Regarding voltage - please do be careful, the SB chips have LOWER max voltages than the older i7s. There's a lot of people on the overclocking sites destroying their chips in the first few days here by thinking they could just crank it up to the old values from their 920s and 860s.Here's the quote from a guy who runs an enthusiast PC shop in Europe about it:You could do 1.4+ on the vcore and 1.65+ on the memory on the older chips and it was safe - not the case here.

Too right. Thats why I've stopped at 4.6Ghz. I suspect I could squeeze out another 200Mhz at around 1.4V but I cant see the benefit. 5Ghz would definitely be possible with this cooling solution but risky on the voltage front.I've limited Vddr to 1.5V as a consequence of the above. Nothing has been tweaked that affects the Southbridge. BCLK is steady at 100Mhz.

So basically, if you own an i7 which is doing 4.2Ghz+ (like mine or others), you have no reason for changing to SB?Quick-exchanging cards would be a sure test for many simmers here, incl. me. You can even let your son use the GTX580 for a while :Big Grin:And come on, dismantling system?? Disconnect - unscrew - pull - push - screw-in - connect power. I bet you I could do it under a minute :Praying:

Too right. Thats why I've stopped at 4.6Ghz. I suspect I could squeeze out another 200Mhz at around 1.4V but I cant see the benefit. 5Ghz would definitely be possible with this cooling solution but risky on the voltage front.I've limited Vddr to 1.5V as a consequence of the above. Nothing has been tweaked that affects the Southbridge. BCLK is steady at 100Mhz.

What CPU cooling solution do you use and what CPU temp do get when stress-testing (OCCT)?

Nice! I stay tuned for your result with your old card...

Thanks for your effects, but until you test with your old GPU results are not useful, IMO. The higher fps/smoothness could be due solely to the new GPU and not SB.

HiYou said: "I transferred the CORSAIR XMS memory from the X58 machine (with the voltage lowered to 1.5V at the moment to minimise the risk to the 2600K. It now runs at 1333MHz"Are you actually saying that you are using triple channel DDR3 RAM on a new system designed for DUAL channel DDR3 RAM or you were using DUAL channel DDR3 RAM on an i7 920 cpu and X58 mobo designed to run TRIPLE channel DDR3 RAM and they use RAM in 3, 6 or 12GB permutations. If the later is the case it may be no wonder you are getting a faster speed with the new rig! :Thinking: RegardsPeterH

HiYou said: "I transferred the CORSAIR XMS memory from the X58 machine (with the voltage lowered to 1.5V at the moment to minimise the risk to the 2600K. It now runs at 1333MHz"Are you actually saying that you are using triple channel DDR3 RAM on a new system designed for DUAL channel DDR3 RAM or you were using DUAL channel DDR3 RAM on an i7 920 cpu and X58 mobo designed to run TRIPLE channel DDR3 RAM and they use RAM in 3, 6 or 12GB permutations. If the later is the case it may be no wonder you are getting a faster speed with the new rig! :Thinking: RegardsPeterH

They are exactly the same modules, they are both DDR3.Dual are paired, triple are...tripled. <_<

What CPU cooling solution do you use and what CPU temp do get when stress-testing (OCCT)?

Cooling solution: http://www.xbitlabs.com/articles/coolers/display/thermalright-archon.html although it is a tight fit due to the height.Prime 95 peak temperature 66C at 4.6 Ghz.

HiYou said: "I transferred the CORSAIR XMS memory from the X58 machine (with the voltage lowered to 1.5V at the moment to minimise the risk to the 2600K. It now runs at 1333MHz"Are you actually saying that you are using triple channel DDR3 RAM on a new system designed for DUAL channel DDR3 RAM or you were using DUAL channel DDR3 RAM on an i7 920 cpu and X58 mobo designed to run TRIPLE channel DDR3 RAM and they use RAM in 3, 6 or 12GB permutations. If the later is the case it may be no wonder you are getting a faster speed with the new rig! :Thinking: RegardsPeterH

I originally had 6 sticks of 2Gb RAM. It left two sticks in the x58 machine and one channel empty. It is designed to do this as long as you put them in the correct slots.The other 4 sticks went into the 2600K machine - two in each channel.

Cooling solution: http://www.xbitlabs....ght-archon.html although it is a tight fit due to the height.Prime 95 peak temperature 66C at 4.6 Ghz.

Thanks for the information I've heard that Prime 95 isn't that well suited for i7 processors and that you should use OCCT instead when stress testing a system (source Nick Needham). Any opinions on that?

Hi AllWith respect to FSX the (fairly subjective) results (on my standard test flight; PMDG 747 in VC, UK2000 Gatwick Approach, Ultimate Traffic 2, GEX, UTX, REX and using an unchanged FSX.cfg) at stock speed it performs similarly or slightly better than my over-clocked (3.78Ghz,12Gb RAM) i7 920. It was noticeably smoother, probably due to the GTX580. Frame rates mid 20's dipping to around 13-15 occasionally. The i7 used to dip into the teens for prolonged periods. With a comfortable over-clock of 4.6 Ghz, frame rates are now in the low 30's to high 20's and hardly ever going below 20. Taxying at Gatwick would have been around 15 and is now in the low 20's.

Those are impressive results! I run Ultimate Traffic II with a mix of WOAI traffic bgls at 100%, what percentage AI do you enable?

Hi David"They are exactly the same modules, they are both DDR3.Dual are paired, triple are...tripled."Sorry I was under the impression that they reverted to single channel or didn't perform optimally in this type of configuration! :Thinking: It makes you wonder why manufacturers bother to make dual channel DDR3 RAM?Thanks for the correction! RegardsPeterH

It makes you wonder why manufacturers bother to make dual channel DDR3 RAM?Thanks for the correction!

Sales!!! :--)(You're welcome Peter)

Thanks for the information I've heard that Prime 95 isn't that well suited for i7 processors and that you should use OCCT instead when stress testing a system (source Nick Needham). Any opinions on that?

I suspect Nick's comments relate to the first generation i7 and not Sandy Bridge. I read somewhere else that the Stress test program does not matter so much on Sandy Bridge. At the end of the day the objective is to run FSX and not blow the chip up. FSX does not generate anywhere near as much heat and is stable. I expect the chip temperatures will be higher in summer. Ambient was about 20C fot the test. Should also say that the temperature was measured in a closed case and not on a bench.

Those are impressive results! I run Ultimate Traffic II with a mix of WOAI traffic bgls at 100%, what percentage AI do you enable?

AI is set to zero but it was also zero on the X58 machine.

Thanks DescendDescend. Much appreciated