Neo2 and an Opteron 165

[Nexus Realized]

Nice results rseven, keep us posted.

 

[rseven]

Thanks guys. I can't say as I did very much on this one. All the work was done when I spent 8 weeks struggling with this board overclocking my Winchester. Now, I knew what the board and ram could do, so I just needed to bump up the htt till I couldn't pass Prime and then pull it back. I just passed 22+ hours of Prime, and I'm going to run Super Pi 32m, but I have really have no doubt that it will pass without a problem. I am looking forward to seeing the temps of my winchester on another board to have a point of comparison as I swear this board just reads a good 5 degrees too high. I will be much more comfortable to be able to verify that on another board. If that proves to be the case then I may experiment further to see how far this CPU will go, but in the meantime I am quite happy to have landed another first rate CPU. Maybe I should buy a lottery ticket tomorrow.
Thank you all for your encouragement and support. It's really great to have a whole gang of overclockers cheering you on.

 

[ixtapalapaquetl]

I know Opterons tend to run hot, but I have never seen one run THAT hot at such low voltage. I really think there is some bad contact going on here. I would first suggest reseating your HS - though you may lose some hours of AS5 settling, you might gain a ton in temps.

If that doesn't do anything for you, I would strongly consider pulling the IHS. In many Opterons, they make awful contact with the core. If you are scared about crushing your core, you could always put some AS5 on it and then reapply your IHS using some super glue on the corners.

In any case, good luck! If you can get those temps under control, you may have a killer chip.

 

[deathman20]

Nice chip, and ouch that CPU runs warm. I'd reseat that beast that should run much cooler. Took me a few tries before I was able to say my HS was seated correctly with the IHS and without the IHS's on.

 

[rseven]

Well. I thought the same thing and took it off and reseated it, but the temps were the same. Also, when I checked the old contact it looked fine. I have remounted XP90's enough times now that I can do it without all the difficulty I had the first couple of times. The temps idle at 42 and load at 51. That idle temp is 9 degrees higher than my old Winchester. Poor contact between the IHS and the cores sounds very likely. I have never removed an IHS before, but it looks like I might have to learn or burn. I would never mount an Xp90 on a CPU without an IHS as it just puts way too much pressure on the side of the core when mounting. I do have a Swiftech that mounts straight down, but it's not as an effective heatsink, so I may just lose what I gain. Perhaps removing it and putting it back is the best option, but I'll need to work up some nerve.
BTW, it passed Super Pi 32m without any difficulty.

 

[evancold]

good job Rich, looking forward to seeing how you get your temp under control

 

[mysterfix]

Hey rseven, you got a killer chip there man! I'm glad to see you got lucky twice and your "Hell Spawn" Neo2 is still holding up for you. I'm going to be moving to the dualcore bandwagon in about 2 weeks here and was debating on whether to take a chance on this stepping.

I think I just might since yours is sooo pimp.

 

[rseven]

Thanks Mysterfix, but I'm not out of the woods yet. This afternoon I decided to switch back to my Swiftech heatsink. I wanted to see how much warmer the temps would be in case I went to remove the IHS and go with a safer heatsink to attach to the CPU. Well, I switched everything and when I went to boot I couldn't get into windows. I tried lowering to stock speeds, safe mode and last known working config. I ended up doing a windows repair installation. After, when I kicked it up to 9x280, I got into windows but then got the BSOD in about five minutes. Then I brought the system back down to stock speeds and reinstalled the XP90. (The temps were about 3 degrees warmer with the swiftech) I put a little extra AS5 as I had noticed something less than a fifty cent piece cover last time. I also slowed by ram a tad to CL2.5. Now, it's back up to 9x285 and running great. I believe the problem was my ram timings. Though they ran fine with a 166 divider and CL2 timings at 9x282 for many months with my winchester, I think I corrupted something in the registry and it was crash and burn time.

Now my big decision is whether I remove my IHS? My load temps are 50 and idle is 41. If I do it and use the XP90 it could make a big difference, but I don't want to crack the core with that sideways mount. I really don't know what to do, but I'll have to do something before the weather gets warm, because this CPU is still hot at stock speeds.

 

[masterofpuppets]

Looking into buying one of these. How do they compare to CCBWE's?

 

[mysterfix]

Maybe you could use a shim like the ones used on the AXP's, you might have to grind it down a bit but it should work.

 

[Reefa_Madness]

Hey bud, weren't you the one telling everyone to quit chasing the Opterons? Now I know why...you wanted to be able to get one for yourself , and from all accounts you got you a good one, too.

Looking good, outside of those temps being on the high side. Waiting until your friend tests your old Winnie might help, but only if he is using the same type of cooler that you were (the XP-90), otherwise the comparison might not be valid.

Lots of guys have have good results removing the IHS (see g0dM@n's thread) and re-installing with some A5, but it does carry risk and sure makes it hard to RMA if the need arises (we wont go into the ethics of RMA'ing an overclocked cpu).

Alternatively, consider a moderately priced WC kit to keep the temps down. I just picked one up myself to play around with...never thought that I would intentionally mix water and electronics (I equate it to jumping out of a perfectly good airplane just to be able to skydive ) but I decided to give it a go in a testing station that I'm setting up. This might just be a good time for you, as well.

