The AMD Phenom II OC'ers Club

[Ev1l_HAF]

Ok so i have a AMD Phenom X4 9950 Black Edition is there a club for that?

[Chicken Patty]

If you search by threads I have started there is a thread for 9950's. It's not a club but it's got tons of info once the thread is revived I'm sure it'll get active again. I'm on the phone if not I would link you to it.

[Ev1l_HAF]

thanks i can look for it

[Kei]

There is also the original Phenom Overclocking/Undervolting thread which you can find with a quick search in the 'Overclocking' forum. Couple that with the thread CP started, and you'll have more information than any one person should ever need lol.

Kei

[Chicken Patty]

I just slapped myself for not remembering that thread. Basically almost same CPU so both threads combined will make you go insane if you haven't already.

[BlackOmega]

Let me fix this up for you.

Originally Posted by jjFarking

Not quite
Core Speed = Reference Clock * CPU Multiplier
Northbridge Speed = Reference Clock * Northbridge Multiplier
HyperTransport Link Speed = Reference Clock * HyperTransport Multiplier
Et cetera.

For all intents and purposes, the reference clock is still the good old FSB
So, using my X4 965 as a very basic example:
Core Speed: 200MHz (reference clock) * 17 (CPU's native multiplier) = 3400MHz
HT: 200MHz (reference clock) * 10 (HT's native Link Speed multiplier) = 2000MHz (which translates to 4000MT/s (MegaTransfers/second)).
Et cetera.

Physically the CPU's memory controller still needs to communicate with the RAM, which is done over the bus, which is why changing the HT/NB does not affect the reference clock, but the HT/NB are affected by a change in reference clock.
This also explains why increasing the CPU's multiplier gives a greater through-put (bandwidth), than does raising the reference clock to get to the same Clock Speed.
To see a simple demonstration of that, run IntelBurnTest, with the CPU @ 4GHz.
At 20x 200reference clock, a 1920mb calculation will take less than 55 seconds on the X4 965 (roughly 42 GFlops).
At 16x 250reference clock, that same calculation will take more than 55 seconds (roughly 40/41 GFlops).
Your mileage may vary, of course, but the disparity will be there.
I just used 4GHz as an example, as it's a relatively easy (..) to get speed on the X4 965. Use whichever speed you like; the effect is still the same regardless

The whole point of AMD utilising HT, is to remove the [potential/encountered] inadequacies that the FSB has with the modern architectures, which can cause bottlenecks and/or other slow-downs.
The way things are going, I can't imagine FSB disappearing completely in a fairly short time-frame.
It is bound to be scrapped though.

BTW, HT is (a few technical exceptions aside) the same as Intel's QPI - both negate the need for many tasks that used to go through the FSB, hence the on-board memory controllers (thus including the NB) of the respective camps' CPUs.

Fascinating stuff

Edit:
This is what I remember from AMD's & the HyperTransport Consortium's respective white papers on the subject.
My memory is not as reliable as it used to be, which is why we have people like erocker & Kei, to name but a few, to help us steer through the murky waters

There fixed .

FrontSide Bus is an intel term, hence it does not apply to AMD's. However, the reference clocks function is similar to the FSB.
It's basically a different name for something that has a very similar purpose. It's the difference between potato (po.tae.toe) and potato (po.tah.toe).

I've gotten somewhat decent at figuring out RAM timings and such. So far my formulae have not failed me. I must say that AM3 has opened my eyes to it more. On my 939's it was somewhat straightforward, and really overclocking the RAM, unless you could get it to DDR 500, was kind of a waste of time for very little benefit. AM3 on the other hand shows considerably more improvement from overclocking its RAM.

[trt740]

how about this boys http://www.xtremesystems.org/forums/...d.php?t=244002

[eidairaman1]

Cherry Picked Engineering Sample it sounds like, then a board that is no frills meaning less power draw, and probably a Modified bios out the ass.

[Chicken Patty]

Originally Posted by eidairaman1

Cherry Picked Engineering Sample it sounds like, then a board that is no frills meaning less power draw, and probably a Modified bios out the ass.

Not necessarily, it ain't the first one that does that.




on the other hand, I just got my Corsair H50, woot woot

[Wile E]

No benches at that speed tho?

[erocker]

Is anyone here running a 955 c3?

