AMD Athlon X2 3800+ OC Problem

[Guest Ryn]

I have a X2 3800+ Toledo. I haven't maxed it out yet coz I'm contented at 2.4GHz. The highest I've tried is at [email protected].

[Thracian]

By the way, is there anyone on this forum who also owns an X2 3800+, I'd like to know their OCs...

 

Another by the way... What is an good stepping for the X2 3800+ , mine is:

 

Family: F

Ext. Family: F

Model: B

Ext. Model: B

Stepping: 1

Revision: BH-E4

(according to CPU-Z)

 

I've seen someone with exact the same (Ext.) Family, (Ext.) Model, Stepping and Revision hit 2.9GHz...

 

Priming 2.76GHz now, I just changed some timings... t.b.h I'm really getting sick of that... :drool:

 

 

First of all...You need the ACTUAL stepping on the cpu itself. The info you posted doesn't help anyone to determine how your cpu will clock.

 

Second, if your only running one instance of prime95, you are using only one core and your temps will not be maxed.

 

Cheers,

Thrac

[Rik.Reinink]

I have a X2 3800+ Toledo. I haven't maxed it out yet coz I'm contented at 2.4GHz. The highest I've tried is at [email protected].

 

X2 3800+ Toledo? I thought they were only available with Machester cores... :confused:

 

I tried to rand 2 Prime95's and one continues without any problems, but the other one stops @ 4 ~ 14 minutes... Just when I thought things couldn't get any worse... :sad:

 

Is there a way I can get my 2.74GHz stable with both cores?

 

Thanks all, and take care!

Rik

[tasr]

Rik,

 

Sorry after my last post this morning (USA time) I was off to rugby. I do have a life after the Street. :nod:

 

Yes you can kill your CPU with over volts even if your temps are low. You need to run prime95 on both cores to call your x2 stable. Check this thread out Setting up an X2 CPU for dual Prime95 stability testing for the setup. Also I would not rush for a new BIOS yet. Your doing pretty good just be a more patient. Get your prime on both cores done and stable. I also like 12hr prime then go for a 24hr to call it complete.

 

Check out the Overclocking Database. That would be a start for other x2 3800 oc settings.

 

I like to use a copy of 939 Overclocking Database Template to keep track of my oc’ing. Hope this helps.

 

Let us know how you’re making out.

 

tasr

[Rik.Reinink]

Hi tasr, and all others!

 

I even had problems running Prime @ 2.55GHz @ 1.45Vcore, trying to get 2.6GHz out of it at least, but is there a way to know for certain whats failing, for example:

 

I get no errors at MemTest -- So my memory must be Ok

I get errors at Prime95 -- So it must be my CPU failing, right?

 

I had this problem, but I found out it actually was my memory!! So, is there some tool, to find out whats not funtioning properly, for example my CPU, or my Memory (timings) or my memorycontroller... Such a tool would be great!

 

Thanks all, and Take Care!

Rik

[tasr]

I even had problems running Prime @ 2.55GHz @ 1.45Vcore

If this is where you are at a little more volts may help get stable. Without see all your settings I can't do much.

 

Get one of these 939 Overclocking Database Template to keep track of your oc. Then you can also post the text. Not much point going to 2.6 in you not stable.

 

tasr

[Rik.Reinink]

Hi tasr, and all other peeps who read this thread, these are my exact specs for the moment, the HTT/FSB is set to 269 becouse I am testing that currently, but 268MHz was 100% stable @ Dual Prime95:

----------

DFI LANPARTY nF4 Ultra-D - //Newest BIOS//

AMD Athlon X2 64 3800+ Manchester

2x512MB GEIL Ultra PC4000 (DDR500)

ATi Radeon X800GTO2 @ X850Pro (16 PP), 480/1000 Radeon Omega Drivers 3.8.231 (Catalyst 6.3)

Western Digital 250GB SATA HDD(8MB cache)

Sony DVD+RW (Dual Layer) Drive

Ernemax Noisetaker 485 Watt Power Supply

*

*

Genie BIOS Settings:

FSB Bus Frequency - 270

LDT/FSB Frequency Ratio - X 4.0

CPU/FSB Frequency Ratio - 16 - 16

PCI eXpress Frequency - 100MHz

 

CPU VID StartUp Value - 1.550v

 

CPU VID Control - 1.550v

CPU VID Special Control - Auto

LDT Voltage Control - 1.20v

Chip Set Voltage Control - 1.50v

DRAM Voltage Control - 2.80v

 

DRAM Configuration Settings:

DRAM Frequency Set - 180=RAM/FSB:09/10

Command Per Clock (CPC) - Enable

CAS Latency Control (Tcl) - 3.0

RAS# to CAS# delay (Trcd) - 04 Bus Clocks

Min RAS# active time (Tras) - 08 Bus Clocks

Row precharge time (Trp) - 04 Bus Clocks

Row Cycle time (Trc) - 12 Bus Clocks

Row refresh cyc time (Trfc) - 15 Bus Clocks

Row to Row delay (Trrd) - 00 Bus Clocks

Write recovery time (Twr) - 03 Bus Clocks

Write to Read delay (Twtr) - 02 Bus Clocks

Read to Write delay (Trwt) - 02 Bus Clocks

Refresh Period (Tref) - 0128 Cycles

Write CAS Latency (Twcl) - Auto

DRAM Bank Interleave - Enabled

 

