• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.
  • The forums have been upgraded with support for dark mode. By default it will follow the setting on your system/browser. You may override it by scrolling to the end of the page and clicking the gears icon.

TPU's Core i7 Overclocking and Feedback

+1 for fits.

a fridge works by keeping the cold in and if you had a object putting off heat then it wouldnt cool at all.
 
:DIm just messing about, keep yer fuur on lol


Speaking of cooling: Coolermaster Z600R any good?
 
:DIm just messing about, keep yer fuur on lol


Speaking of cooling: Coolermaster Z600R any good?

I hear good things. Unless you're getting some kind of deal you might check out the Scythe Mugen 2. For the price your getting a lot of cooler and performance. Outside of those, the only other option I would recommend would be either the Prolimatech Megahalems or Noctua NH-D14. Though, the Thermalright Venomous X is getting a lot of good mention lately by those using it. Though those last three are close to each other after you factor in fan cost, there really isn't any reason to get much else, from a performance perspective. Yea you jump a lot in price, but all the others in between the first group and second are not really worth the price increase since performance for noise will be more.

As for me, I initially had the OCZ Vendetta 2. I used 3 different fans on it, even my 3000 rpm Kaza, until I settled for a while on the CM R4. Still, I couldn't push vcore much past 1.275v with LLC on without temps getting into the high 80s. I finally convinced myself to drop some real money when the Noctua NH-D14 come out and I can tell you that I was not disappointed. I saw a full 15C drop, with the same cpu overclock, after AS5 cured. The real shocking thing to me though is how quite it is compared to the OCZ setup while still being cooler. I can run 3.6Ghz (21x181) core and uncore passive with no other help that the stock HAF fans! Ambient room temps were in the 22-24C area. I was able to make 10 loops of LinX using almost all memory and temps never got over 80C. It took about 30 minutes to finish and HT was on too. 1.425v on the core with vdroop is not problem with Noctua fans at full speed, all while stilling being quiet. Temperatures never get over mid 80Cs.

When it comes to cooling an i7, if your going to overclock, go all the way or nothing at all. The huge gains seen from the expensive coolers are worth every penny. Even if they equal similar performance temperature wise to the cheaper coolers, they will usually be a lot quieter doing it. I know not everyone can just coin up 60-90 USD, but if you can save up the money, it is worth waiting.
 
I hear good things. Unless you're getting some kind of deal you might check out the Scythe Mugen 2. For the price your getting a lot of cooler and performance. Outside of those, the only other option I would recommend would be either the Prolimatech Megahalems or Noctua NH-D14. Though, the Thermalright Venomous X is getting a lot of good mention lately by those using it. Though those last three are close to each other after you factor in fan cost, there really isn't any reason to get much else, from a performance perspective. Yea you jump a lot in price, but all the others in between the first group and second are not really worth the price increase since performance for noise will be more.

As for me, I initially had the OCZ Vendetta 2. I used 3 different fans on it, even my 3000 rpm Kaza, until I settled for a while on the CM R4. Still, I couldn't push vcore much past 1.275v with LLC on without temps getting into the high 80s. I finally convinced myself to drop some real money when the Noctua NH-D14 come out and I can tell you that I was not disappointed. I saw a full 15C drop, with the same cpu overclock, after AS5 cured. The real shocking thing to me though is how quite it is compared to the OCZ setup while still being cooler. I can run 3.6Ghz (21x181) core and uncore passive with no other help that the stock HAF fans! Ambient room temps were in the 22-24C area. I was able to make 10 loops of LinX using almost all memory and temps never got over 80C. It took about 30 minutes to finish and HT was on too. 1.425v on the core with vdroop is not problem with Noctua fans at full speed, all while stilling being quiet. Temperatures never get over mid 80Cs.

When it comes to cooling an i7, if your going to overclock, go all the way or nothing at all. The huge gains seen from the expensive coolers are worth every penny. Even if they equal similar performance temperature wise to the cheaper coolers, they will usually be a lot quieter doing it. I know not everyone can just coin up 60-90 USD, but if you can save up the money, it is worth waiting.

If you look at these pictures, the Venomous X is very similar to the Megahalem.

http://en.expreview.com/2009/12/10/...-heatsink-pictured-in-all-its-glory/6063.html
 
attachment.php
 
If you look at these pictures, the Venomous X is very similar to the Megahalem.

http://en.expreview.com/2009/12/10/...-heatsink-pictured-in-all-its-glory/6063.html

There is a thread over on OCForums about seating pressure, ways of taking ambient temps, and how this all correlates to differences between cooler to cooler and even from seat to seat with the same cooler.

JoeC, an admin over there, made a very interesting comment that when he has done cooler reviews he takes his ambient reading from right in front of the intake fan and in several locations. He stated that he has seen coolers that circulate the air locally and can cause the fan ambient to rise 5C while stressing. This is something that would never be picked up by reading ambient temps for the room and can severely skew results even though ambient appears to be constant throughout testing.

