Dual Core Opteron Thread #2

[QuietIce]

I recently picked up a 165 Opty LCBQE 0712 and I'm loving it. I have it up to 2.9 so far but I think I may be limited by my FSB. The computer won't boot if I raise the FSB past 325 no matter what the cpu voltage is set at. Any ideas for boosting it even further? Do I have to bump up the chipset voltage more?

Will any memory timings affect the FSB? I have BH-5 so the timings are super tight. The memory doesn't seem to be the limitation because I've had the stuff running stable at 260mhz before but with dividers it's currently only at 225mhz.

Drop the CPU multi, Ram Divider, and HT Link multi down then turn the FSB up to see how high it'll go.

On the other hand my LCBQE 0711SPAW hit top end about there, too @ 1.4 vCore ...

[RVWinkle]

Drop the CPU multi, Ram Divider, and HT Link multi down then turn the FSB up to see how high it'll go.

On the other hand my LCBQE 0711SPAW hit top end about there, too @ 1.4 vCore ...

Good idea, I have the HT and Ram Divider pretty low already, I'll try dropping the cpu multiplier and test pure FSB overclocking. I'm at 2925 @ 1.4 so I'd really like to round it out with another 75 mhz.

[Cheffy]

if its much like my cpu, 2920@1.4v, 2966@1.425, 3000@1.55 - the scaling goes to pot at just over 2966 making the last 33mhz not worth chasing....

the ram should be ok, tho my mem controler gets very funny at the same point at core volts jump - 3000 is much happier with ram on 180 div than 200 div.

try whats been recomeneded already to check the htt limit - it might jsut be the limit of your board/the cpu (very high cold-bug?)

[Mord]

Hey, guys!

Yesterday I received my opty 165 from newegg. It's a LCBQE 0722RPAW!

First checkpoint - 3GHz@1.35V. Not absolute stability testing, but here's an hour of orthos.

Strange thing is it boots and benches at 3200@1.4-1.5V, but fails orthos in a second.

A few questions. No one helped me in the Liquid cooling section. Obviously I'm popping the ihs off. Is it ok with ApogeeGT block and lanparty mobo? Hope so. I mean if there would be a good contact between the block and the die. Also I don't have any springs in the bundle! Is that a problem?

[RVWinkle]

A few questions. No one helped me in the Liquid cooling section. Obviously I'm popping the ihs off. Is it ok with ApogeeGT block and lanparty mobo? Hope so. I mean if there would be a good contact between the block and the die. Also I don't have any springs in the bundle! Is that a problem?

If the attachment is bolt-on like most water blocks then you are fine and will just need to take off the bracket surrounding the cpu socket.

[RVWinkle]

if its much like my cpu, 2920@1.4v, 2966@1.425, 3000@1.55 - the scaling goes to pot at just over 2966 making the last 33mhz not worth chasing....

the ram should be ok, tho my mem controler gets very funny at the same point at core volts jump - 3000 is much happier with ram on 180 div than 200 div.

try whats been recomeneded already to check the htt limit - it might jsut be the limit of your board/the cpu (very high cold-bug?)

Further testing on mine yielded some interesting results. I found that increasing the memory divider allowed me to raise the FSB. With this I was able to raise the FSB all the way to 350 before I ran out of memory dividers. This yielded 3.15ghz with the unfortunate side effect of my mem running at 175.

What's happening is that the memory controller seems to get weaker the higher it's clocked. The higher the fsb, the lower I needed my memory to be clocked at to boot. What I want to find is a setting that allows me to bring my memory back up to the 220-240 range while allowing for a high fsb, anybody have any ideas?

Right now I'm priming at 3ghz @ 1.4 and I'm thinking I'll top out somewhere around 3.1-3.2 if I can find what's causing this instability.

[PhilDoc]

Wow, nice chip. What bios are you running. On mine I found that different bioses worked better with certain dividers and my memory.

[RVWinkle]

I'm at work now but I think it's the 704-2bta bios. I'll try the 702-2 this weekend. Aren't those the only two that are compatible with this chip?

