The Official Dual Core Opteron Overclocking Thread

Well havnt posted for a while, been real busy, but ive had a little time to play with the 165 VPMW tht Absolute_0 sent me and i can say it is looking good

:)

After some initial teething problems (it was crapping out at the same place as my 170 UPMW) i think i may have a chance of getting prime95 stable at 3Ghz. Ive got the CPU naked and using 1.55v (CPU-Z, real vCore probably closer to 1.6v) and have been priming for 20 minutes. Load temps are hitting 42-44C, this is with my watercooling, fans always running at 5v (for burning in process i think im going to increase fanspeed, just to bring temps down a little).

I dont think i will be able to hold this overclock in the summer though, as the volts seem to be a little too high for when the ambient temps get into the 40's. Hopefully a new mthbrd will give me some headroom with regards to vCore, im keeping my fingers crossed tht the RX520 will provide this. Will update with pics when i get some prime completed.

Thanks to Absolute_0 for hooking me up with a great chip, looks like its time for my 0550UPMW to meet eBay, heheehhee

:p:

She's doing 3 Ghz? Good to hear mongoled! :)

Quote:

---

Originally Posted by mad mikee

I on the other hand call the XP-120 'the core nibbler' since it made nice shiny dust of the edge of my CPU.

---

Hehe, well...that depends on how you do it;)

Done right it's still safer mounting the XP-120 in my opinion.:slap:
:)

Quote:

---

Originally Posted by mad mikee

Go search for my posts w/ Big Typhoon or TTBT, shows how I did it. Just have to do a couple of things and be CAREFUL. I on the other hand call the XP-120 'the core nibbler' since it made nice shiny dust of the edge of my CPU. For best performance, you DO want to compensate for the torquing by compensating w/ support from above or below, whichever is easier.

Only thing about going IHSless is fact that core is slightly convex, so final mounting adjustments need to be done while looking @ temps in bios (or you can lap your core to flatten it like I do :eek: :D )

Lastly, you should lap whichever sink you use (If you haven't already) since I've had both and they were BOTH notably concave, bad match for bare or IHS. In fact if you haven't done that, try it first, since you can usually drive a truck though the machine marks on the bases... (Gain back a degree or 2 there :D)

---

Hmm, what do you think I can use to give support from below? Electrical tape on the chip?

I think I'll probably do this today, with results tonight or tomorrow. A guy on EOCF showed me his X2 4400 and FX57 running naked, vertically with a big typhoon and NO support at all, but I'd rather be safe.

edit: OK, I searched some of your posts and I will definetly use your idea of epoxying the h thing. I'm still looking for something to protect it though :slapass: I'm so damn impatient with this stuff though, I'll probably end up not using anything.

I doubt my reported temps will even get lowered at all, considering they're reported WAY wrong.. it's pretty much always between 38-42C, no matter what. It's 38 at 1.300 and 1.350, it's 40 right now at 1.400 and it goes to 42 at 1.45.. yeah right.

This arrived at my door about an hour ago so here are some initial results. This was on my first boot into windows with stock voltage...

http://img460.imageshack.us/img460/3...133v0is.th.png

More results coming soon...

You can go much further than that! Try 2.9, if it fails back down to 2.8

After you mount it, push straight UP SLOWLY / GENTLY on the fan while watching bios temps, you may notice that the temps go down by a degree or 2.

Then you just have to figure out how to do the same thing WITHOUT THE FINGER :grr: :D Either from above (lifting) or below (like a jack).

Quote:

---

Originally Posted by Repoman

Hmm, what do you think I can use to give support from below? Electrical tape on the chip?

I think I'll probably do this today, with results tonight or tomorrow. A guy on EOCF showed me his X2 4400 and FX57 running naked, vertically with a big typhoon and NO support at all, but I'd rather be safe.

edit: OK, I searched some of your posts and I will definetly use your idea of epoxying the h thing. I'm still looking for something to protect it though :slapass: I'm so damn impatient with this stuff though, I'll probably end up not using anything.

