Jump to content
Corsair Community

Overclocking results for CMZ8GX3M2A1600C8


dark_skeleton

Recommended Posts

Hi all,

Just wanted to share some overclocking results I did lately on CMZ8GX3M2A1600C8. That means 8GB of RAM, 2 sets of 4GB each.

 

After every change, memory was tested with Memtestx86+ v4.20 for errors and for those that passed the test, Prime95 and SuperPI were launched.

Just to mention, I started with lowering the voltage and was pleasantly surprised that I could run them with just 1.35V @ X.M.P. (I didn't test lower voltage).

Now, to the real OC.

I wanted to achieve the fastest speeds supported by by MB, CPU and RAM. Earlier I OC'd my CPU a little and that resulted in 103MHz BCLK (instead of standard 100MHz). As we know, BCLK in SandyBridge-motherboards is kind of a taboo, because you can only change it within a very tight range so I left it as it was after auto-tuning of my MoBo.

Because of that OC, the speeds I achieved aren't standard ones.

 

First thing that I learned was: you can't achieve 2472MHz. At all. Since max safe RAM voltage is 1.65V and I didn't feel like increasing VCCIO, my system wouldn't boot even with timings as loose as 15-15-15-40 (you can't loosen them more than that). [EDIT: Even after increasing VCCIO to 1,25V and DRAM voltage to 1,725V, it still wouldn't boot).

 

I moved to testing @2197MHz then (since I didn't want to modify my BCLK, I could only use RAM freq multipliers).

I started by loosening timings until I could start my PC at the default RAM voltage (1.5V) and ended up being able to boot and pass memtest86 at 11-11-11-33. Now that's something, I can boot up @ nearly 2200MHz (600MHz more than X.M.P., nearly 40% frequency OC!) using default voltage!

 

Here's where the real OC testing started. Since I didn't like the timings at default voltage @ 2200MHz, I wanted to push them as tight as possible. The thing is, sometimes a whole test run can pass in memtest86, but after testing it again, some errors may show up. After one pass I thought I could run my memory at 10-11-10-33 @ 1.55V, but I was wrong. I went back to testing and the best result of that is, by far:

 

2198MHz @ 10-11-10-24-T2 @ 1.65V

 

Pretty neat improvement over stock 1600MHz :)

 

Confirmed stability with 3 full passes of memtest86+. However, these timings are giving me rounding and SQRT errors in SuperPI and Prime95 even after increasing voltage to 1,75V and VCCIO to 1,25V. There are no problems at all during PC usage though (gaming), except for occasional bluescreens or game crashes that happen maybe once a week.

 

It seems I can run them at 11-11-11-33, 10-11-10-30 and 10-11-10-24 @2198MHz at whatever RAM/VCCIO voltage I want and they boot, but they hardly ever pass a SuperPI compute or a few minutes of Prime95 test. I think I'll have to look into secondary timings or try to loosen some of them even further...

 

Please share your opinions and what speeds, what timings, at what voltages you managed to achieve :)

 

Here's the spreadsheet of OC settings I have tried. [will keep updating it]

http://dl.dropbox.com/u/5650076/OC%20RAM.png

 

Next step: Getting stable with 2200MHz and trying to achieve 7-7-7 timings :)

Link to comment
Share on other sites

  • 3 months later...

Thanks mate :)

Ah, that's what I have read too. I guess I'll just have to experiment or I'll try to go down to 7-7-7 first ;)

I found some definitions for those other timings and even found a thread where people are clearly saying that some of them are giving nice performance gains, but then again they were experimenting on 7-8-7 2200MHz sticks ;)

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.

×
×
  • Create New...