[PhilDoc]

I'm at work now but I think it's the 704-2bta bios. I'll try the 702-2 this weekend. Aren't those the only two that are compatible with this chip?

NO, there are at least two from DFI, I think the 406 and 329 on their web site.

[RVWinkle]

NO, there are at least two from DFI, I think the 406 and 329 on their web site.

Stupid me I didn't realize that FireDragon's bios guide wasn't up to date and that lower numbers would be more recent. I'll be sure to try out the updated versions this weekend.

[QuietIce]

If the attachment is bolt-on like most water blocks then you are fine and will just need to take off the bracket surrounding the cpu socket.

Not trying to contradict but are you sure? Seems like some of the guys over at OCF had to shave the socket a little for a good fit. IIRC the Apogee very slightly rubs against the high side of the CPU socket, where the hinge for the lever is.

I know for a fact the Storm doesn't rub there but it's very close ...

[RVWinkle]

Not trying to contradict but are you sure? Seems like some of the guys over at OCF had to shave the socket a little for a good fit. IIRC the Apogee very slightly rubs against the high side of the CPU socket, where the hinge for the lever is.

I know for a fact the Storm doesn't rub there but it's very close ...

I should qualify my post by saying I've never tried it. The processor package sits on top of the socket as opposed to Socket A where part of the socket was raised. I don't see how there could be any touching.

[QuietIce]

I should qualify my post by saying I've never tried it. The processor package sits on top of the socket as opposed to Socket A where part of the socket was raised. I don't see how there could be any touching.

With an IHS there's no issue at all because the CPU's tall enough that everything clears the socket. Take the IHS off, though, and you've lowered the top of the CPU by ~1/8" - just barely enough to get into that hinge.

The Storms clear the backside hump where the hinge is because they have beveled edges on the bottom. The Apogee has a flat bottom and I'm pretty sure it's just big enough to get into the hump. I'm not saying you can't mount an Apogee but it's not a "clean" mount and temps will suffer slightly if you don't shave just a little off the CPU side of the hump. That's why I never use an Apogee on a de-lidded chip - beside the fact that Storms perform much better.


BTW - An Apogee works on Socket A. My stock 2100+ runs ~28°C at load with a single BIP on the rear fan. I expected to OC it but the POS board doesn't have a way to adjust the PCI bus speed ...

[autobot]

Hey there Folks. I have an uncapped Opty 165 Stepping CCB1E 0609FPMW and I'm having a bit of trouble getting a decent overclock.

My system is as follows:
AMD Opty 165 (CCB1E 0609FPMW)
ASRock Dual939-SATA2
OCZ Platinum PC3200
PNY 7900GT (volt modded 637/718)
SB Audigy2ZS
WD SATA 80GB
Maxtor IDE 120GB
Lite-On DVD CD/RW combo IDE
Antec NeoHE 550
Swiftech Storm (rev1)
Swiftech MCP-655
DangerDen Maze4GPU
Swiftech MicroRes
Swiftech MCR-320
Yate Loon D12SM012 (model 705)
Masterkleer 7/16"ID - 5/8"OD tubing
FluidXP+ Blood Red

Here is a pic of my settings in CPUz


I have issues with getting my CPU oc past 2.5Ghz while remaining Orthos stable. I set the CPU voltage to its max of 1.4v in the BIOS. According to the first post in this thread, this CPU should be able to reach 3Ghz. I would greatly appreciate any help in accomplishing at least 2.8Ghz or 2.9Ghz. Thank you!

[QuietIce]

Hey there Folks. I have an uncapped Opty 165 Stepping CCB1E 0609FPMW and I'm having a bit of trouble getting a decent overclock.

I have issues with getting my CPU oc past 2.5Ghz while remaining Orthos stable. I set the CPU voltage to its max of 1.4v in the BIOS. According to the first post in this thread, this CPU should be able to reach 3Ghz. I would greatly appreciate any help in accomplishing at least 2.8Ghz or 2.9Ghz. Thank you!

First off, have you checked to make sure your board is stable above 285 MHz? If so, what is your "FSB"/"HTT" top end?

Second, how high has your RAM checked out and have you tried lowering it even farther?

