• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

AMD® Phenom™ II Overclocking Thread

No one wants to know of my failures last night - morning.
It didn't BSOD, it didn't GSOD..

It shut down.. Why? Clever Martin here left flashget downloading with "shut down when complete" checked....

My GF comes at half 7, so I've got a slot to stress test then, till 8 am.. So.. Over a 12 hour Blend..

There's a plus side to this, as the ambient in my room raises from the PC, the fastest the clothes come off. XD.

Having read the above post however, I'm going to reduce my RAM speed to 1333MHZ. I'm going to put my Vcore to 1.525 (Even though I'm sure, I'll only need 1.475v when it comes down to "fine tuning".), my NB and HT to 2.6GHZ.

I'm running 1.35v on my CPU NB, does anyone think I'll need more?.
 
Problem is people are using all sorts of benching programs - there's benches on here that show 3.9Ghz but when you look at the same persons sig they're only actually running 3.7 now! If it aint prime95 blend stable - it ain't worth having :p

Plus no-one is posting they're actual voltages and settings from BIOS - CPU-Z only gives a tad bit of info compared to what we overclockers have to deal with in BIOS.....[/QUOTE]

if you read whole thread matey you might see and like ive said before yes a lot can run higher than they do but would you run them clocks 24/7 ?

its all ok saying in my sig 3.9 or 4.0 but i wouldnt run that 24/7 and 3.9 is fully stable but i dont like the jump in voltage to do it or temp increase. the speed increase is minimal but the risk of damaging stuff is a lot greater = not worth it

i have a c2 955 runs at 3.7 with 1.4v to run at 3.9 i need 1.5v i can boot and do windiows apps and general light stuff at 4.1 with 1.5 but...if i try to run ibt it fails pretty quickly.

so if i went for some high voltage like 1.55 or above i believe it would do 4.0 stable allday. like you see in these threads the wall and real tricky bit is gettinng past 3.8 stable 24/7 and this is where most come undone. after that its mostly a bit of luck with a good chip mobo combo.

so i settled on a sweet spot of 3.7 which is fine and more than fast enough without the silly risks and epeen for sigs :D

anywhos do any of these 6 cores coming out work on our boards (785) ?
 
Last edited:
Martini: I would only over-volt if I were sure the increased temperatures weren't themselves going to cause instability.

In terms of the clock I would go for just a vanilla CPU frequency clock first (NB at CPU Frequency - 1200mhz), keep the RAM at 1333mhz and loose. From there you can tighten up the NB and RAM.. NB for up to around 2.4 - 2.6ghz should be fine on stock you might need + .025 / + 0.05
 
@ >Dg< Never been one for 'suicide overclocks' as I call them ;) I could probably do a couple of runs of SuperPi etc @4ghz but see it as pointless due to the extra vcore etc - the extra volts ain't worth it like you've pointed out mate. I wasn't referring to anyone in particular dude - don't think I was picking on you! :)

I've got an 890 board mate which I know will defo work with thuban (6 core chips) - CPU-Z is misreporting it as a 785. As far as I'm aware Thuban CPU's will be supported on some of the older 78x boards with BIOS updates.

So is it safe to assume guys as with the Q6600 chips - The lower the VID reports in coretemp the better for overclocking? (trust me to get a 1.4v VID chip!)
 
Last edited:
I'm not entirely sure how AMD determines VID....but for Intel having a low VID does not always mean it's a good clocker...in fact it has more to do with how hot the silicon runs. For example the best Wolfdale I've ever had also had the highest VID I'd ever seen on an E8400 (1.2625V), it did over 4.5GHz totally stable on only 1.344V...it was also the coolest. In contrast I had an E8500 once with a 1.175VID and it ran hotter than the sun....it was a great clocker but even on watercooling temperatures were extremely hard to keep under control (i'm talking 120C then shut down after about 5 seconds of IntelBurnTest....and this was with a pretty decent WC setup)
 
Last edited:
Yes I know, I think there was a definite correlation between low VID and good clockers, but I'm just saying that (with Intel chips at least) it goes hand in hand with high temperatures, thus limiting the overclock....and also that I have had a few low VID chips that didn't clock nearly as well as some very high VID chips.

...but anyway, to answer your question, yes I think there's a correlation between lower VID AMD chips and good clockers :p Martini's chip has a decent VID and seems to be getting on pretty well with it :D
 
lol - thats exactly what I was looking at too - martinis VID!

I tried to email you on your email address gurusan - got bounced back - any chance you can post/email the settings your using for 3.8 please? ta.
 
Last edited:
emailed your hotmail account mate - thats what bounced back ;)

how did you get 3.8 - was it 200x19 or something else? also did u have to upp any other volts. thanks fella.
 
My settings if it helps:

Clock: 205
Multi: 18.5
cpunb: 2600
ht: 2600
RAM: 1600

Vcore: 1.4875
cpu-nb v: 1.3125

loadline calibration: 16%

Remember that all chips are different. e.g. I cant get 200x19 stable even at 1.5 vcore
 
thanks guys - your stars!

I've proven I can do 200x19 on 1.425v - so to get my memory back to 1600 I'll need to up CPU-NB and HTT to 2400-2600 I'm guessing. Also noticed the last time I ran 200x19 on DDR3-1600 the settings were slightly lax-er for memory. Will report back probably saturday after some more tweak 2nite :)

thanks fellas.
 
gurusan: - I notice that my cpunb v is much lower than yours but my vcore is quite a bit higher. I wonder if I could get up to 3.8 (from 3.79) by increasing nb.
Might be worth a try... :)
 
Maybe I should be aiming for the next cpu-nb step up. I seem to have plenty of voltage head room. Anyone know if the performance increase is worth it?
 
Well in my brief testing I found 2800 to be the limit of both IMCs I've tried (705e & 555). I've not given them more than 1.45V though.

Personally I would try to aim for the highest CPU clock you can first, then work on tweaking memory and NB.
 
Annoyingly I do seem to be at my highest cpu clock. Any increase of core clock or multi causes a fail. I have tried:

19x200 - fail
18.5x205 - pass
18.5x206 - fail
18x212 - fail
17.5x218 - fail

I have just had a flash of inspiration though. I have a feeling that when I did my 19x200 test my memory was clocked at 1600 speeds but the cpu-nb was at stock 2000 speeds. Perhaps this caused the fail. I should try 19x200 again with cpu-nb at 2400/2600.

Also very stupid question: why when we talk about 1600mhz memory does it say 800mhz in cpu-z?
 
Yeah it's a good idea to put your memory to 1333 when trying to find the CPU limit as to eliminate potential problems with the memory or IMC.

...and it's DDR (double data rate), so it has a real clock of 800MHz with an effective clock of 1600MHz.
 
You must log in or register to reply here.
Back
Top Bottom