Asus P6T Deluxe Discussion Thread

[CookiePuss]

You guys that are running 4ghz, try using multiplier of 19 x 210, and you can probably drop your vcore down about .05 - .1.

I dont know if its just me, but it takes me around 1.375vcore to run 200x20 stable, but 1.275 to run 211x19 stable. LinX 20 loop stable, with 12 gigs takes about 30 mins a run.

[Nocturno]

I´ve been testing this mobo and found out this about C-STATE:

System: Asus P6T Dlx - i7 920@3700Mhz 1.16v - 3GB GSkill@1560Mhz 9-8-8-21 / CPU is Watercooled




C-STATE OFF







C-STATE ON [AUTO]








C-STATE Conclusions:

ON:
Larger memory bandwith but less L3 Cache
Less Power consumption
CPU is cooler about 3º Celsious
Less CPU performance in some aplications. in video encoding the performance is the same.

Tested with Tomb Raider and the FPS are equal with our without C-STATE

[tjelaw]

I´ve been testing this mobo and found out this about C-STATE:

System: Asus P6T Dlx - i7 920@3700Mhz 1.16v - 3GB GSkill@1560Mhz 9-8-8-21 / CPU is Watercooled

C-STATE Conclusions:

ON:
Larger memory bandwith but less L3 Cache
Less Power consumption
CPU is cooler about 3º Celsious
Less CPU performance in some aplications. in video encoding the performance is the same.

Tested with Tomb Raider and the FPS are equal with our without C-STATE

Nice find!!

I was comparing some bench results of my own, I also used the same Everest memory bench. And I had to same differences. Lower DDR3 bandwith / higher L3 cache...

I couldn't find out what caused it (I tested at same clock speeds) It was driving me crazy!! But now im very glad I know what it is. Thnx!

Maybe when clocking higher than 4.2 and the multiplier begins to throttle etc, this could help.

[speckled10]

I´ve been testing this mobo and found out this about C-STATE:

System: Asus P6T Dlx - i7 920@3700Mhz 1.16v - 3GB GSkill@1560Mhz 9-8-8-21 / CPU is Watercooled

C-STATE Conclusions:

ON:
Larger memory bandwith but less L3 Cache
Less Power consumption
CPU is cooler about 3º Celsious
Less CPU performance in some aplications. in video encoding the performance is the same.

Tested with Tomb Raider and the FPS are equal with our without C-STATE

Brilliant find nocturno !!!! U = genius

I too was wondering why the hell my L3 Cache was so much slower than other similar benchies .. havent tested yet but I'm sure this will be it ..

I noticed your m/b temps there are 41C .. you got any cooling on the Northbridge ?? That little ASUS sidefan thing ?

I was getting mid-30's Idle .. but its gone up 2-3C since I put dust filters on the HAF intake fans ..

[speckled10]

OK Confirmed !

C-State On


C-State Off


I also went a little further and tested with various C-State limits

C-State C6 Limit
As above 'C-State On'

C-State C3 Limit


C-State C1 Limit


C-State C1E Limit


So basically this proves its the actual C6 State (lowest power) is causing the hit to L3 Cache speed .. interestingly while the CPU is in C6 the L3 Cache is the only thing left *on* .. see here ..

http://www.bit-tech.net/hardware/200...tecture-dive/3

Either way its obviously got something to do with it ..

I have no clue tho about the strange memory speed variations .. max with C-State On .. min with C-State Off or C3 limit .. moderate with C1 or C1E limit On

Ran some other benchmarks but CPU scores didnt seem to change much .. so I'm still sticking with C-State On full and enjoy those nice chilly Idles

[Nocturno]

At least someone liked it!