 

[DaWiper]

Now my big decision is whether I remove my IHS? My load temps are 50 and idle is 41. If I do it and use the XP90 it could make a big difference, but I don't want to crack the core with that sideways mount. I really don't know what to do, but I'll have to do something before the weather gets warm, because this CPU is still hot at stock speeds.

I read(think it was here at ocforums) that the trick with mounting a xp90 or similar on a a64 without a ihs is to loosen the frame from the motherboard then hook up the xp 90, then tighten the bolts. I've never tried it myself but it sounds like it could work. This way you should not need as much force to mount the xp90 on the frame.

 

[deathman20]

Well at least mounting a XP-120 might be slightly different then a XP-90 but I had no issues mounting it on my naked core. Heck it was even easier to mount it after I took the IHS off on the mobo. But still a pain in the butt removing the ram hehe.

Oh ya, No shim used here eaither. The way at least my HS sits is it rests on the plastic piece on the socket and im able to lift it slightly and set it right down ontop of the core and clamp the last 2 clips. Works like a charm.

 

[dave_graham]

just an FYI, the current 165s that have been sent out are 0550VPMWs. I'm waiting with bated breath for the oc data to come back.

dave

 

[nutpo]

I just got a 0550 in on thursday morning... So far i've been able to push it at 2500 with vcore at 1.45. I'm gonna be trying to push it some more over the weekend.

 

[rseven]

Thank you for all the valuable input, gentlemen. DaWiper, that idea of loosening the plastic frame sounds like it would work quite well. At this point I'm putting it on hold till I get the results with the winnie. Also, I have so more evidence that it's just this dumb Neo2. Last night I was idling at 41 and loading at 50. This morning I turn on my computer and the temp is 27! I rebooted several times and the idle is in the high 20's. I Prime it and it goes up to 37. I was out all day and when I came home and turn on the system, now the temps are 40 degrees! Since I'm convinced it's the board, and since my studio is a very cool 65 tonight, I pushed the voltage to OMG, 1.35! I am now priming it at 2610MHz and it's been stable for 2 hours. The temp is 50, but I don't really believe it. Since 50 is still not dangerous even if it's correct, I will probably try to push it higher. I believe the reason I failed Prime at 2601 the other night was my ram was still at CL2, and TCCD just can't run 235MHz at CL2, or at least mine can't. Since 1.35 is really still stock voltage, (I believe stock is 1.30-1.35) I have to say this is one amazing CPU. I don't think this board is capable of much more than 300htt, so I will never know what it's fully capable of doing. I am also wondering, what temp it will report tomorrow. Another thing the board is doing is reading hot on cold boots. This was a flaw in the earlier versions of the Neo2 bios. Well, I guess it's back.

Well, that's it for now. Again, I thank you all for the help and support. What a great forum! Cheers!

 

[rseven]

Well, here's an update on my saga. Last night, I ran Prime stable at 9x290=2610 with 1.35 vcore for 2 hours and 20 minutes. At that point I stopped it and turned it up to 9x295 and ran Prime another 30 minutes without a hitch and then turned it off before going to sleep. All this time my idle temp was 41 and load was 50. This morning I booted up at that same speed and now my idle temp is 28 and load temp is 37. Go figure! I cranked it up to 9x300 and failed Prime in 4 minutes. I increased the voltage to 1.37 and failed Prime in 2 minutes. My conclusion is that my board does not want to run stable at 300htt. Now it has in the past with a different bios. In fact, this is not the most stable bios I could be running and I should go back to the Sideeffect 1A version I had been using, which seemed to be the best. In the mean time, I turned down the htt to 298 and put the voltage back to 1.35 and Primed it. It failed Prime after 23 minutes at a load temp of 37. I am thinking that I can get it stable at htt 294 or 295 at stock voltage. Hey, I will be quite happy with 2.65GHz at stock voltage. (Happy is an understatement!)

About these goofy temp readings. Maybe if I add them together and divide by 2, I might get an accurate reading? Perhaps a different bios will help, but my experience with the Neo2 is that it marches to the beat of it's own drummer no matter what bios it has in it.

 

[Bhugarov]

I hate you ...

No seriously, I'm very happy for your continued streak of luck with the CPU's. It couldn't happen to a better guy.
I've been AWOL for the holidays and just found this thread. I think that you are correct in assuming that the Neo could be reporting bogus temps, but some of the blame may be due to the IHS. If the speader is is as "good" as some of the Winnies I've seen you may just have a contact issue. In any case it looks like you are getting a real nice clock at the moment, (and knowing you it will be even better with time).

P.S. You aren't the only one who was disappointed with the v1.b. bios.
Congrats! and keep us posted.

 

[rseven]

Thanks Bhugarov. I can feel thoughs jealous vibes. I currently 2+ hours into Priming at 9x295=2655MHz and so far so good. Cheers!

 

[deathman20]

Thanks Bhugarov. I can feel thoughs jealous vibes. I currently 2+ hours into Priming at 9x295=9655MHz and so far so good. Cheers!

9.6Ghz sweet I want one hehe