) i raised my htt to 220 mhz and lowered multi to x10 and booted into windows. My cpu speed was 2.2 ghz and my memory speed was 200 mhz on the mark...

then i went with a diffrent option... 3:2 or something and it had nothing next to it...now i'm able to do 240 mhz htt with ram staying at 200 mhz BUT in cpu-z the cpu ratio changed from 11 to 12 so i dunno... any advice?

 

[KillrBuckeye]

Hey magx, I am the author of the thread that risk_reversal linked. I should point out that at the time I created that thread, the latest BIOS was 1014. I am going to update that thread to point out that to the best of my knowledge, this memory divider issue has not been resolved by ANY BIOS version to date. The observations you described in your latest post demonstrate that the problem still exists in BIOS 1017.

The [5:3 DDR333] setting is extremely screwy. DO NOT USE THIS SETTING. This setting was extremely unstable for me, even when memory was at stock speeds, and CPU-Z couldn't even read the memory frequency when I had this setting enabled. Further, other people have had stability problems and/or incorrect memory frequencies show up when using this setting. In your case, running 220x10 with the [5:3 DDR333] setting *should* yield a memory speed of 2200/CEILING(10/(5/6)) = 183.3 MHz! Instead, you're getting 200 MHz--major problem there. When you used 240x10 with [3:2] divider, your memory speed *should* be 2400/CEILING(10/(3/4)) = 171.4 MHz. Instead, you're getting 200 MHz, which is indicative of a true [5:3 DDR333] divider: 2400/CEILING(10/(5/6)) = 200 MHz.

So, what we've learned from this is that everything I stated in my original thread (linked by risk_reversal) still holds true with the latest BIOS version:

- Do NOT use [5:3 DDR333] divider setting in the BIOS because it does something screwy
- If you want a 5:6 divider, use the [3:2] setting in BIOS.
- If you want a [3:2] divider you're out of luck
- Asus BIOS support is horrible and they never responded to my inquiry into this issue

 

[risk_reversal]

magx, if you want working dividers and a stable bios for a 3500 Winchester core, then the 1010.003 bios is a good choice although others users speak highly of the 1009.

KillrBuckeye, it was following your original thread that I opted for bios 1010.003 for my Winnie with correctly working dividers and 1T command that also worked. For complete clarity, with these screwy bioses, to get [5:3 DDR333] you need to select 3:2 in the bios is that correct.?

Reason I ask is because I am about to upgrade my cpu to a SanDiego 4000 and I want to run it at 250x12 ie 3000Mhz (if posssible) and use the 5:3 dividers to get my ram running in spec at DDR400.

All the best

 

[KillrBuckeye]

KillrBuckeye, it was following your original thread that I opted for bios 1010.003 for my Winnie with correctly working dividers and 1T command that also worked. For complete clarity, with these screwy bioses, to get [5:3 DDR333] you need to select 3:2 in the bios is that correct.?

Reason I ask is because I am about to upgrade my cpu to a SanDiego 4000 and I want to run it at 250x12 ie 3000Mhz (if posssible) and use the 5:3 dividers to get my ram running in spec at DDR400.

Yes, for every BIOS since 1011, to get [5:3 DDR333], i.e. a 5:6 divider, you need to select [3:2] in the BIOS.

FYI, 1009 worked fine even with my Venice core. There were only a few annoying issues, like my TT Big Typhoon fan speed was below the warning threshold and it gave me a warning message on every boot. Small things like this were corrected in subsequent BIOS versions. Further, to get support for some of the later CPUs, a newer BIOS version may be required. (Technically, my Venice wasn't supported by 1009, but it seemed to work). I'm not sure what a San Diego requires. You might want to try to research that a bit before you install it. It would suck if you got a no-POST and had to swap it out for your old CPU just to perform a flash.

 

[risk_reversal]

Many thanks for your reply KillrBuckeye. According to the Asus website, the SanDiego E4 requires the 1011 min. I was actually trying to find out from other users in respect of requirements by posting but I think that this whole A8V bios fiasco has upset many users who have since switched mobos.

