Sign in to follow this  
z999z3mystorys

My adventures in Overclocking!

Recommended Posts

I'm going to be OCing a i7 4770k with a cooler master V8 GTS air cooler, and a Asus Maximum VI formula Motherboard.

 

I'll be using Prime95 to stress test it, and Realtemp to monitor temps.

 

Goals:

A. Find a "24/7" OC that runs at or below 67c under prime95 for the first two sets of tests (about 30 minutes) without errors

B. Find a "benchmarking" OC that runs at or below 85c under prime95 for the first two sets of tests (about 30 minutes) without errors

 

 

 

 

First attempt:

CPU voltage: set too 1.15v (1.52 in CPU-Z)

CPU clock 100mhz X 40, 4.0 GHZ

 

Results: no errors, no warnings, max temp was 71c

 

a bit hotter than I wanted, so I'm going to back off on the voltage a bit for the next attempt.

 

Second attempt:

CPU voltage: set too 1.125v (1.120 in CPU-Z 1.136 under load in CPU-Z)

CPU clock 100mhz X 40, 4.0 GHZ

 

Results: no errors, no warnings, max temp was 69c

 

still a bit hotter than I wanted, going to decrease the voltage some more

 

Third attempt:

CPU voltage: set too 1.1125v (1.120 in CPU-Z)

CPU clock 100mhz X 40, 4.0 GHZ

 

Results: got an error from Prime95 saying it couldn't detect all the threads correctly, not sure if that's related to the voltage or not, as it ran for awhile without any errors showing up in the stress testing.

 

Will continue later...

Share this post


Link to post
Share on other sites
world    84

If I could make a suggestion, I think a 24/7 OC would be more useful if you tested stability with prime95 for a longer period of time. Some people insist a 24 hour period, but I think a couple hours is probably good enough

 

Good luck, have fun! thumbsup.gif

Share this post


Link to post
Share on other sites

Yeah, this is just to find it, not the final testing. I think 30 minutes should be enough to find something that has a good chance of being stable, then do the final test for a much longer period.

Share this post


Link to post
Share on other sites
knud    120

You could also try playing around with the memory to see if that makes any difference.

Share this post


Link to post
Share on other sites

I'm changing my objectives a bit, I'm going to see how low a voltage I can get at stock speeds, with the XMP profile for my memory (at 1600mhz for that)

I upgraded to Prime95 version 28.1 that has an update for my CPU type, so the results and the way it's tested have changed a bit

 

 

 

Attempting 100x39 for 3.9 GHZ at 1.1v (1.104 in CPUZ)

 

Results: 3 sets of tests, no errors or warnings, max temp was 66c

Going to attempt 1.05v next... (note that 1.05 is the lowest I can set it too)

 

 

Attempting 100x39 for 3.9 GHZ at 1.05v (1.056 in CPUZ)

 

Result: error followed by BSOD shortly after starting prime95

I guess minimum voltage allowed by the bios is a bit too low :)

 

 

Attempting 100x39 for 3.9 GHZ at 1.075v (1.072 in CPUZ)

 

Result: error followed by BSOD shortly after starting prime95

Voltage seems to still be a bit too low.

 

Attempting 100x39 for 3.9 GHZ at 1.1v (1.104 in CPUZ) (going to put it through more and heavier tests)

 

Results:

A. Small FFTs test, 15 tests, 33 minutes, no errors or warnings.

B. In-place large FFTs tests: ~30 minutes, no errors or warnings

C. Blend test: ~30 minutes, no errors or warnings

 

At this point I'd put it through a long term stability test, if I was interested in running it at stock but undervolted. I'm not however.

 

 

Attempting 100x42 for 4.2 GHZ at 1.15v (1.152 in CPUZ)

 

Results: error after less than 1 minute, stopped it then.

Need moar volts!

 

Attempting 100x42 for 4.2 GHZ at 1.175v (1.184 in CPUZ)

 

Results: error after about 7 minutes, stopped it then.

Need extra moar volts!

 

 

Attempting 100x42 for 4.2 GHZ at 1.2v (1.200in CPUZ)

 

Results: Blend for 1.5 hours, no errors or warnings

Seems to be stable enough, but I did get to 90c, going to try it out with normal use for a bit.

Share this post


Link to post
Share on other sites
RKDxpress    0

Is 100x39 at 1.1v stock for you if not what is? also could you run something like 3dmark 11 for a benchmark at stock then when you get your 24/7 goal run it again? I could compare to that as a gamer. Thanks RKDxpress.

Share this post


Link to post
Share on other sites

100x39 @ 1.2v is stock, the 100x39 @1.1 is stock speed, but undervolted. I'll run 3dmark to see what the difference in score is between 3.9 and my current 4.2 OC

 

here's a link to my 3dmark scores:

http://www.3dmark.com/compare/fs/1566484/fs/1566429

 

the physics score jumped by about 1000 points, but the rest was pretty much unchanged.

 

basically give my current CPU and GPU, the CPU is so much more powerful, even at stock speeds, that OCing doesn't do much. If on the other hand I had say 3 GTX 780 ti cards in my system, I'm guessing the CPU overclock would have a lot bigger difference. as far as gaming goes.

Share this post


Link to post
Share on other sites
Sachin    0

I hope your adventure won't ruin your motherboard. :P

All the best for next adventures. ;)

I'm also thinking to do the same with i5. :D

 

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
Sign in to follow this