Can't Bust 285 FSB...

[Battle_Rattle]

Hi all...

 

I'm sitting at an overall Opteron 165 OC of 2430 and I'm decently pleased with it but 2.6-2.7 would be nice. I can't seem to go past 279-285 FSB (I found this limit by isolating the FSB from the CPU speed and Ram.) I'm using a Ultra-D with the 623-3 Bios (all specs below.)

 

Is the fact that I have a dumpy Seagate IDE Hard Drive hooked up holding me back? Should I take "Auto" off of my Ram and LDT settings? Should I go above 1.41v Core volts?

 

Check my stats and realize I have a Si-120 w/ Panaflo that is keeping my CPU load temps at 35c (im actually suspicious of that, so add 10degrees to that value) It was my first computer build but I studied well on how to apply the Arctic Silver 5 (I put a tad more than the size of a grain of rice)

 

In the interest of completeness here are my Genie settings ...

 

 

FSB Bus Frequency --- 270

LDT/FSB Frequency Ratio --- Auto

LDT Bus Transfer Width --- 16 16

CPU/FSB Frequency Ratio --- x9.0

PCI eXpress Frequency --- 100Mhz

K8 Cool n Quiet Support --- Disable

CPU Vid Startup Value --- 1.4v

 

---CPU Core Voltage --- 1.41v

---LDT Bus Voltage --- 1.22v

---Chip set Voltage --- 1.52v

---Dram 2.5v Voltage --- 2.88v

 

CPU VID Control --- 1.4v

CPU Special Control --- AUTO

LDT Voltage Control --- 1.2v

Chipset Voltage Control --- 1.5v

Dram Voltage Control --- 2.8v

Dram +.03v…. --- Disable

 

Mac Lan --- Auto

Mac Media Interface --- Pinstrap

Machine Mac(NV) Address --- Disable

Mac Lan Boot ROM --- Disable

Internal Phy Sata ½ --- Enabled

Internal Phy Sata ¾ --- Enabled

Via 1394 Control --- Enabled

Marvell Giga LAN Control --- Enabled

 

RAM Settings Here

 

DRAM Frequency Set --- 180 (DRAM/FSB: 9/10)

CPC --- Enabled

Cas Latency Control (Tcl) --- Auto

Ras to Cas Delay (Trcd) --- Auto

Min Ras # Active Time (Tras) --- Auto

Row Precharge Time (Trp) --- Auto

Row Cycle Time (Trc) --- 07 Bus Clocks

Row Refresh Cyc Time (Trfc) --- Auto

Row to Row Delay (Trrd) --- 02 Bus Clocks

Write Recovery Time (Twr) --- 02 Bus Clocks

Write to Read Delay (Twtr) --- 02 Bus Clocks

Read to Write Delay (Trwt) --- 03 Bus Clocks

Refresh Period (Tret) --- 3120 Cycles

Write CAS Latency (Twcl) --- Auto

Dram Bank Interleave --- Enabled

 

DQS Control --- Auto

DQS Value --- 0

DRAM Drive Strength --- Auto

DRAM Data Strength --- Auto

Max Async Latency --- Auto

DRAM Response Time --- Normal

Read Preamble Time --- Auto

Idle Cycle Limit --- 256 Cycles

Dynamic Counter --- Disable

R/W Queue Bypass --- 16x

Bypass Max --- 07x

32 Byte Granularity --- Disable (4 bursts)

 

 

THANK YOU FOR YOUR TIME! I DO APPRECIATE IT!

[Neuroff]

You probably have to raise VCore

[operaman]

Run the 704-2BT and put the RAM in the orange slots. Bump up the chipset voltage to a nice 1.7. I am getting 2520 running a 9/10 RAM divider at stock volts with the stock HSF from AMD primed.

 

Also your RAM timings are way too tight for what you are realistically going to get. Back it off to 3-3-3-8 or even 3-4-4-8. The latency does not really help or hurt at this point. The A64 wants more bandwidth and latency is a non-issue. Try those things. Then run dual super pi for 30 minutes (32M) and then start dual priming. You should be good to go.

 

When necessary then bump up the cpu voltage in small chunks if you need it.

[Guest laydback]

We need to see a A64 Tweaker screenie. Since you left lots of settings to AUTO. We need to actually know what setting is actually in use. Defintaly lower your LDT multi from AUTO to 4x or even 3x. 35c under full load isn't bad, bump up the vcore. Also search around the forum for other users with a similar setup and see what that has to offer. Take care.

[Battle_Rattle]

Ok sweet.... Im at 2600Mhz and PC4200 now....

 

I changed 4 things...

 

1) I upped Vcore to 1.45v

2) I changed LDT multi from "auto" to x3

3) I took Dram timings off auto and put in my stock Mushkin timings of 3-4-3-8

4) I put Dram CPC to 2t

 

Instant mobo FSB gain from 285 to 300 (I stopped at 300 because I was so damn surprised and I wanted to test! lol)

 

I have to think that it was the LDT multi on "auto" that was giving me problems... I'll isolate that later. I think i got it in my head that you can keep it on auto from the DFI OC guide.

 

Temps never surpassed 37c ...again I'm suspicious of that.

 

One last thing though.... what test would you guys do to ASSURE stability for a dual core. Dual small-fft prime for 12-18 hours or so?

[operaman]

The 2t can be changed. You were running at well over 1000 HTT. Remember this: Rule of thumb on HTT is never go over 1000. SO for example on Auto you were at 5 X 285... WOW... That is high. I may be wrong but I always to ake the multi there( 1,2, 2.5,3,4,5,) and multiply that by my FSB. Generally speaking I just lower it to 3X as a force of habit to get things going. Try that and use 1T and I betr you will likey... Good luckwith your oc man.

[operaman]

One last thing though.... what test would you guys do to ASSURE stability for a dual core. Dual small-fft prime for 12-18 hours or so?

 

Start with dual Super pi at 32M. Then once you get a stable run of that run dual instances of prime. 12-18 hours should be good. I run for 24 hours but I am a stability freak so there ya go....

[Battle_Rattle]

Thanks guys...ill do long testing soon...

 

UPdate >>> Running a 180 (9/10) divider and 1t does indeed seems to work (no memtest failures through 1 full test and 5 passes of test 5. So memory is short steady @ 261 DDR522