I doubt my reported temps will even get lowered at all, considering they're reported WAY wrong.. it's pretty much always between 38-42C, no matter what. It's 38 at 1.300 and 1.350, it's 40 right now at 1.400 and it goes to 42 at 1.45.. yeah right.

---

Based on the 2 I have, 2.7-2.75Ghz on Default (Check w/ DMM, since most boards overvolt ) 2.8-2.85 w/ .025 more, then it starts climbing. I'm @ 2.9 on both machines w/ 1.4 - 1.425 vcore (real/dmm) :D

Quote:

---

Originally Posted by TheMeatFrog

This arrived at my door about an hour ago so here are some initial results. This was on my first boot into windows with stock voltage...

http://img460.imageshack.us/img460/3...133v0is.th.png

More results coming soon...

---

Quote:

---

Originally Posted by mad mikee

After you mount it, push straight UP SLOWLY / GENTLY on the fan while watching bios temps, you may notice that the temps go down by a degree or 2.

Then you just have to figure out how to do the same thing WITHOUT THE FINGER :grr: :D Either from above (lifting) or below (like a jack).

---

Are you talking about mounting it while in a vertical position? Because that's just crazy :eek: I just cut the IHS off, took 1 minute and everything went well. (besides a bent pin, nothing a small flat screwdriver couldn't fix :) ) Now I'm waiting for the epoxy on the BT to set (thanks for the tip!) I hope I don't crack the core

Mount w/ board/case horizontal.
Then get best temp like that, per my methods.
THEN put back in final position and gently nudge whole assembly up @ the fan while watching the temps.

You are then supporting the HSF @ TWO points, the board mounting and the fan.
Show me a pic of your setup, then I can describe in more relevent terms :fact:

Quote:

---

Originally Posted by Repoman

Are you talking about mounting it while in a vertical position? Because that's just crazy :eek: I just cut the IHS off, took 1 minute and everything went well. (besides a bent pin, nothing a small flat screwdriver couldn't fix :) ) Now I'm waiting for the epoxy on the BT to set (thanks for the tip!) I hope I don't crack the core

---

I know what you mean. I'd get a pic anyway but I don't have a digital camera

I will be running the whole thing horizontal for a while anyway, I'm too damn lazy to stand it up. Plus I change out RAM a lot these days, so I don't even use a side panel. Thanks for your help, I will update once this damn epoxy sets :slapass:

edit: :mad: First mount, temps were the same, I let prime run for a minute and remounted. Second mount, temps were ~1C better, but core 0 failed in 2 minutes where it usually lasts ~1 hour.. remounted again, this time core 1 fails in a few seconds where it's usually 100% stable. :( About to do my 4th mount, I'm hoping the issue is because part of the die didn't have any thermal paste on it when I removed the HSF.. I will be pissed if I ruin this CPU

edit2: remounted again, this time spreading the AS5 evenly to make sure there were no gaps, and I also put electrical tape over the SMD packs. Not so much to prevent the cooler from touching them, but to make mounting easier (with just the naked core, the cooler would shake all over the place when trying to screw it down). Temps seem to have gone down a degree or two vs. previous method (Actually they still hit 40 supposedly but it jumps down to 39 more often. IMO the range from 38-40 is actually like an 8-10 degree C range of temps, as this sensor sucks). Well both prime's are on test 3 in prime now, and I'm just glad I haven't fuxxored my chip completely :woot: for me!

So if you take off your IHS, you MUST 100% spread the thermal paste evenly, or you could have gaps causing just certain part of the core to overheat!

edit3: before, core 0 failed on test 3 in prime.. went to remount again, knocked the chipset cooler off so I had to redo that.. now temps on it are ~1C worse :mad: BUT I finally got a perfect mount, temps are 38C and go down to 37 every once in a while. And prime is now on the 7th test on both cores :woot: These things are uber picky on how you mount them.. :stick: more hassle then it's worth IMO.

