The Phenom II List of Overclocks

[[XC] Synthetickiller]

Since I haven't seen any 945s here, here's mine. I was really hoping for 4.0ghz, but I can't get the bus that high and a locked multi at 15 is a PITA.

Here's my best 24/7 stable OC so far using a scythe mugen II:


http://valid.canardpc.com/show_oc.php?id=833556

Anyone doing better w/ a 945?

At least I don't think my K9A2 Plat is holding me back. Hell of a board considering how old it is.

[AlleyViper]

Just reporting the results on my "old" 955BE build:
MB: DFI 790FXB-M3H5
RAM: OCZ 1600 CL7 LVAM
CPU: 955BE 0911 EPMW ...090006 (as picking low numbers seemed like a good practice while ago)
Default voltage: 1.325V (quite a low one!)
Cooling: TRUE Black with BTK rev2 in vertical mount, 2xSFlex G at 1.5K
Case: PC-B70 (front in 2x14cm, rear out 2x12cm, closed case)
PSU: PCP&C Q750W
VGA: 4890 8+6 pin (ugly hot idle under cpu!)

Fixed settings:
RAM: 1333 6-6-6-24 1T Unganged, 1.657V (strangely this cpu can't even POST at 1600 divider using orange slots in two M3H5s, only on yellow it does fine, even at 1600-7-6-6-24, else there seems to be a barrier at 1425MHz when upping HTT)
NB:1.250V (def), HTT 2.0GHz (upping to 1.290V didn't help for 3.8 with less vcore)
CPU-NB: 1.275V, 2.6GHz (haven't tried less yet, 2.4 needed 1.225V)
SB 1.2V PLL:1.20V (def)
NB HT: 1.270V (def)
NB PCIE : 1.120V (def)
ACC: ON (auto all cores)

For 12h Prime Stable on Windows Vista 64bit; temps before stopping CPU/SYS/NB ºC:
3.6 GHz (18x200) 1.375V, (when this was tested NB was at 2.4GHz, 1.225V, didn't take temp notes)
3.7 GHz (18.5x200) 1.400V, 48/43/54ºC
3.8 GHz (19x200) 1.475V, 50/43/56ºC (actually made 24h)

My target was 1.450V given that this cpu looks to be leaky, and I don't know if 1.475V is too much for 24/7 use...
I was afraid my PWM would blow, but temps seem to be decent with this case and cooler after long primes, even if in a closed room.

I had BSODs bccode: 124 around 8-10h at 1.450V, and in less than 3h at 1.425V. At 3.8GHz 1.45V I tried both to up NB to 1.290V, or set CPU-NB to 2.4GHz at 1.275V or 2.6GHz at 1.30V without any change to the usual 8-10h BSOD, so I settled at 2.6 1.275V which seems stable and unrelated to this crash. Upping to 1.475V vcore seemed to solve it instantly, even if doubling testing time to be sure, so I guess temps weren't directly the issue.

For what I see around, 1.475V seems way too high for 3.8GHz on a 955be, and it seems to be quite a jump from 3.7GHz at less 0.075V. But I don't know if vista 64 is always in use, or if users are considering the old rule of at least 12h prime. As for a couple hours of prime 1.425V would suffice, while 1.45V would seem to pass a user ended 8h test with this cpu.

If anyone has suggestions on how to drop .025V I'd be grateful! Thanks!

[Bullet92]

AMD Phenom II x4 955BE - 0915 BPBW (CACYC)
Gigabyte MA790FXT-UD5P - F3L BIOS
2*1GB Crucial Ballistix 1800MHz
TRUE + 120mm fan - ~17-20°C room temperature

CPU-Z suicide run.