I have to confess that I am about to put Asus in the Abit camp, it being that I do not buy Abit period.

A user has responded to one of my posts confirming that he ran the 1015.002 bios with an o/ced SD with no issues as regards working dividers and 1T command.

Still many thanks for the input. Hopefully, this has given magx food for thought.

BTW magx, how are you faring with your o/c?

Cheers

 

[magx]

i just upped my voltage 2 notches in the bios (i think 1.45; half way to 1.5v) with htt at 240 mhz and ram divider at 3:2 and it does seem like the memory divider is screwy...my ram is running at 200 mhz (on spec) at this setting. i've been running prime95 for around 3 hours now...no errors yet. when i raised my htt by 2 mhz i failed prime95 after a while so i backed down. this cpu seems to need high voltages. this still doesn't solve my 2.8 vdimm voltage problem. Any voltage increase over 2.7v on any type of ram results in a system freeze...

Temps are sticking around 49 c (120 f)

 

[magx]

oh, i'm buying the 'DFI LANPARTY nF4 SLI-DR' and a 4400+ Toledo DC CPU with 2048 MB of OCZ ram (same timings as this ram but 1024 MB sticks) and other junk, so i'm hopin to get 2.6 ghz min out of it. **** this mobo

 

[risk_reversal]

Looks like you were almost their with this board and your cpu o/c.

In case you reconsider and decide to keep the A8V, my suggestions would be as follows.

1. What are the steppings to your 3500 Winnie. Some steppings o/c better than others. There are lots of threads in this respect. My Winnie 3200 has the reputed CBBHD steppings (best) and is o/ced modestly to 2400 from 2000 stock with vcore at 1.48. I have run it at 2600 with prime95 stable. If your o/c fails Prime over a 24hr period then you need to crank up the voltage. Also it is always a good idea to burn in your components ie cpu and ram at stock speed as this will improve your chances of getting to a higher overclock.


2. Using the 3:2 divider you got your cpu to run at clock speed 2400 (240x10) and the ram fell in at DDR400.

3. In respect of your ram having probs, I would have relaxed the timings a bit from 2,3,2,5 to 2.5,3,3,6. The performance loss would have been nominal. Obviously you have good quality ram and you could have taken your 3500 to 2500Mhz and run the ram at 208.

4. Your temps were also good.

However changing your rig completely to a DFI and X2 4400 Toledo would be a major step up especially as the X2 4400 has 1Mb caches for each processor.

Good Luck

 

[magx]

stepping 0..i upped my htt by 10 mhz from 240 to hit 2.5 ghz and passed mem test with 2.5-3-2-6 timings but even upping my voltage to 1.525 it still failed prime95 and temps were around 54c (ouch) am i doin somethin wrong? this thing is voltage hungry...

 

[KillrBuckeye]

stepping 0..i upped my htt by 10 mhz from 240 to hit 2.5 ghz and passed mem test with 2.5-3-2-6 timings but even upping my voltage to 1.525 it still failed prime95 and temps were around 54c (ouch) am i doin somethin wrong? this thing is voltage hungry...

Every CPU is different. Some will require 1.55V to hit 2.5 GHz stable. Be careful though, because passing Memtest86+ doesn't rule out a memory instability issue. This is why it's best to find the capabilities of your CPU and memory independently. Right now you have too many variables as you're overclocking both your CPU and RAM, and you have no way of knowing which is causing the instability.

To find the max speed of your CPU, put the memory on a more aggressive divider (try the next one down--I think it's [4:3 DDR266] ?), and raise the HTT until it becomes unstable. To find the max speed of your memory, lower the CPU multi to something like 5x, and raise the HTT until Memtest starts giving errors. To be safe, you should clock your memory well below the speed at which it gives Memtest errors. I'd recommend backing up anything important before trying this, because you can quickly corrupt your OS if you boot to Windows with very unstable memory.

