Coolermaster V8 test results - new install

[Jamin43]

I put in my Coolermaster V8 last night - figured I'd share the results with the board - in case anybody was looking at buying one.

My CPU with stock cooler was running a little hot as my CPU was an earlier batch.

Core i7 920 C0 Stepping at Stock speed

Intel OEM CPU cooler
Cores idle 48-52 ( was 39-45 degrees before I broke one of the fasteners - D'OH )
Prime 95 79-82 ( on my original build test )

Coolermaster v8
Cores idle 33-40
Prime 95 58-66

* Idle, I got a bout around 6-8 degree drop from my original install temps
* Full load - I get around 16-19 degree drop when all 4 cores are cranking at 100 % ( which I don't think I'll ever hit as an end user ).

For the overclockers out there - how much do you think I could safely overclock ( considering temp as limiting factor ) with my current setup. I was someday hoping to go around 3.2 - would that be pushing the envelope with this CPU / Cooler combo - or should this be able to handle it. After owning this CPU - still love the thing - but I can appreciate the cooler running of the D0 stepping a bit more - with heat being a limiting factor if I decide to overclock.

How do these temperature drops compare with others who have this cooler - or even another brand of cooler running on a higher wattage quad core.

It was a bit tight - to screw in the board - in the Antec 300 - but doable. And the V8 Clearance was far enough away in the first set of triple channel RAM to avoid the Memory heatsinks. If you wanted to use all 6 DIMMs - the second set of memory would probably need a lower profile for the second 3 DIMMs. The V8 is quieter than my case fans - so I don't notice the sound at all.

Edit: If you buy / install a V8, Make sure you can read the " COOLERMASTER " top to bottom on the top of heatsink - when you view in your case. Reading a few videos - I was under the impression it was tough to decide air flow direction - so overthunk it following the arrow on the plastic protector tab as my guide. Had to install it twice to get it right. There is no mention of this in the directions - but overall - it's a pretty simple process. Universal backplate for all CPU's - and a separate mounting bracket for AMD and Intel. The screws on the mounting bracket are repositionable when you depress them to move from 775 to 1366 to 1156. Pretty simple stuff.

Thanks

 

[linkin]

congrats on getting that cooler. i want one, i guess that's why we have christmas

My CPU's stock speed is 2.2ghz, right now it's on 3.02ghz, at 1.4v. I'm pretty sure you could hit 3.2 at least, probably around 3.4 or 3.6 with the right voltage, regardless of the different CPU. remember people have got these around 4ghz already.

I guess it's just the usual of upping the FSB (QPI or whatever intel calls it on i7) until you get instability, then it's on to the voltage. If you can, set your overclock to unlinked so your RAM won't get overclocked with the CPU, so you can do it seperately.

 

[Jamin43]

My CPU's stock speed is 2.2ghz, right now it's on 3.02ghz, at 1.4v. I'm pretty sure you could hit 3.2 at least, probably around 3.4 or 3.6 with the right voltage, regardless of the different CPU. remember people have got these around 4ghz already.

Nice, my specific limitation concerns are regarding the CPU heat more than anything. What kindof ceiling will the heat give me in terms of where I should set my expectations at how high I can reasonably expect to achieve - with my current CPU / heatsink temps I listed above as a starting point.

The 4+ ghz are mostly D0 step which runs a few degrees cooler out of the box. But I have zero experience on OC - so I could be wrong on that one maybe the c0 step can run that fast too. I'm naive on the topic til I jump in and start to tweak myself.

I guess it's just the usual of upping the FSB (QPI or whatever intel calls it on i7) until you get instability, then it's on to the voltage. If you can, set your overclock to unlinked so your RAM won't get overclocked with the CPU, so you can do it seperately.

If / when I decide to overclock - that's good to know.

Thanks

 

[Jet]

The 4+ ghz are mostly D0 step which runs a few degrees cooler out of the box. But I have zero experience on OC - so I could be wrong on that one maybe the c0 step can run that fast too. I'm naive on the topic til I jump in and start to tweak myself.

http://forums.techpowerup.com/showthread.php?p=1642991

http://www.xtremesystems.org/forums/showthread.php?t=205688

http://www.overclock.net/intel-cpus/538439-guide-overclocking-core-i7-920-4-a.html

Check some of those out. Having just gotten my i7 920 D0, I have been overclocking the last few days. From what THERMAL_REACTOR has said, the C0s overclock very decently as well--he has most of the processor speed records on CF (my motherboard is limiting me, I cannot go above 4530Mhz or so, while he can reach 4.6-4.7). The C0s just require more voltage. Depending on your motherboard, you should be able to reach 3.5 very easily (I run 3.8Ghz on stock everything); most i7 920s can reach 3.8, but then after that it's harder and harder as you need to do fruther tweaking, etc. The jump from 3.8 to 4.0 required a processor voltage increase from 1.2250ish to 1.2750ish...still well under the 1.35V Intel spec, though.

