Stumped! Opteron 170 is unstable!

ShogunCS

Limp Gawd
Joined
Jan 22, 2005
Messages
176
Hey guys,

I'm having some major issues getting my Opteron 170 to stabilize at any decent speed since Super PI and Prime95 both fail constantly within minutes. I just received my new Opty 170 and swapped it with my previous CPU which was a Opty 148. The 148 could easily run at 2.8ghz (255x11) at 1.45v. Therefore, I know my RAM isn't the culprit here since it could run even at 270 at 2.8v with no problems at 1:1 with my 148. The 170 I have has the CCBBE 0610DPMW stepping which is supposed to hit 2.8ghz and all I am trying to accomplish is 2.5ghz (250x10) to start. I'm not new to overclocking as I've had pretty good success with other CPUs but this is my first dual core CPU and I'm not sure if I'm going about it properly. I'll list off some of my BIOS settings I am currently set at:

FSB: 250mhz
LTD/FSB Ratio: 4x (1000mhz)
LTD Bus Transfer: 16 16
FSB Mult: 10x
PCIE Frequency: 100mhz
CnQ Support: Disabled
CPU Vcore: 1.45v
LTD: 1.2v
Chipset: 1.52v
DRAM: 2.8v
-----------------------------------------------------------------------------------------------------------------------------

DRAM Frequency Set (Mhz): 200MHz (1:1)
Command Per Clock (CPC): Enable
CAS Latency Control (Tcl): 3.0
RAS# to CAS# Delay (Trcd): 3
Min RAS# Active Time (Tras): 8
Row Precharge Time (Trp): 3
Row Cycle Time (Trc): 10
Row Refresh Cyc Time (Trfc): 16
Row to Row Delay (Trrd): 2
Write Recovery Time (Twr): 2
Write to Read Delay (Twtr): 2
Read to Write Delay (Trwt): 3
Refresh Period (Tref): Auto

DRAM Drive Strength: Auto
Max Async Latency: Auto
Read Preamble Time: Auto
Idle Cycle Limit: 64
Dynamic Counter: Enable
R/W Queue Bypass: 16x
Bypass Max: 07x
32 Byte Granularity: Disable (4 bursts)

As you can see, my settings are not very tight as I've had my DRAM settings even tighter with my 148. As for the BIOS itself, I'm using the latest version from DFI and I have a clean Windows install (doesn't really matter) with all the latest drivers for my components. My temps are all good across the board, all well under 40 degrees (Celcius). I'm not sure what else to post but any advice will be greatly appreciated.

Thanks!

P.S. Refer to my sig for hardware details.
 
trying increasing chipset and ltd voltages.

Idle Cycle Limit: 256
Dynamic Counter: Disable
Command Per Clock (CPC): Auto

And I would suggest updated to the current bios if your not. And if you are up to date, I would clear the CMOS.

Make sure you have good contact between the H.S. and CPU. Is your temps fine?
 
I've tried increasing and decreasing all voltages on the mobo to no avail. My BIOS is up to date and my temps are all good. I'm not sure what it could be.
 
I have finally figured it out, my RAM kept crapping out at 2.8v so I dropped it down to 2.7 and I can now run stable. The funny thing is, my 148 ran just fine with my RAM at 2.8 so this was kind of odd. Anyways, thanks for the suggestion but all is good now!
 
when you are at or above 250FSB set LTD/FSB Ratio to 3x.
 
well the ram issue is more of the memory controller that is ondie.. so changing the cpu will affect mem speeds... I have to run my 3-4-4-8 OCZ 2gb kit at 3-4-4-10 to have my CPU stable at 334*9 with a 5/6 divider to get DDR500 if i run it at 3-4-4-8 it will not prime for longer than 1 hr but run it at 10 and i get 23hrs orthos stable...
 
Yep, I had this exact same problem with my Opteron 170 and 2 gigs of Samsung UCCC. On my old 146 it would run stable at 3-4-4-8 @ 260MHz. Popped in the 170 and suddenly my system would lock up, crash, or power itself down within seconds or minutes of booting, even at 2.5GHz, just like you. In disgust, I ended up selling the CPU on the HardForum, with the warning it might not be stable at 2.5GHz. Guy that bought it was like "Thanks! Chip runs great at 2.95GHz!" So of course my reaction was "Nooooooo. Whyyyyyy?!?" Then I decided to go ahead and suck it up and buy another one from Newegg. I got the same stepping as the first one CCBBE 0610DPMW. Pop it in, boot up to 2.4GHz, good. 2.5GHz, system powers down. So now I'm thinking "no way it's the chip" so it HAS to be something else. Most likely culprit is the RAM, so I start experimenting with timings, even though it was stable before at these timings.

What I found to make a HUGE difference in stability with the 146 (this is on an Asus A8N-SLI Premium) was to increase Trc (Row Cycle Time) and Trfc (Row Refresh Cycle Time). From reading around the net, it seems that setting Trc to 2 less than Trfc is the most efficient way to do it. I think mine are currently set to 11 and 13, with my standard settings at 3-4-4-10. I was ready to bump them to 12 and 14 but I ended up getting it stable, so I'm happy. I think it's worth a shot in your case to try this if you're interested in tweaking it a bit.

You could try setting your Trc value to 14 and Tfrc to 16 and see if that helps your stability. If it does help, you can slowly lower them down to find the sweet spot.
 
Back
Top