 

[QuietIce]

Hey magx, I am the author of the thread that risk_reversal linked. I should point out that at the time I created that thread, the latest BIOS was 1014. I am going to update that thread to point out that to the best of my knowledge, this memory divider issue has not been resolved by ANY BIOS version to date. The observations you described in your latest post demonstrate that the problem still exists in BIOS 1017.

The [5:3 DDR333] setting is extremely screwy. DO NOT USE THIS SETTING. This setting was extremely unstable for me, even when memory was at stock speeds, and CPU-Z couldn't even read the memory frequency when I had this setting enabled. Further, other people have had stability problems and/or incorrect memory frequencies show up when using this setting. In your case, running 220x10 with the [5:3 DDR333] setting *should* yield a memory speed of 2200/CEILING(10/(5/6)) = 183.3 MHz! Instead, you're getting 200 MHz--major problem there. When you used 240x10 with [3:2] divider, your memory speed *should* be 2400/CEILING(10/(3/4)) = 171.4 MHz. Instead, you're getting 200 MHz, which is indicative of a true [5:3 DDR333] divider: 2400/CEILING(10/(5/6)) = 200 MHz.

So, what we've learned from this is that everything I stated in my original thread (linked by risk_reversal) still holds true with the latest BIOS version:

- Do NOT use [5:3 DDR333] divider setting in the BIOS because it does something screwy
- If you want a 5:6 divider, use the [3:2] setting in BIOS.
- If you want a [3:2] divider you're out of luck
- Asus BIOS support is horrible and they never responded to my inquiry into this issue

IDK about the other issues with your BIOS but the memory numbers I've seen in this thread lead me to believe that the BIOS has been updated to use the 183 RAMclock but now the settings are all off by 1 notch so your DDR333 is really DDR366, DDR300 = DDR333, etc. At 220x10 w/DDR366 the RAM should be running at 2200/CEILING (2000/183.3333) = 2200/11 = 200.0.

Just a thought ...

 

[KillrBuckeye]

IDK about the other issues with your BIOS but the memory numbers I've seen in this thread lead me to believe that the BIOS has been updated to use the 183 RAMclock but now the settings are all off by 1 notch so your DDR333 is really DDR366, DDR300 = DDR333, etc. At 220x10 w/DDR366 the RAM should be running at 2200/CEILING (2000/183.3333) = 2200/11 = 200.0.

Just a thought ...

It has been 8 months or so since I've used my A8V Deluxe, so I can't remember everything I observed. I'm pretty sure, however, that the [4:3 DDR266] setting did actually apply the 2/3 divider, and the [2:1 DDR200] setting applied the 1/2 divider, so it's not just an issue of everything being shoved down a notch. Further, I recall extreme instability when using the [5:3 DDR333] setting, even with CPU at stock speeds. As I mentioned before, CPU-Z couldn't even report the memory frequency when this setting was enabled. However, given that other people seem to be getting speeds indicative of a DDR300 divider when using this setting, maybe it depends on the CPU core. The OP has a Winchester, and I've seen someone else with a Winchester report this behavior.

I don't know what it is with Asus and screwed up memory dividers... On my current board (A8N-SLI Deluxe), I absolutely CANNOT run any memory divider (other than 1:1) on my Venice core without extreme instability, i.e. it's unstable even with memory underclocked by 40 MHz. If you read the long sticky about the A8N-SLI, you will see that this behavior has been observed by others, and it only seems to affect Venice cores. Still, it makes me very nervous about getting a CPU with only a 10x multi, because my OC may be limited if I have to run 1:1.

 

[QuietIce]

It could be that CPU-Z wasn't programmed for the 183 setting at that time - isn't that relatively new? But not owning one of those boards I really couldn't say. I just noticed that the DDR333 settings in this thread result in a DDR366 Memclock. Who knows what ASUS is up to when they release a new BIOS version - they do seem to have their share of problems. The A8N32 OCs very well and the settings are flexible but it has some issues with legacy USB and a major problem with the secondary RAID controller. I guess ASUS is too used to selling OEM to get it right for us peons ...