[p_o_s_pc]

biostar boards clock like mad its been proven time and time again. they just don't hold up that great sadly but that is a nice clock

[Paulieg]

Originally Posted by erocker

Is anyone here running a 955 c3?

I'm running a 965 C3. I have a 955 C3 still in the box, but it's on it's way to System Viper tomorrow.

[jjFarking]

Originally Posted by BlackOmega

FrontSide Bus is an intel term, hence it does not apply to AMD's. However, the reference clocks function is similar to the FSB.
It's basically a different name for something that has a very similar purpose. It's the difference between potato (po.tae.toe) and potato (po.tah.toe).

I've gotten somewhat decent at figuring out RAM timings and such. So far my formulae have not failed me. I must say that AM3 has opened my eyes to it more. On my 939's it was somewhat straightforward, and really overclocking the RAM, unless you could get it to DDR 500, was kind of a waste of time for very little benefit. AM3 on the other hand shows considerably more improvement from overclocking its RAM.

I did say that it was 'for all intents and purposes', meaning that it isn't quite the correct term, but it's what most people remember.
It's just a bus speed, of course

I just wish AMD was a little clearer on the subject & explain once & for all the proper relationship between the various buses & what they affect.

Would you happen to have a link to something a little more up-to-date than the old white papers I referred to?
It's annoying when trying to glean correct info, when the only info I can rely on is a little out-of-date

Cheers mate

EDIT:
I might add that the continuing incorrect use of 'FSB', would have to be, in part at least, blamed on the mobo/BIOS manufacturers, as they continue to use this terminology, even when it's not at all applicable.
2¢

[Chicken Patty]

For now the max load temps while crunching are about the same at 47-48ºc. However I had the case opened with the Megashadow, with the H50 case is closed. Tomorrow I'll try to swap rotation of the fan out and see how I get better results. Right now I have it blowing in. I'll keep you guys posted.

[BlackOmega]

Originally Posted by jjFarking

I did say that it was 'for all intents and purposes', meaning that it isn't quite the correct term, but it's what most people remember.
It's just a bus speed, of course

I just wish AMD was a little clearer on the subject & explain once & for all the proper relationship between the various buses & what they affect.

Would you happen to have a link to something a little more up-to-date than the old white papers I referred to?
It's annoying when trying to glean correct info, when the only info I can rely on is a little out-of-date

Cheers mate

EDIT:
I might add that the continuing incorrect use of 'FSB', would have to be, in part at least, blamed on the mobo/BIOS manufacturers, as they continue to use this terminology, even when it's not at all applicable.
2¢

I knew what you meant .
I'll have to see if I can find the info I searched to gather all of that. But you're absolutely right, AMD is very ambiguous about the buses or even why the reference clock stock is 200. For all I could tell, they just liked the number.

From what I gather about the ref clock is that it's sole purpose is exactly that. A set number that all of the other pieces (CPU, GPU, etc) use to reference and keep in sync with the rest of the system.

And you're totally right about the BIOSes, the BIOS on my board (790FX-GD70) has certain voltage options like CPU PLL voltage. It gives a formula of how to obtain this number which is something like .5 RAM voltage + CPU VDDQ voltage but no more than x. Well that's great and all but there is no CPU VDDQ voltage. (It might not be VDDQ, I'll have to check to verify). So I'm still like wtf is this? Then when looking over some intel overclocking I see that this is a voltage setting commonly adjusted. Yet this is the first I've seen it on an AMD board. But then agian this is my first AM3 board.

SLightly OT here but look at this.

[mR Yellow]

Originally Posted by erocker

Is anyone here running a 955 c3?

I am. I can easily get the CPU to boot into windows over 4ghz but can't get the system stable under stress testing situations. I still don't know all the settings on my mobo (M4A79T Delux).

[exodusprime1337]

Originally Posted by mR Yellow

I am. I can easily get the CPU to boot into windows over 4ghz but can't get the system stable under stress testing situations. I still don't know all the settings on my mobo (M4A79T Delux).

lol same thing here, i'm finding though without load line calibration getting over 4 ghz is gonna be near impossible for me, i shouldn't have cheaped out and got the m4a, and got the crosshair instead, the vdroop on this board is just effin insane.