Quote:

---

Originally Posted by TheMeatFrog

This arrived at my door about an hour ago so here are some initial results. This was on my first boot into windows with stock voltage...

http://img460.imageshack.us/img460/3...133v0is.th.png

More results coming soon...

---

what cooling?

just sold my laptop for 750, water cooling and 170 on the way!

where have goood water cooling stepping 170's come out of recently. i always hear good steppings of all cpu's come from monarch, id get from them just need clarification

I finally got my PWM IC temps down by mounting a PCI slot dual fan cooling card perpindicular to the motherboard on the side closest to my harddrives. (looks a little ghetto but so what) Max PWM IC temp is now 44C instead of 59-60C.

However I am having the strangest problem... I can dual core prime stable for all night long at 2.8Ghz, 1.35V BIOS, max CPU temp 40C.

But to boot into windows at 3.0Ghz stable I need 1.45V BIOS. Everything is fine until I try to dual prime then temps shoot up very very fast on CPU to 49-50C then core 1 always quits :stick:

So do you think it will work if I get watercooling?? I would hate to buy a watercooling setup and then it just does the same thing. 1.45V is not that extreme is it? Seems like other people here are doing like 1.55V on these
CCB1E 0550VPMWs no problem. Don't know why mine is so picky :(

watercooling would definitely help, air doesn't hold the load down well enough on these thhings. They get too hot and fail, and your only option is more volts, which exacerbates the problem. These VPMWs want water.

Quote:

---

Originally Posted by Absolute_0

watercooling would definitely help, air doesn't hold the load down well enough on these thhings. They get too hot and fail, and your only option is more volts, which exacerbates the problem. These VPMWs want water.

---

Definetly, lowering my CPU temp 1 degree looks like it's gonna give me at least an extra 18-27MHz

I just got a 170 0551 upmw. When I get back home late next week I am going to put my Corsair Hydrocool EX on it ( now running a 146 at 2900 ). What can I expect with this stepping ?

Thanks:banana:

Now you know why I lap the core to flatten it on all of them :D Never have a mounting issues (unless I'm being stupid :rolleyes: )

Any info on 165s 0551 WPMW? Or anyone has a 06 DC Opteron? I've seen 146, 144, etc but can't remember seeing a single 165/170/175 etc...

I'm looking for a 165, maybe 170 that makes 2700 on stock, as the summer is coming really fast here! :D

I just picked up 2 165's with a DFI motherboard.
I am a bit perplexed on the memory settings.

I hit a wall around 290 with my mushkin, not sure it its my cpu or my mem thats carpping out. I was still at 1:1. Are there any settings/timing that needs to be adjusted to reach the 2.8 range, or is it really up the quality of the ram?

put it at 166 or 180 divider and see if it will push

Quote:

---

Originally Posted by KViper19

put it at 166 or 180 divider and see if it will push

---

Okay I tried running at X3 and lowering my dividers to see if I could establish my ram limits. I was at 325 and I decided that was enough.

I swapped chips to test and see if I could find a higher default voltage while I was at it, the second chip (0546) let me in windows at 2.52 with default while the (0550) crapped out at around 2.4.

However on either of these chips, I cannot seen past 2.7Ghz at any rate. But I'm not convinced that its the chips though, I'm thinking more of some settings to accomodate the opterons running at higher HTT's or something.

Not very sure on these motherboard settings though. I was pretty confident with memory timings etc. until I loaded this up. :slobber:

I will try the 166/180 dividers in a bit and post my results

Quote:

---

Originally Posted by KViper19

put it at 166 or 180 divider and see if it will push

---

Alight, I tried the lower dividers and looser timings and no go. at around 2.7 both chips crap out no matter how much voltage I feed 'em. I quit trying at 1.5v

Any other ideas?

Okay, I don't mean to spam the thread but...
I figured out that the main cause of my poor clocks is heat.

First off the DFI thermal sensor is complete crap.
Secondly the stock HSF with the 165 is okay but since its running off the DFI thermal controller it was not removing much heat.

I'l now dual priming at 2.5 on stock volts.
I guess a water cooling setup would be a good investement.