i5-750 stock vcore overclocking

S

spotpuff

Gawd
Joined
Aug 15, 2001
Messages
522
Just wondering what people have gotten their core i5-750 (or i7-860 if you prefer) up to on stock voltage.

Mine has a stock vcore of 1.125v (as indicated in the BIOS) and I did ~80 runs of Linx at 3.5Ghz (167x21) and everything was fine.

At idle it clocks down to 9x167 and something like 0.88v (vcore left at "normal" instead of auto for this; seems to work pretty well!)

I know Tom's Hardware clocked their i5-750 at 3.6Ghz on stock voltage of 1.25v, just wondering if anyone else tries this or just goes straight to 4GHz at 1.4v :)

As an aside I was using the H50 and with the default fan full blast I was getting the following temps are reported by RealTemp. Fan was blowing into the case. I replaced this fan with 2 Gentle Typhoons @1450 rpm in push pull pushing out of the case and added a 20cm fan to the side panel. Temp reported is highest of the 4 cores; usually one core is 2C hotter than the rest for some reason.

167x21 @ 1.3vcore - 70C load, stock fan full RPM (~1600 I believe?)
167x21 @ 1.3vcore - 65C load, 2xgentle typhoon 1450 RPM
167x21 @ 1.125vc - 54C load, 2xgentle typhoon 1450 RPM

Nothing new other than voltage has a tremendous effect on temperatures, which [H] has reported on in the past and I am always interested in knowing. I wish more sites did testing on Core clock, voltage, power draw and temps; we've all heard temp increases are not linear, but is it a square relationship for both clock and vcore? Or just one but linear for clock?
 
Last edited:
i ran with stock heatsink at 3.5ghz for about 2 weeks. =D

i was able to run a 3.5ghz clockspeed with 1.175v

i5build_stockheatsink_oc.jpg
 
I can get up to around 3.4 stable with stock voltage with my 750. Running a little lower than that right now due to heat issues in my tiny ass case.
 
Fantasysage said:
I can get up to around 3.4 stable with stock voltage with my 750. Running a little lower than that right now due to heat issues in my tiny ass case.
Click to expand...

Yeah the cooler master HAF 922 with an extra 20cm fan on the side is pretty ludicrous. 700rpm so they're quiet, but not silent. I wouldn't classify the noise as annoying but it's not silent. Unobtrusive? It's a shame they didn't include the extra fan, though, as I had to buy it, and it's made out of some VERY hard resin that self-tapping screws are not a fan of. I would recommend 1.5" long screws (the 6-32 ones looked good). The fans are 30mm thick FYI.

I think 55C load at 3.5 is a pretty good temp; I might push higher later but I'm happy with that as an OC either way especially at stock vcore.

Haste said:
i ran with stock heatsink at 3.5ghz for about 2 weeks. =D

i was able to run a 3.5ghz clockspeed with 1.175v
Click to expand...

Your settings are almost the exact same as mine; I did 167x21 :) Same motherboard too :D How bizarre! Are the 75C temps with the stock cooler? I have my 5850 plugged into the bottom PCIE x16 slot, though, because it blocks PCI slots that I don't use and lets me use the other PCIE without as many issues. x8 is fine for the 5850.
 
spotpuff said:
I think 55C load at 3.5 is a pretty good temp; I might push higher later but I'm happy with that as an OC either way especially at stock vcore.



Your settings are almost the exact same as mine; I did 167x21 :) Same motherboard too :D How bizarre! Are the 75C temps with the stock cooler? I have my 5850 plugged into the bottom PCIE x16 slot, though, because it blocks PCI slots that I don't use and lets me use the other PCIE without as many issues. x8 is fine for the 5850.
Click to expand...

55c is a nice temp for full load.

the screenshot i posted is with the stock cooler.

i loving this motherboard, its rock solid. this is the 2nd gigabyte board(recent s775 build had one too) ive owned during my computing career and am still very impressed with gigabyte motherboards.
 
Haste266 said:
55c is a nice temp for full load.

the screenshot i posted is with the stock cooler.

i loving this motherboard, its rock solid. this is the 2nd gigabyte board(recent s775 build had one too) ive owned during my computing career and am still very impressed with gigabyte motherboards.
Click to expand...

I agree; I went from a DFI lanparty to gigabyte 965p-DS3 to this.

Not sure where DFI is now, but Gigabyte's features are definitely making them stand out.

