OCCT is in beta fo i7 support and is currently not supporting the platform 100%

RealTemp 2.9 is also beta however it is the only readout right now that is trustworthy for i7.. but it is STILL beta as well

63 is nothing to i7. They run hotter than Q's of the past however you do need to make sure you installed that heatsink correctly and it is mounted well. They can be a bit tricky due to their size and at one time there was some issues around the top bracket and a small washer that sits between the center of the bracket and the top of the base plate missing. I thought they corrected that some time ago though.

At this point lets assume it is mounted correctly

Of course you may have to make manual memory multipler and timing changes to ensure the memory is running at the correct speed and timing as you increase BCLOCK however since you purchased the 920 you may not hit 1600 exactly as it will depend on the BIOS math

This is overclocking and its not 'plug-n-go'


TJMax is 100c on i7 and you want to be 20 under that at least

I would say you are OK in RealTemp under max load up to 80c but no more.

If you find you are seeing high 70's with CPU speed stable at 3.4-3.6, something is not right assuming you are not pushing Vcore too high

QPI/DRAM voltage and DRAM Voltage are 2 different things.. do not mix those up. DRAM will be max of 1.66 and QPI/DRAM will most likely sit at 1.35 or a bit above

never mind on the memory timings.. I just tested it and confirmed it with CPUz. the frequency is indeed 1600 and the timings are as should be.



the only thing I have to worry about now are the temperatures and getting a stable overclock... I'm on my way to test the CPU @3.4ghz for temps



thanks again

Make SURE you disable TURBO MODE and also HYPERTHREADING in the BIOS along with the CPU TM Function. FSX will not take any advantages with HYPERTHREAD and that can drop your temps as well

Hyperthread is needed for applications like Video Encoding and other engineering software that is hyperthread aware. If you do not use such software disable hyperthread in the BIOS and you may get the drop in temps you need

Andy I take it that fan is the one that came with the unit... funny because mine is a 4pin, 4wire PWM fan and not a 3 wire

The 3wire fan.. if I am seeing that image right, can not be controlled from a 4 pin PWM plug and it must be running 100% at all times unless the BIOS is set to control it through the hardware monitor.

Am I correct in that assumption?


I cant tell from the image exactly.. it looks like a Rd-Wht-Blk wiring and is missing the blue wire PWM fans require


As I said earlier.. OCCT is being updated for i7 and right now I can not place confidence in it as test solution. If it PRIMED clean then I would assume OCCT may be having some problems however OCCT is also checking memory, not just processor cores


Have you run MEMTEST?

http://www.memtest.org/#downiso

I think I would make that my first test to run

5 passes (9 tests per pass) probably take about 1-2 hours or so.

If you fail that, its either memory settings, memory voltage, vcore, or defective memory product

You should always start from the ground and work your way up ... in that we dont slam it to the wall .. you need to verify your components one at a time to make sure you are not dealing with defective product or settings off-base

So if you pass MEMTEST with the system running default CPU and memory speed and timing on the right settings, then you move up and test it again, and then again

You need to find out if you are unstable in DOS directly to the memory subsystem be it a setting or defect. Typically a defect will show up without clocking... and since eveything in the area of memory remains a constant as you increase the CPU speed the only value you may need to work around that is QPI/DRAM and Vcore


If you find problems with memory stability with little or no clock then shoot over to the Mushkin forum and let them know what you find. Your best bet in locking that down is directly with them.

LOL  


no, Nick  


this is the memory I'm using now.. I should have cleared this up with you earlier! sorry about that!

http://www.newegg.com/Product/Product.aspx?Item=N82E16820227381



Nick! I shut off all those Intel TM crap and the TurboV and whatnot, and the temps shot down, but still not low enough to run even 3.8ghz cool enough for our liking here... at 3.4ghz they'll play around at 67-68 hottest core and at 3.8 they went up to 82 but took a while to get there with PRIME95... I'm not comfortable with that though... however I'm quite relieved that the temps did shoot down as much as they did! the memory at 3.8 was running at 1523 I believe with the 7-7-7-20 timing but I haven't ran any memory tests. will do that tonight though!



as for the fan on the HS.. it is indeed the one that came with the HS. it's a 3 pin power connector (the connector on the motherboard has 4.. I also thought that was a little weird but I thought what the heck) I'm not sure if I've seen any settings for the fan on the BIOS but I'll double check that tonight after work. do you recommend I get a different fan than that to throw on the HS?

Nick,


finally some good news... although it seems that this 920 won't be able to reach 4ghz, I just tested for a very stable cpu OC to 3.7ghz. again, OCCT failed it quite quickly with temps going over 73, but I ran PRIME for well over 1 hour and check out the results:











what do you think?

all I have to do now is test the memory which seems to be working fine but you never know..

I'll get back after the memtest is done.


thanks again man!  

Andy

Andy

your QPI DRAM is only 1.275. Typically that takes 1.35 to get to 4GHz but that also depends on the memory too

is hitting 4GHz a pure temp issue or stability?

The OCZ sticks may do 1800 but you will most likely need to raise the timing to 8-8-8-24. Thats problematical and something one must work out. You can not raise the DRAM Voltage but you could raise QPI/DRAM if 8-8-8 1800 was not stable

I doubt you would get them over 1800