AMD's ACC software optimized for BE cpus?

poohbear

Platinum Member
Mar 11, 2003
2,284
5
81
Hey all, i was reading this article that says amd's ACC overclocking software was created to use w/ BE cpus cause it does something "special" w/ the cpu multiplier inparticular, and hardly does as good when doing fsb overclocking:

http://ixbtlabs.com/articles3/...n-a7da-s-790gx-p3.html

its saying BE cpus used to be non-essential for overclocking since most cpus had high multipliers and only needed a slight increase in fsb, so the price premium for BE processors was not worth it. W/ the release of ACC this all changed cause w/ BE cpus it allowed for much higher overclocks:

Up until recently buying Black Edition processors has been a matter of taste. What concerns Phenoms, owing to high default multipliers they usually require an insignificant raise of FSB clock even for serious overclocking (with increased voltage). Most motherboards can do that, at least among models for overclockers. Besides, any processor can be overclocked by raising the reference clock, while only several models come as Black Editions.

However, with the appearance of "mysterious" Advanced Clock Calibration in SB750 Southbridge, processors with the unlocked multiplier become a technically expedient choice (taking into account democratic prices for Phenom X4 9850 Black Edition, such processors are economically expedient as well). We've set ACC +2% (for all cores) and obtained extra 400 MHz to the result without ACC. We were surprised to see the overclocking potential so expanded and full stability at such frequencies. With disabled ACC, further overclocking resulted in startup failures, and raising the voltage by any value was of no help (these facts suggested that the overclocking potential was squeezed dry). Note that ACC is effective only when you overclock the system by increasing the multiplier. At least we've seen it in two Phenom processors (see the table above). This technology had absolutely no effect on Phenom X4 9550, overclocked by raising the bus frequency. The same with the 9850. When we tried to overclock them above 2900 MHz by raising the bus frequency, the system immediately responded with instability. By the way, when we installed an Athlon processor, the ACC option disappeared in BIOS menu. So we couldn't check whether this technology can squeeze extra MHz from unlocked models of this series. From the look of it, the official recommendation from AMD to enable ACC only for Phenom Black Edition processors is well grounded.


any truth to that? It says they managed a whopping 400mhz extra overclock w/ ACC? they were using an older phenom1 cpu, but im guessing that still carries over to phenom2's no?

thanks in advance.
 

thilanliyan

Lifer
Jun 21, 2005
12,060
2,273
126
It does help with the OCing (ie. I'm Prime 95 small FTT stable at 3GHz at 1.35v with ACC but not without)...and yes from reviews I've seen it does help greatly with older Phenoms. I don't have a Phenom II so can't test.
 

Flipped Gazelle

Diamond Member
Sep 5, 2004
6,666
3
81
ACC is reportedly not as necessary when overclocking PhII's and PhI's.

I had a 9850 non-BE, and I hit 3.2 ghz with my 790GX motherboard, and also an Nvidia Geforce 8200 motherboard.

My "guess" at this is that ACC was necessary to correct an insufficiency in AMD's chipsets, whereas Nvidia never had that problem to begin with.
 

Flipped Gazelle

Diamond Member
Sep 5, 2004
6,666
3
81
Originally posted by: cusideabelincoln
Did you hit 3.2GHz on the 790GX board with ACC turned off?

No, ACC was on when I had the 9850 in that board. I wish I had thought to make a comparison.

Of course, there is no ACC on Nvidia chipsets, and I got to 3.2 on that...
 

poohbear

Platinum Member
Mar 11, 2003
2,284
5
81
@ Flipped Gazelle, did ACC make a difference w/ your x3-710?


Originally posted by: dmens
eh, i still think ACC is PLL skew manipulation. just my guess.

what's PLL Skew?
 

heyheybooboo

Diamond Member
Jun 29, 2007
6,278
0
0
Originally posted by: poohbear
@ Flipped Gazelle, did ACC make a difference w/ your x3-710?


Originally posted by: dmens
eh, i still think ACC is PLL skew manipulation. just my guess.

what's PLL Skew?

PLL = Phase Lock Loop

Remember the old b&w Sci-Fi movies with the mad scientists and oscilloscopes in their labs?

The various timings in a computer have to be in 'phase' - think of the waves on that oscilloscope. 'Skew' (and jitter) is when the waves do not run congruently. This is really important in over clocking - all the individual timings of the computer must be in phase or you will crash.

The '600' south bridge revealed some PLL problems when over clocking that would push timing out of phase. SB700 was only marginally better.

SB750 (and ACC) go a long way in correcting the timings problems - It is hoped the new SB800 (due out late summer IIRC) will build on these improvements.
 

Flipped Gazelle

Diamond Member
Sep 5, 2004
6,666
3
81
Originally posted by: poohbear
@ Flipped Gazelle, did ACC make a difference w/ your x3-710?


Originally posted by: dmens
eh, i still think ACC is PLL skew manipulation. just my guess.

what's PLL Skew?

The only difference ACC made with my X3 710 is enabling the 4th core. Otherwise, the max overclock is the same with/without ACC. From what I've read, it seems that ACC is just not as necessary with PhII.
 

poohbear

Platinum Member
Mar 11, 2003
2,284
5
81

thanks heyheybooboo for the explanation, interesting stuff. Flipped gazelle good to know PhIIs are'nt effected one way or another, one less hassle when overclocking.:)