I am having problems with putting the computer to sleep sometimes, though, and some features like ErP (formerly EuP, some European power standard) are not familiar to me. Also, sometimes the BIOS will do weird stuff like not detect my USB keyboard, even though it had to at boot for me to hit DEL to get into the BIOS.

Still, overall I'm happy, and positioning the video card lower was great for me from a layout perspective. I'm still perplexed as to their choice to make the chip cooler block the top x1 slot, though.
 
I have mine at 3.6 ghz (180x20) with vcore set to 1.1875v, or whatever the exact number is in the BIOS. LLC is off and max voltage as reported in Windows is slightly lower at 1.15v. I haven't had time to run the full suite of stress tests so I may need to bump the vcore slightly higher, but hopefully this will be stable.

Temps: idle at 22 C, load at 55-56 C with a Mega + Noiseblocker.
 
Have mine at 1.101v and 3.5 (175x20), can run burntest for 100 iterations on High with 0 hiccups or issues, maxes out at about 68c with 2 fans on my H50, but I draw air out of my case through the radiator, not in like they would like me to so that prolly has something to do with temps. Run the fans like this because I would have only 1 exhaust fan if I didn't.
 
Loadline calibration doesn't equal stock voltage, fyi.

Anyways temps increase with both vcore and clock speed. The big difference is power draw. Your chip's wattage under load will increase linearly with clock speed but not by that much but with more vcc it really starts to go up. There is a good recent article about this floating around somewhere.

I personally would be more concerned with VTT than VCC with i5/i7.
 
BababooeyHTJ said:
Loadline calibration doesn't equal stock voltage, fyi.
Click to expand...
Were you referring to my post? It looks like technically anything over 1.1v for VCC isn't "stock", though anything under 1.2v is a pretty minimal increase. As for VTT, I still don't really understand it; I just left it at the default (1.1v) but I suppose it may need a bump as well if things aren't stable in high-memory usage situations?
 
A little update on my situation. I realized my board was overvolting the shit out of my CPU and I re-applied TIM.

So, with a vcore of 1.175 @ 3.4ghz I am loading with prime95 at 56C stable after an hour. Ambient temperature is around 22c. Not bad at all for my dinky little cooler. I think I might bump vcore up to ~ 1.2 and see how high I can go.

EDIT: 3.6ghz @ 1.175 volts stable at 56c after 45 minutes of prime95. Looking pretty good.
 
Last edited:
Well, 5 months later (has it really been that long?) I bought 4GB more of RAM and installed it.

I flashed a new BIOS (if it isn't broken...) so had to redo my settings. The new BIOS won't allow for a x21 multiplier at all unless turbo boost is enabled, so I tried 20x175 and the system booted but failed Linx w/ 6gb of RAM and everything at "normal" voltage in the BIOS.

I realized I set my DRAM voltage to 1.5v instead of 1.66v so tried again, still failed. Backed off to 3.4GHz (170x20) and linx seems to be happy now. Having more than double the RAM to run linx seems to make runs take much longer and it's 02:30 so I don't think I'll run 80 runs today but I will try it tomorrow and hopefully everything passes. Not sure if the 100MHz frequency loss is due to the CPU dying, the TIM degrading, the extra 4gb of RAM, or what.

Linx just passed the 2nd run w/ 6103MiB of RAM (problem size 28232) so I'm going to bed and I'll run more tomorrow. The most annoying part of OC tweaking has got to be rebooting and waiting for Windows to boot...
 
my 750 only does 3.2 at stock voltages. Corsair claims you can't even run their 1600 memory without going to 1.3vtt. I found that is pretty much true with my stuff. I imagine that a lot of people have to raise the vtt first to get anywhere. It seems to me that memory plays a huge roll here. The last two I setup with this exact memory, both needed 1.3vtt to run the memory at rated speeds.

VTT is the memory controller voltage. If you left that at stock then you should probably goose it a bit. I'm not sure where RB is right now, but I read some tech notes from Corsair that basically came right out and said that 1600Mhz on thier memory requires 1.3vtt for dual channel and 1.4vtt for triple channel. I realise you don't have corsair memory but it might be worth a shot. It was pretty much confirmed to be the case when helping a friend OC his i7. Anything below 1.4vtt wasn't stable unless we were below 3GHz.

Changing my VCC doesn't seem to help much. I would leave that one alone for now. I think mine might be at 1.125 It needs that little bit to be stable at 4GHz so i just left it.
 
You must log in or register to reply here.
Back
Top