AOD is not working on my XP so I think this beast can go a little bit more if I can open AOD and rise 1-2 more steps for HTT

[I.nfraR.ed]

Some benches on FryRender, just wanted to beat my old score @ 4026, so i guess that's a success (9 sec faster) @ 4114



Btw, I can't change HT Ref. Clock in AOD even with 1MHz, no matter up or down. It was possible with my Brisbane 5400+, but since then I can't. It's the same with all my 3 Phenom II CPUs. That's why It's harder for me to achieve high frequencies just for validation. I'm only setting the values in bios and rebooting.

[feniksgod]

Heh very very nice townsman but here isn't appreciated this test very much . SuperPI and wPrime are major tests.
We with you are a small fishes in the sea......, here(XS)

[Rippthrough]

Hi guys, long time no see!
Anyway, 955 Black edition, on air, completely stable on XP 32bit - it'll go to 3700mhz on this voltage but the temperatures start pushing it a bit - it's still on the stock cooler.
Northbrigde and HTT will both go to 2.2Ghz with a touch more voltage but I couldn't see the point and dropped them back to stock for now.
I've a way to go yet but I think 4ghz stable on air might be in reach, I'm only on a 4 pin CPU supply rather than the new 8 pin connectors so that should help a little in that regard too:

[skandal]

Is the CACYC AC 0940FPMW step any good?

The CPU is a Phenom II 945 revision C2 :|

[[XC] Synthetickiller]

Is the CACYC AC 0940FPMW step any good?

The CPU is a Phenom II 945 revision C2 :|

What is your 945 clocked to? I'm looking for someone to break 3.9ghz or even hit 4.0ghz.

If you can do 275x15, I'd go nuts!

[afireinside]

Old school 940BE here



Watercooled, 1.6375Vcore 1.425Vnb

[Zeus]

Damn, that's a lot of Vcore!

[skandal]

What is your 945 clocked to? I'm looking for someone to break 3.9ghz or even hit 4.0ghz.

If you can do 275x15, I'd go nuts!

So far... 3.45Ghz with 1.47v. With 1.5v I can't get it stable at 3.6 :| damm chip... it sucks

[afireinside]

Damn, that's a lot of Vcore!

Yeah, I need a reason to buy a C3

Seeing as the phenom 2s have been out a year it may be a bit late to prove a point, but I don't think ~1.65v will seriously effect the life of them. I'll see though.

[BeepBeep2]

I wonder why voltage tolerance has gone UP since 90nm Athlon 64...

I didnt dare to push more than 1.55v into my A64 X2 for 24/7... :P

[afireinside]

I wonder why voltage tolerance has gone UP since 90nm Athlon 64...

I didnt dare to push more than 1.55v into my A64 X2 for 24/7... :P

Well 1.55 is what AMD claims is the max safe for 45nm... 90nm were monsters. I was running like 1.7v on those.

[BeepBeep2]

Well 1.55 is what AMD claims is the max safe for 45nm... 90nm were monsters. I was running like 1.7v on those.

?????

I think the most I saw on ripping was 1.65 on air and 1.75-1.78 on LN2 or, close to 0c...

Nice to know you were doing 1.7v on air/water. :P

[afireinside]

?????

I think the most I saw on ripping was 1.65 on air and 1.75-1.78 on LN2 or, close to 0c...

Nice to know you were doing 1.7v on air/water. :P

Pretty sure people were pushing 2V on extreme cooling. I was in the 1.9 range on dry ice with my opteron 146. That thing was a beast. Chipped 2 corners of the die, smashed a few SMD caps, and the guy I sold it to still got it prime stable at 3.2ghz on water.

[[XC] Synthetickiller]

So far... 3.45Ghz with 1.47v. With 1.5v I can't get it stable at 3.6 :| damm chip... it sucks

What are your settings?

I noticed for my rig, I had to bump the SB voltage really high to get any stability. If I drop it at all, I lost my OC. Could be due to old hardware though.

PM me if you want my settings, granted, having an am2+ board (MSI K9A2 Plat) that has the old sb600 and ddr2 1000 will change things, or maybe I lucked out with my chip.

