VERY early opty 146 impressions

Ingonuts13 said:
Aren't the Opty's noted as being poor overlcokers anways. They werent really designed for that purpose in mind. Keep the info coming too, I am interested to hear more results.
Click to expand...

As it's been said the S939 Opterons are pretty good overclockers. The S940 Opterons can overclock, it's just that most of the motherboards won't let you do it in the BIOS. You have to use Clockgen or NTune. I've done it with my setup in Windows XP 32bit, the 64bit version doesn't work with Clock gen or NTune. At least not very well. It blue screens in XP x64 the second I apply the settings in Clockgen and NTune doesn't do anything. It worked alright. I hit 3GHz without any issues. I did have to put a divider on my memory though. My ECC memory isn't the best about overclocking. 220MHz is all it will do.
 
well, just to keep everyone up-to-date. speaking for myself only. :) after messing with memory timings, voltage and every setting you can imagine.... my proc can post into windows at up to 3.1ghz. it is SuperPi 1M stable at up to 2.8ghz. however, for me it still comes down to 24 hours of prime95, at which point my proc is only stable at 2.7. i believe anyone on air, should expect similar results. anyone on water should expect the 2.9+ range. just my opinion. i've seen a lot of results on xtremesystems forums and although there are some great clocks on air, no one seems to be testing 24hr prime95. if anyone sees different, let me know, as I would much like to contact that person for advice in getting to higher clock speeds. :)
 
wadec22 said:
well, just to keep everyone up-to-date. speaking for myself only. :) after messing with memory timings, voltage and every setting you can imagine.... my proc can post into windows at up to 3.1ghz. it is SuperPi 1M stable at up to 2.8ghz. however, for me it still comes down to 24 hours of prime95, at which point my proc is only stable at 2.7. i believe anyone on air, should expect similar results. anyone on water should expect the 2.9+ range. just my opinion. i've seen a lot of results on xtremesystems forums and although there are some great clocks on air, no one seems to be testing 24hr prime95. if anyone sees different, let me know, as I would much like to contact that person for advice in getting to higher clock speeds. :)
Click to expand...


Good to hear this news. I completely agree with the Prim95 for 24hrs issue. I may actually consider getting an Opteron now instead of an X2 just to see how fast it will O/C on my rig.
 
I would agree. 2.7 is pretty sweet for the opty, even better if you only get the 144. :)
 
Just to update my sig and such...

My Opteron 146 is perfectly prime 95 24 hour stable at 2.75 ghz at 1.488 V.
I can post at 3.0 ghz but not super pi 1m stable at practically any voltage. It seems not all opterons are created equal. You might have better luck with batch that's closer to x0000 number. since htt (or even ram) doesn't seem to limit my oc but rather the actual ghz of the cpu, i decided to run my 146 at 306x9 :)

I was debating on whether to remove th ihs because this thing runs hot (44-45 C idle); not sure if it'll help significantly.
 
yeah, the CABNE's run rediculously hot. my 144 is getting around the same results as you. with normal air temps on water (38c load) and 1.5v (1.47v reported), my max 32m is around 2.85ghz, prime stable needs around 2.75ghz

the 146 that i'm getting soon has a batch number of 001 :cool:
 
also, i'm finding that a high speed burn is helping my cpu out a lot. get it so that you fail prime in about 1-2 min, run burn-in overnight, then find max.. mine jumped up ~50mhz.. and i don't think it's due to ambient room temp changes ;)
 
No. If you read the full thread, there is one other guy on this particular forum that has one. If you go to xtremesystems forums, there are a good number of people, say twenty, that have them. There are even more than that over there that have the 144. Still need to decide if I'm keeping mine or sticking with my 3200 winchester.
 
i have one but i'm slowly burning in from undervolting and just running a quick OCCT to check what the max HTT is with the lowest volts so far mine is occt stable at 225x10 w/ 1.1v I think. I'm just doing it for the heck of it to see how well it can do undervolted. At stock clocks it will only undervlt to 1.075 (cpu-z reports 1.072) each time OCCT fails I up the volts only buy one increment of .025v and it gets OCCT stable....strange. Every 10mhz I up i need to increase 0.025v on the vcore for it to pass the OCCT stability test.
 
whats the deal with OCCT stability test? what are its advantages over any other standard test?

and you there wadec, what choice is there to make between winnie and 146?
 
asianguy80 said:
I'm just doing it for the heck of it to see how well it can do undervolted.
Click to expand...
great fun ;) i did this for my review.. how low could you get the cpu voltage at stock? :D


als.. apparently OCCT is more demanding than prime95
 
choice between the two, ummm.... let's see 220.00 that I paid for the 146 vs. 130 for my winnie vs. paying 160.00 for the 144 when it comes out :)
 
(cf)Eclipse said:
great fun ;) i did this for my review.. how low could you get the cpu voltage at stock? :D

als.. apparently OCCT is more demanding than prime95
Click to expand...

Stock clocks 1.075 in the bios to get it OCCT stable. I started off with 1.000v but it wouldn't even post past the "Update DMI ... " thing

1.050v I could boot but then I tried running OCCT and opening speedfan after i started the stability test and the system rebooted itself. ram was at 2-2-2-6 with 2.75v
 
You must log in or register to reply here.
Back
Top