Fun with a 6600K (Richland)

[Daveburt714]

I bought this chip/board to replace my server, but for some reason I just can't quit tweaking it long enough to replace the old rig...

I've been following the threads on here (as always) and have seen some folks talking about AMD's performance/power consumption, and their imminent doom.
Personally, I'm not impressed with the high priced, power hog FX-9xxx chips! Granted, they would be fun to play with, but I couldn't justify the price performance ratio considering that SR isn't that far out... But I digress...

OTOH, after testing this 6600k all the doom&gloom about AMD's demise seems like BS as well! After playing with this Richland it's pretty obvious to me that AMD is back on the right track and, once again there are better things to come!

Anyway... The point of this post was my testing of the A8-6600k that I purchased to replace an old PHII 550 server...
I'm no GURU, but here are some results from me trying to find a balance between performance and power consumption using this settup.

These tests were done with aftermarket air (Xigmatech Thors Hammer/single fan) at ~77f.
First, lets go with my current stable clocks:


I'm not sure why this chip has such a low IMC clock (stock/1800), or why the stock CPU VID is so high (1.425)...
Maybe, they're just trying to be safe, or I just have really good chip... regardless, I managed to lower the V's and raise the IMC quite a bit!

System Power consumption during Prime Blend (at the above clocks) is actually damn good too!


Here's a pic of the rig:


If anyone's curious as to why I went with a 6600 rather the the 6800. It's because this was meant to replace a server/Bedroom HTPC and video performance isn't all that important in this case...

[PatRaceTin]

What is estimate voltage @ 5.0GHz ?

[imamage]

I dunno why ,
since Socket FM1 board ( I mostly toy around ASUS , sometimes Biostar)
from cheap A55 board to expensive A75 (ASUS F1A75-M for example) , they tend to detect the VCore as high as
1.4125V , but in reality the chip work just fine at 1.3125V over a year

Things get worse on Mini-ITX board as most ITX board doesn't even have option to lower VCore ( only ASRock have it )

I was hoping the situation improve when FM2 board is out, but this is not the case, the OEM just LOVE to "play safe' and pump the very same high VCore to APUs.

For FM2 , I got ASUS F2A85-M PRO (mATX) , Zotac A55 ITX , MSI A75 ITX board , and a ASRock A85 ITX board
they all LOVE to pump excessive VCore (~1.4125 to 1.42V ) to my APUs , and it's "Default/Auto" setup !

[Daveburt714]

I'm still working on this (even though, in the end I'm looking for low power stability )!

I haven't pushed V's above 1.5, but PatRaceTin asked about 5Ghz, so I thought I'd offer a CPU-z valid @ low(er) volts...
Here's 5.3 Ghz @ 1.46v's:


This definitely won't be my final clocks and I doubt it'd be prime stable, but it shows the potential of this $110 chip...

[ah_khoo]

nice clock sir... have u tried how high can ya pair of gskill go? been tryin hard to get my khx beast to run 2500MHz but no luck so far.

nice imc freq there, may i know what's d Vcpunb?

[tbone8ty]

nice clocks dave!

win another one here!

http://techreport.com/news/25010/ent...a8-6600k-apu#0

[Daveburt714]

nice clock sir... have u tried how high can ya pair of gskill go? been tryin hard to get my khx beast to run 2500MHz but no luck so far.

new imc freq there, may i know what's d Vcpunb?

Best 24/7 I could get out on my RipJawX (stable) was 2225 9-11-10 @1.65v. That was with an 8350/CH5V combo though.
The IMC voltage on the 6600k was 1.250.

Tbone: Thanks for the link.