Play around with it--as long as you read those links up top and know what the voltage specs are for your adjustments, you should be able to get a very decent overclock. For overclock testing, 80C isn't terrible--the i7 TDP is 100C instead of the 85C-95C of other processors. For general use, though, I'd keep it under 70-75C (all fans on low, my 4Ghz is right around there).

 

[funkysnair]

im running 3.6ghz on my co stepping, reached it easy peasy, but to go further i have not tried....

i can only take so many blue screens

 

[Jet]

im running 3.6ghz on my co stepping, reached it easy peasy, but to go further i have not tried....

i can only take so many blue screens

You know you been overclocking too much the last few days that your heart stops whenever the mouse pointer freezes for a split second....and then you realize you're not on the computer that you are overclocking.

 

[CdnAudiophile]

Every processor is different and each board is different so it's hard to give exact numbers but here is a good reference to start with. When I overclock I always change all the voltages. I never set them to auto because that can create instabilities in your OC's. I see you have 1600 ram so a 200bclk would be a nice round number for a ratio to get 1600 on the ram. Now if you want 3.6ghz you would go with an 18 multi. These voltage should get you running at that speed.

Vcore (cpu) 1.35v
Qpi VVt (cpu/memory) 1.30v
dram (ram) 1.65
IOH (NB) 1.2
ICH (SB) 1.2
CPU PLL 1.88

Now after using something like prime95 and keeping a close eye on temperatures you can go from there. If heat is a concern try less CPU voltage and test for stability again. If your temps are ok try maybe a 19 multi. Try more voltage if it's not stable and keep an eye on the temps. Just keep repeating that until you find a balance of speed/heat. You can also play with the bclk but that will result in your ram being OC'd a bit or under clocked. My C0 ran very cool it just needed alot of volts to run. For 4GHZ my C0 needed 1.44v (CPU) with a 1.35 QPI with all the other setting I suggested above.

 

[Jamin43]

Every processor is different and each board is different so it's hard to give exact numbers but here is a good reference to start with. When I overclock I always change all the voltages. I never set them to auto because that can create instabilities in your OC's. I see you have 1600 ram so a 200bclk would be a nice round number for a ratio to get 1600 on the ram. Now if you want 3.6ghz you would go with an 18 multi. These voltage should get you running at that speed.

Vcore (cpu) 1.35v
Qpi VVt (cpu/memory) 1.30v
dram (ram) 1.65
IOH (NB) 1.2
ICH (SB) 1.2
CPU PLL 1.88

Now after using something like prime95 and keeping a close eye on temperatures you can go from there. If heat is a concern try less CPU voltage and test for stability again..

So I'm gathering from your post - it's a finesse that you look at the variables and can tweak to compensate for heat / stability / and other things in a more or less trial and error approach til you reach what you feel to be an optimal point in terms of performance / temp / stability? If it gets too hot - there are variables that I can compensate for the heat - while still increasing speed of processor.

So it sounds like there is really no rule of thump in terms of X increase in processor speed ( or percentage increase ) = Y degrees temperature increase - I'm oversimplifying things - and in reality Overclocking doesn't work that way?

Thanks

 

[Jet]

So I'm gathering from your post - it's a finesse that you look at the variables and can tweak to compensate for heat / stability / and other things in a more or less trial and error approach til you reach what you feel to be an optimal point in terms of performance / temp / stability? If it gets too hot - there are variables that I can compensate for the heat - while still increasing speed of processor.

So it sounds like there is really no rule of thump in terms of X increase in processor speed ( or percentage increase ) = Y degrees temperature increase - I'm oversimplifying things - and in reality Overclocking doesn't work that way?

Thanks

The only true information you know is that there is some relation between watts = so much heat and that voltage increases watts exponentially while frequency increases watts linearly. You could theoretically estimate how many watts your processor is going to use at X Vcore and Y frequency, but in practice it's much easier to just mess with actual numbers and real-world observations.

 

[CdnAudiophile]

So I'm gathering from your post - it's a finesse that you look at the variables and can tweak to compensate for heat / stability / and other things in a more or less trial and error approach til you reach what you feel to be an optimal point in terms of performance / temp / stability? If it gets too hot - there are variables that I can compensate for the heat - while still increasing speed of processor.

So it sounds like there is really no rule of thump in terms of X increase in processor speed ( or percentage increase ) = Y degrees temperature increase - I'm oversimplifying things - and in reality Overclocking doesn't work that way?

Thanks

Exactly, 2 indentical computers with exact same hardware down to the exact same batch cpu are not going to produce the exact same results. Each component introduces it's own variance that will effect the outcome.