[mR Yellow]

Originally Posted by exodusprime1337

lol same thing here, i'm finding though without load line calibration getting over 4 ghz is gonna be near impossible for me, i shouldn't have cheaped out and got the m4a, and got the crosshair instead, the vdroop on this board is just effin insane.

Are u running stable @ 4ghz (profile says u are running @ 4ghz) under stress testing?
If so, would u mind sharing settings?

Thanx

[Hunt3r]

Originally Posted by BlackOmega

I knew what you meant .
I'll have to see if I can find the info I searched to gather all of that. But you're absolutely right, AMD is very ambiguous about the buses or even why the reference clock stock is 200. For all I could tell, they just liked the number.

From what I gather about the ref clock is that it's sole purpose is exactly that. A set number that all of the other pieces (CPU, GPU, etc) use to reference and keep in sync with the rest of the system.

And you're totally right about the BIOSes, the BIOS on my board (790FX-GD70) has certain voltage options like CPU PLL voltage. It gives a formula of how to obtain this number which is something like .5 RAM voltage + CPU VDDQ voltage but no more than x. Well that's great and all but there is no CPU VDDQ voltage. (It might not be VDDQ, I'll have to check to verify). So I'm still like wtf is this? Then when looking over some intel overclocking I see that this is a voltage setting commonly adjusted. Yet this is the first I've seen it on an AMD board. But then agian this is my first AM3 board.

SLightly OT here but look at this.

I never knew you had to sell the i7 980x .. it must be very strong for the price..hehe

[jjFarking]

Originally Posted by Chicken Patty

For now the max load temps while crunching are about the same at 47-48ºc. However I had the case opened with the Megashadow, with the H50 case is closed. Tomorrow I'll try to swap rotation of the fan out and see how I get better results. Right now I have it blowing in. I'll keep you guys posted.


http://img.techpowerup.org/100127/P1274404resized.jpg

So.. are you still getting the Obsidian?
Despite its sharp corners, I'm actually starting to get won over by its possibilities.
Sure as hell beats the Element S in every way

Also, why aren't you running 2x 120mm fans on that rad?
Space constraints perhaps?
Not required maybe?


EDIT: very tidy looking mate

[jjFarking]

Originally Posted by Hunt3r

I never knew you had to sell the i7 980x .. it must be very strong for the price..hehe

It's an engineering sample, which means it's créme de la créme in virtually every aspect.
It would be awesome, though the price asked is a little .. steep

[fullinfusion]

Originally Posted by Chicken Patty

For now the max load temps while crunching are about the same at 47-48ºc. However I had the case opened with the Megashadow, with the H50 case is closed. Tomorrow I'll try to swap rotation of the fan out and see how I get better results. Right now I have it blowing in. I'll keep you guys posted.


http://img.techpowerup.org/100127/P1274404resized.jpg

Gee I been away for a while, didnt expect the H50. Looks good though. How does the block mount?

[Chicken Patty]

Originally Posted by jjFarking

So.. are you still getting the Obsidian?
Despite its sharp corners, I'm actually starting to get won over by its possibilities.
Sure as hell beats the Element S in every way

Also, why aren't you running 2x 120mm fans on that rad?
Space constraints perhaps?
Not required maybe?


EDIT: very tidy looking mate

Thanks man, I will get the obsidian later down the road , yes.

I had no timeast night, today I will try to test different fan configurations. I'll keep you'll posted.

Originally Posted by fullinfusion

Gee I been away for a while, didnt expect the H50. Looks good though. How does the block mount?

thanks man, I got my taxes ahead of time so I went for it. The pump mounts super easy, what gave me some trouble was the backplate, I think the screws that bolt up to it were not threaded correctly. However I got it to work.

[fullinfusion]

Hey bro dont it mount to the stock rear plate?

obsidian? who said that? damm you ppl!

also to Kei! your a bastard lol... Im running stock cpu clocks to see how low I can low volt the cpu. I however refuse to run and NB lower than 2600MHz... The nb is running 1.20v (for now) and the cpu the same, the memory is pumping @1600MHz @ 1.60v 1T 7.7.6.15.25.... so far it passes Intel burn test
I find running stock cpu clocks to be just as fast as an o/c'd cpu for a 24/7 rig....
Well you did it! you rubbed your low volt crap on me after all lol!!!!
I guess that makes us even for the H2o shit hey bro lol!