Third, how fast are you running your HT Link?

[PhilDoc]

Are you sure you're making good contact with the core?

[autobot]

The latest BIOS (2.30) has no HTT cap. I chose my ram timmings based off the guy I bought the processor off of. He is an avid OCer and he played around with the settings for about an hour until everything was smooth. When he left, it was running smoothly at 2.4Ghz. I have since been very slowly upping the HTT until I hit the wall of 2.5Ghz with those ram timmings. I really don't know much about adjusting ram timmings to fiddle with them. The HT link right now is at 600Hmz and the core is making great contact as I have reseated it 4 times, all with the same result. I'm using good 'ole AS5 as the TIM.

BTW, I thank you for your responces.

[QuietIce]

The latest BIOS (2.30) has no HTT cap. I chose my ram timmings based off the guy I bought the processor off of. He is an avid OCer and he played around with the settings for about an hour until everything was smooth. When he left, it was running smoothly at 2.4Ghz. I have since been very slowly upping the HTT until I hit the wall of 2.5Ghz with those ram timmings. I really don't know much about adjusting ram timmings to fiddle with them. The HT link right now is at 600Hmz and the core is making great contact as I have reseated it 4 times, all with the same result. I'm using good 'ole AS5 as the TIM.

BTW, I thank you for your responces.

You really need to turn down the CPU multi (to maybe x7), set a lower RAM divider, and then start turning up your HTT so see how high it will go and remain stable.

No board goes forever - even my better A8N32s top out from 360-380 ...

[autobot]

Wait, I can lower the multiplier on my 165? I thought the 165's were locked in at 9x.

[csimon]

Wait, I can lower the multiplier on my 165? I thought the 165's were locked in at 9x.

They are locked upward not downward.

[autobot]

Well I tried out lowering my multiplier to 8, dropped down the ram divider to 133 and stability wise I topped out at 313 which of course is 2.5Ghz.
I dunno what the deal with that was until I thought that perhaps my memory was getting too high. I put everything back to my original settings and dropped the divider down to 133 instead of 166. Low and behold, that was the issue. My OCZ Platinum memory receiving 2.7v and at 3-3-3-8 doesn't really like to go over its stock speed of DDR400. I'm so far testing stable at2655Mhz with a load of 45c on core 1 and 42c on core 2.
I'll post again when I hit the wall again. Thank you!!!

[QuietIce]

Well I tried out lowering my multiplier to 8, dropped down the ram divider to 133 and stability wise I topped out at 313 which of course is 2.5Ghz.
I dunno what the deal with that was until I thought that perhaps my memory was getting too high. I put everything back to my original settings and dropped the divider down to 133 instead of 166. Low and behold, that was the issue. My OCZ Platinum memory receiving 2.7v and at 3-3-3-8 doesn't really like to go over its stock speed of DDR400. I'm so far testing stable at2655Mhz with a load of 45c on core 1 and 42c on core 2.
I'll post again when I hit the wall again. Thank you!!!

Glad you figured it out!

Before I OC the whole system I always check out each part by itself; CPU, RAM, HT Link, and HTT - that way there's no surprises later ...

[autobot]

Well, I hit the wall once again at 2789Mhz with a HTT of 310 at the stock 9x multiplier. I tell ya, this OCZ Platinum is really picky! It doesn't like running any faster than DDR400 speeds. Should I drop the memory divider once again (currently at 133) and try to up the HTT some more?

[QuietIce]

Well, I hit the wall once again at 2789Mhz with a HTT of 310 at the stock 9x multiplier. I tell ya, this OCZ Platinum is really picky! It doesn't like running any faster than DDR400 speeds. Should I drop the memory divider once again (currently at 133) and try to up the HTT some more?

You bet!!! NOTHING is more important than CPU speed. I'd drop the RAM until you know what the top end on the CPU is going to be. After that you might want to adjust things but maybe not.

RAM speed is a distant second for OC'ing ...

[biohead]

Ok guys, my 180 is now officially out of order. I'm giving this hobby a break until Phenom hits. I suddenly realized how much money I wasted on short term pleasures.