[Roger20]

Hey
anyone know anything about this 965 C3
stepping

CACAC AC 0947FPMW

[BeepBeep2]

Pretty sure people were pushing 2V on extreme cooling. I was in the 1.9 range on dry ice with my opteron 146. That thing was a beast. Chipped 2 corners of the die, smashed a few SMD caps, and the guy I sold it to still got it prime stable at 3.2ghz on water.

Sheesh...alas my 5600+ has been cleaned (92% alcohol) and remounted so many times the watermarks have worn away...and scratches have appeared...

I'm gonna end up lapping it someday, I just know I will.

It looks nothing like the pic at left anymore...

[bhokuto]

New To OC'ing, so be kind, beep beep..

Going for the Scythe Mugen 2 soon..

1st Validation:


2nd Validation:


*The first one was done, adjusting in BIOS, stock voltages; CPU multiplier, HT Link Freq, and NB Freq. No CPU/HT Multiplier w/DDR3 @1333MHz in the BIOS
**The second one was done, adjusting BIOS>memory, stock voltages 'manually set'; AOD>CPU multiplier to 17x, CPU/HT Multiplier to 218 and dropped the Memory from 1333MHz to 1066MHz
to adjust memory timings to 7-7-7-20-27-1T @1066MHz, from the 'out-of-the-box' timings and speed of: '9-9-9-24-34-1T @ 1333MHz'

Just getting started, probably not much more can be squeezed out may hit 3.9MHz with this microATX board

[rodrigo.rras]

Damn I very upset with my processor.
It´s phenom II CACAC 0952 C3 stepping.

It´s very good because it´s requeires very low voltage to reach STABLE high clocks.

Like :

3.9ghz (1.35v) + 2.6ghz (NB at 1.25V) STABLE at Prime 95

But There´s nothing i can do. My processor is not prime 95 stable under 4ghz.

I tried so many configurations :

1.39V 4ghz (System crashed then reboot)
1.40V 4ghz (System crashed then reboot)
1.42V 4ghz (System crashed then reboot)
1.44V 4ghz (System crashed then reboot)

please help....

[I.nfraR.ed]

I have 2 CPUs 0952 - one 955 (from the signature) and one 965 (0952GPMW).
965 is stable at 4GHz with 1.425V, but it was impossible for me to get it stable with NB higher than 2.6. The other one (955) is good for 24/7 usage, makes 4.1 with 1.5V and 4170 with 1.55, but it sucks in benches, doesn't like higher voltage at all. Its IMC is better, but not as the one of my first 965 (0945DPAW).
You can try to push some more voltage on the NB and CPU-NB.
You can also try to lower everything except the CPU frequency and see if you could get it stable. Also watch your temps, I don't know what's your cooling, but that could be the problem.

[BeepBeep2]

Damn I very upset with my processor.
It´s phenom II CACAC 0952 C3 stepping.

It´s very good because it´s requeires very low voltage to reach STABLE high clocks.

Like :

3.9ghz (1.35v) + 2.6ghz (NB at 1.25V) STABLE at Prime 95

But There´s nothing i can do. My processor is not prime 95 stable under 4ghz.

I tried so many configurations :

1.39V 4ghz (System crashed then reboot)
1.40V 4ghz (System crashed then reboot)
1.42V 4ghz (System crashed then reboot)
1.44V 4ghz (System crashed then reboot)

please help....

Your results are actually quite good, what cooler are you using? 3.9 Ghz @ 1.35v is amazing. (Mine does about 3.7)

[rodrigo.rras]

Man imagine, I live in Rio de Janeiro. It´s summer.
I think it´s a good processor.
But i really wanted to break the "magic 4 ghz"...


I think I don´t have a Good PSU (Coolermaster 460w)
and also, I don´t have a good motherboard.

It´s a gigabyte GA-78GM - US2H ... no SB750 or SB710...
it´s SB700...

bye
thanks for the response...

[BeepBeep2]

Thats amazing.