OCForum thread

I know the discussion is about the thermal paste and being able to get an average expected result given lots of data input from different sources. BUT!, he still makes some very interesting comments about other variables that can skew review results that are not initially apparent to most. RGE makes some comments too later in the thread that is worth a read.

Also, read SPCR's newest heatsink test results. They make some interesting comments in there as well that might be worth taking into consideration when looking at a roundup of heatsinks from any given website.
 
Psh! Westmere was sooo '10. Looks like they already have A1 Sandy Bridge working and running W7.

11.jpg

12.jpg
 
OK another question - DRAM voltage

I have the option od 1.64 or 1.66. I am using 1.66V now and runing the ram at 1800MHz.

Is 1.66V safe even though it has this message in the bios? Or should i just stick with 1.64.


The XMP profile puts it to 1.66V and it puts the QPI voltage to 1.4 (QPI too high)
 
my corsair dominator 1600mhz CL8 DDR3 ram cannot pass 4000 uncore with 2000mhz using a DFI DK X58. I using different setting on memory timing and voltage but its still failed to reach 2000mhz and 4000 uncore. anyone here succeed to pass the 4000 uncore using the same board and the same ram?
 
OK another question - DRAM voltage

I have the option od 1.64 or 1.66. I am using 1.66V now and runing the ram at 1800MHz.

Is 1.66V safe even though it has this message in the bios? Or should i just stick with 1.64.


The XMP profile puts it to 1.66V and it puts the QPI voltage to 1.4 (QPI too high)

DRAM voltage must be at most X+.5V where X=QPI voltage. My QPI is running at 1.45V right now to support 2000MHz memory.
 
So at 1.66V, the QPI must be atleas 1.16V in order to keep it safe, is that correct?
 
So at 1.66V, the QPI must be atleas 1.16V in order to keep it safe, is that correct?

Yep, you're getting the algebra :D
 
OK another question - DRAM voltage

I have the option od 1.64 or 1.66. I am using 1.66V now and runing the ram at 1800MHz.

Is 1.66V safe even though it has this message in the bios? Or should i just stick with 1.64.


The XMP profile puts it to 1.66V and it puts the QPI voltage to 1.4 (QPI too high)

Perfectly safe bro.
 
OK, now im at 190 x 19, ram at 1520 on 7-8-7-20-1T.

Interesting thing i found, that even if i downclock the ram to like 1333 @ 9-9-9-24, BCLK at 133 for default cpu speed, if i put the UCLK to like 3420 or whatever it is, it wont boot with the QPI voltage at 1.2V (default). 1.25V is ok, but i havent tested whats minimum in order to boot. Is that normal then? I've put it on 1.4V, like XMP suggests and i booted at 1800MHz on 9-9-9-24 no problem, which is nice.
 
OK, now im at 190 x 19, ram at 1520 on 7-8-7-20-1T.

Interesting thing i found, that even if i downclock the ram to like 1333 @ 9-9-9-24, BCLK at 133 for default cpu speed, if i put the UCLK to like 3420 or whatever it is, it wont boot with the QPI voltage at 1.2V (default). 1.25V is ok, but i havent tested whats minimum in order to boot. Is that normal then? I've put it on 1.4V, like XMP suggests and i booted at 1800MHz on 9-9-9-24 no problem, which is nice.

Well remember you are running the uncore approximately over 1Ghz quicker than stock. That's what you meant by UCLK right?
 
The thing that needs to be atleast double the ram speed - Uncore Clock Frequency or whatever it is
 
The thing that needs to be atleast double the ram speed - Uncore Clock Frequency or whatever it is

yeah, the default multiplier is 16x. so 133*16 is 2128 MHz You said you were running it at stock clocks but at over 3Ghz, that's why you need the additional voltage.`

However, that doesn't have to be twice the RAM speed, I remember running it at defaults, maybe I don't recall correctly???
 
Well if i leave the CPU at its stock speed, the ram goes at 1600MHz which needs atleast 3200Mhz on the Uncore. Next step is 3333MHz, then the one following is 34XX. at 34xx it just wont boot if i have the QPI/Uncore voltage at 1.2 (which is stock). I was just asking if thats normal or not :)
 
Well if i leave the CPU at its stock speed, the ram goes at 1600MHz which needs atleast 3200Mhz on the Uncore. Next step is 3333MHz, then the one following is 34XX. at 34xx it just wont boot if i have the QPI/Uncore voltage at 1.2 (which is stock). I was just asking if thats normal or not :)

yeah it is normal in the way of voltage, but I don't know if that has to be twice the speed. I'm just not sure about that.
 
Thats what it says in the bios and it doesnt allow it to be set lower than 2 x ram speed. Well, atleast on my board
 
Thats what it says in the bios and it doesnt allow it to be set lower than 2 x ram speed. Well, atleast on my board

maybe because I ran my RAM really low most of the time so it didn't affect me :laugh:
 
Back
Top