DQS Skew Control - Increase Skew

DQS Skew Value - 255

DRAM Drive Strength - Auto

DRAM Data Drive Strength - Auto

Max Async Latency - 7.0 Nano Seconds

DRAM Response Time - Fast

Read Preamble Time - 5.0 Nano Seconds

IdleCycle Limit - 008 Cycles

Dynamic Counter - Enable

R/W Queue Bypass - 16 x

Bypass Max - 07 x

32 Byte Granularity - Disable(4 Bursts)

 

Short Description:

269x10 @ 2690MHz, 1.550V

3-4-8-4 @ 2.8V DRAM Voltage

----------

I am pretty happy here, just got 2.68GHz stable, (setting it to 1.55V did the trick, and my temps are great, 35*C under full load, thanks for that one tasr ) trying 2.69GHz, but I don't expect much problems.

 

I also tried 2.72GHz just after 2.68GHz, but the second core failed after 45mins. but one core just went on and on, any suggestions?

 

Oh, and by the way, does "overclocking" the PCI eXpress Frequency improve my FPS, becouse I can put it all the way up to 145MHz without any problems (I think...)

 

Thanks all, and take care!

 

Rik

 

P.S.

Sorry for the long post, I hope I didn't break any rules or something??

[RLM]

Allegedly the PCI express makes no difference at all, though some will beg to differ.

 

Would be interesting to know exactly what your stepping is? It would certainly give us a rough idea as to whether your OC seems about right or not.

 

As someone else said here the stepping is written on top of the chip and would read something like LDBHE 0602XPCW as an example.

 

Just read on another forum how a guy with an x2 3800 managed to jump from 2.66 to 2.9 by loosening up his async latency and read preamble settings. Don't know if that was prime stable though.

 

Few things. Personally I wouldn't bother increasing the chipset or LDT volts unless it really is making a difference. Just creating unecessary temp increase if not.

 

Secondly don't know what your memory recommended voltage is, but please bare in mind that the Ultra-D settings can be a bit off. Insome cases they can be up to .2v higher than what you actually set i.e. 3.0v instead of 2.8v

 

If you read above where you set the Dram voltage you should get a close to actual reading. On my ultra-d I had to set it to 2.7v to get 2.8v. Does this make sense?

 

Anyways there you go.

 

RPG

[Rik.Reinink]

Hi RLM!

 

I just lowered my Chipset and LDT Voltage, and indeed, it didn't make any difference. I can get it stable @ 2.7GHz on 1.55V, but on 100% load, both cores, my temps reach 48*C max. (45*C Most of the times) and idle only 27*C.

 

Is 48*C (too) high? Thanks for advice people!

 

Take Care!

Rik

[Rik.Reinink]

I was playing around with my OC becouse I had those timings RLM recommended me lowered a bit, so I put my PC on 1.5V and 2.77GHz (277*10x)

 

The following happened:

- Windows booted, BSOD -> reboot

- BIOS, memory voltage to 2.6V (becouse 277/1.1=250MHz so it might be a bit hard on such low voltage I guessed)

- Windows boot (bit slow :S) -> logged in -> started PC Wizard 2006 -> hang -> reset

- BIOS, memory voltage to 2.7V (didn't want to give up, since GEIL guarantees that 3V is 100% safe)

- Windows boot -> login, bla bla -> MemTest -> error, error, error -> Pi: SQR NOT CONVERGENT

- Reboot, BIOS, memory voltage 2.8V

- Windows boot, bla bla -> MemTest: 100% Done -> PI: SQR NOT CONVERGENT

 

And THAT I didn't like... MemTest was ok, but PI failed on me, but again, I didn't give up so:

 

- Reboot, Bios, memory voltage to 2.9V (as RLM said, it actually is (as my Ultra-D says) 2.96V, which must be ok...)

- Windows bla bla -> didn't even bother memtest -> PI, 1M: Without problems

 

Then I was glad for sure, but just don't understand why MemTest (a tool for memory testing :|) could pass, but PI, a CPU test program wanted more Voltage to the memory.

 

If anyone could explain... Please be my guest!!

 

Thanks in advance all, and Take Care!

Rik

 

P.S.

One more thing, I always wondered:

 

280*10 = 2800MHz

350*8 = 2800MHz

 

Is there a better one of the two, or is it just that I can set my memory speed more optimal?

 

Another P.S.

It didn't appear to be stable anyways, but I'll keep y'all posted!

[dephconn]

I'm 100% stable at 1.39v and 240 fsb (2.4 ghz). I'm not sure what stock is - I have it on auto.

 

a 104% voltage got me stable at 2.5 ghz, and 110% almost stable at 2.6 (I think I needed a memory divider at that point), but for my processor the "sweet spot" seems to be 2.4 ghz. I just got unlucky, I guess.

 

It's really luck of the draw with OC'ing, especially with these dual cores, so you may have found your limit.

[Rik.Reinink]

Hi dephconn!

 

I think I indeed hit my max, I think I'll lower the voltage a bit, and put it back on 2.55 ~ 2.6 GHz, so I could try my mem without a devider, this is optimal, right?

 

Any more suggestions still are welcome for sure!

 

Thanks all, and take care!

Rik