I guess everyone has their own priorities. HT serves absolutely no purpose for me so it stays off. I'm spending every moment of free time playing Fallout 3 and it's DEFINITELY not necessary for that. ;)
Printable View
I guess everyone has their own priorities. HT serves absolutely no purpose for me so it stays off. I'm spending every moment of free time playing Fallout 3 and it's DEFINITELY not necessary for that. ;)
Pisklink, nice results. What's your cpu batch #?
Batchnumber is 3836A875
Thanks for the overview of voltages. It seems I don't need more Vcore with current settings because if you look at the screenshot I already was benching on 4Ghz @ 1.35v. Only if I would put on HT I think I will need more voltage. I will try to put CPU PLL lower and see if it's still stable.Quote:
a couple ideas for you.
i think the max cpu pll should be no greater than 1.89 (according to intel)
you may need more than 1.35 vcore for 4 ghz, I needed 1.4v
Why is it so quiet in here?
Is the max. voltage for the CPU 1.55v according to Intel specs in the overview above?
I still don't get what PLL is for.
I was priming @ 4Ghz, 1.35 vQPI and 1.88 PLL which are the maximum safe values set by Intel.
Then i decided to put PLL back at 1.80 and it kept priming for 6 hours until I had to stop.
weird
I keep seeing how some of you have better stability when raising CPU vtt compared to core voltage.
Well raised CPU VTT from 1.25v to 1.30v and at same vcore went from no stability at all @ 4.1 Ghz, to at least running 3dmark, Will test stability further over the next few days, see how it behaves :)
http://img.techpowerup.org/090204/Capture029.jpg
I finally got ahold of a kill a watt and measured my OC power numbers =)
Very interesting to say...and system not using anywhere near what I thought it would.
My major hardware is listed in my signature.
Here are my readings 3822mhz @ 1.27v 182x21 VTT/QPI@ 1.25v 3095mhz .. all other voltages on auto:
Computer off but plugged into wall: 2watts
POST to windows an average of 160-190watts
Windows idle:160-175watts
WCG number crunching: 290-300watts
WCG+3dmark06: 370-395watts
3dmark06 only: 260-290watts
Prime95 blend: 307watts
CoreDamage: 325watts
Now the next point is...the PSU effeciency is rated @ 80+. I am not even using its full load capacity meaning its more likely less effecient then that. So, assuming 75% effeciency...is it correct to take 25% off the above numbers and that number is the true power draw of my PC? (obviously I know the socket sees the full numbers listed above, but Im talkin internal usage after the effeciency of the PSU is taken into affect...)
Hey huys! I tried to overclocked my core i7-920 and the only stable until now is 21x166 in 1.128V! I run LinX and it passed and the max temps are 65-75. I ordered a few days ago 4x1 OCZ GOLD edition pc3-14400 1800mhz and i will try to overclock when they arrive. My problem is if i use 3x1gb of OCZ is it going to work trible channel?
http://img516.imageshack.us/img516/7767/linxbu9.jpg
I'm not sure but I always thought the efficiency raises when not using the full potential of a PSU. So if a PSU is rated for 800W it will provide e.g. 85% efficiency at that max capacity. If less power is drawn the efficiency will raise because all the PSU components are less stressed.
@ All
Is it correct that 1.55v is the max. allowed vCore according to Intel specs?
I don't get it, what's the speed on my computer? :rofl: I guess, at stock speed, and idle cpu one core is overclocked, and the rest are taking a break? So tweaker reports the multi of the cores that are on the break, and CPU-Z- the single working core. Is that a reasonable assumption?
Thanks for your reply. Luckily I'm living in an iglo on the Antartic so cooling shouldn't be a problem. :D
Nah, my chip doesn't get that hot really. It's watercooled by a D-Tek Fuzion version 1 with 5,5mm nozzle and at 4080Mhz it gets around 48 - 50 degrees celcius in Realtemp.
I wan't talking about realtemp, it's the discrepancy b/n the tweaker and cpu-z I was talking about.
1.65v as stated by Intel
Well, the jury is still out on this, intel says - some where (I do not know where) not to exceed 1.65v but then there's the intel electrical spec - see below - not sure what to believe. I try to stay below 1.65, a lot of folks are going by the rule of keeping memory volts within 0.5v of qpi/vtt. So if your qpi/vtt is 1.35v, you want to keep your memory volts to leas than or equal to 1.85v. I know of no one that had fried their cpu (yet).
anzial: RealTemp reports your average CPU multiplier. Depending on your C1E / EIST bios settings and how the Windows power options are set up (Minimum processor state), you might see the multi jumping up and down at idle. I left RealTemp like this because it best shows you what's going on. If EIST and your power options are set up properly, at idle, you should have a consistent multi of 12.0 on Core i7. If you don't want your multi dropping at idle then you need to make sure SpeedStep is off and Minimum processor state is set to 100%.
Sometimes CPU-Z does some averaging and reports the multiplier at its full value but the displayed MHz will be based on some sort of average. Your screen shot is a good example. CPU-Z shows 133.3 X 22.0 which doesn't equal 2907.9 MHz which is displayed.
rge did a lot of Core i7 testing for me so I think you should be able to trust RealTemp MHz as well as temperatures. Here's the most recent release:
http://www.xtremesystems.org/forums/...postcount=3166
RealTemp 2.70 does not properly support Core i7 so don't use that version.
what are the max temps for the 920? is 75c too hot on core 1-2 and 69 on 3-4 running 2 instances (8 threads) of p95? running 194x21 turbo on.
ok cool im not so bad then.