I picked mine up same day as Tony (we were in the same town ) but I've had tons of problems and been pulling my hair out over them.

I'll explain fully later but let me make one thing clear:

My old Phenoms did NOT die! It was a MB/BIOS or less likely, a coldboot problem

When I got my last 9600BE I tested it in my 790FX/780G/770 boards. All three gave me C1 error (memory initialization). It was brand new.
The BIOS I use is P0J and it defaults to 800 RAM, so it wasn't 1066 mode problems.
3 PSUs, two brand new, they weren't the problem nor the rails, nor temperature.
When I got the new 9850, I made sure to test in a working DFI 790FX at my uncles firm. A pain to borrow it off someone, but I wanted to make sure it works fine - booted first time, primed 5mins fine, job done.

Came home, tried all 3 boards I had (MSI 790FX, Abit 770, GBT 780G), all 3 did not boot 9850, same C1 problem.

Make note: other people may experience this problem -it ONLY occurs if you reset CMOS or shutdown and leave it off to, well, become cold

So how did I resolve it?

I just left it on, and let it restart on and on for 2 days. Yesterday night late it started up all by itself.

I can't oc, because if I do it without flashing a new better BIOS, and the oc fails -> clear CMOS, it will very likely not start up for another 2 or more days.

I have to go off, am very busy elsewhere, but let me say it is definitely much better than the B2 in terms of max speed. 2900MHz benched stock voltage however it was not stable. 2800 was rock stable though.

Power draw is higher than an equivalent clocked 9600BE B2 at the same voltage, much higher. They've not just increased voltage but increased current.

I haven't got a heatsink on the CPU yet so can't oc, sorry

Snapped the AM2 mount for HSF and sliced my full finger in a 152CFM Delta - just waiting to get my Xigmatek + Delta on here, which is in a different town

CPUZ has problems, it isn't validating again for some odd reason http://valid.x86-secret.com/show_oc.php?id=340474

Ah, it only validates upto 14x multi: http://valid.x86-secret.com/show_oc.php?id=340468

4th try, it validated 2900: http://valid.x86-secret.com/show_oc.php?id=340489

Many tools seem buggy with 9850. Only increased multi yet, not even tried 3000/NB/HT/RAM.

BIOS, 3200 no HSF 1.47v gets stuck at Windows logon.

Stock CPU VID 1.300
Stock CPU Volts 1.296/1.288
Stock NB VID 1.300
Stock NB Volts 1.296/1.288
Stock CPU Speed 2500
Stock CPU Multi 12.5x
Stock HT Speed 200
Stock NB/HT Speed 2000

Some screenshots (I have not tried much, only had a few minutes uptime) and some 2500/2600 comparisons with old 9600 and 9600 BE on the same EVEREST version (no tweak, just normal).


























I'll push it much later - have too much to sort out. For reference, this is what my last 9600 BE did max stable and it was my fully tweaked 24/7 setting idling less than 3W DC for CPU=>






This time around, I'm going to start off by undervolting at stock speed rather than the other way

BTW its a hot chip, you'll need good cooling for it (under load) but you should know that by its 125W heat output rating. 6400 was a hot chip too. I suspect all air oc's will be limited by cooling, water should do a little better. The Tctl degrees it idles at 2600 stock volts is 14C more than my 9600BE did at 2691 1.352v with stock HSF lying on it (not clamped, but still on).