Post your C3 Test Results! (new title)

[Zeus]

Mike, how much did you need to shove through your ram to get 960? Mine didn't even boot with 1.74V at 960 7-7-7, 940 does though.

[neo mike]

Mike, how much did you need to shove through your ram to get 960? Mine didn't even boot with 1.74V at 960 7-7-7, 940 does though.

Hey Eugene i was running about 1.81 to get to 1920mgz, i did get to 1930mgz but i couldnt run 1m its would just BSOD on me. My board has a overclock dial on it, so i set ram to 1860, boot into windows and then turn up the dial on board till i reach my rams speed i want.

Its weird when i hit 1921 mgz i cant run 1m, but 1920 will run fine , but im going to get more out of it. Im thinking i need to tweak drive strenghts, cause i cant up anymore voltages, there is nothing left but drive strenghts. I think i need some better ram, the 2250 GTS Corsairs, to have any chance of going much higher, i dont know.

[chew*]

It was 8m stable 1950 was sucide

You know the drill...

Screens or it didn't happen.

Wouldn't surprise me in the least zeus, MSI has been tuning for artificial clocks that don't stack up perfromance wise with recent bios's......

980 is already done though

1.81 is alot of volts for corsairs......I usually need 1.75 max to do those speeds and thats 3 diff sets......

[neo mike]

What do u mean artificial clocks ?, maybe i dont need 1.81, i just figured to pump some juice into it to help keep it stable. So your saying the MSI board is subpar in performance in ? 1m 32m ?.

[shaolin95]

Latest results for my C3.
4.1Ghz on Windows 7 64bits.

[chew*]

What do u mean artificial clocks ?, maybe i dont need 1.81, i just figured to pump some juice into it to help keep it stable. So your saying the MSI board is subpar in performance in ? 1m 32m ?.

Preety much. Msi lately has been tuning for "screenshots" but the performance per clock is not there.

Older bios's like 13b6 and 14b2 are faster clock per clock.

Honestly the only thing it has a use for for me is 3dm03 and 4x gpu's.

I like the layout they use but not the direction they have taken.

[neo mike]

Im just trying different boards out, i admit i like the UD5P better than this MSI, and should have never got rid of it, but i got this board half price. I have noticed that the performance isnt there clock for clock with other post im seeing with same settings.

I dont see it as artificial clock, you cant boot into windows at 1900 mgz ram, so you have to use something to tweak inside windows to up the speed, so whats the difference what it is, unless im missing something and you can boot into windows at 1900 or more.

EDIT, i will run 1.36b and compare clocks with 1.84b, i was running 1.36b before, but had lockups alot, random freezes.

[beta]

maybe they're trying to make the board a little less incompatible with ddr3 modules by sacrificing performance

[chew*]

maybe they're trying to make the board a little less incompatible with ddr3 modules by sacrificing performance

Thats actually exactly what they are doing, which is why certain modules that won't boot at 1600 c7 divider magically work at rated on MSI with recent bios's....It's not a patch to bios to fix compatibility with XX Ic's like normal, it's more them killing performance just to bandaid the issues.

[neo mike]

This is 13b6 bios MSI, i guess i would have to run 32m to see a clock for clock difference ? cause i dont see it here.

[chew*]

All you guys use 1m to measure ram performance..........when ram plays hardly any roll in the result

Your going to have to play with alot more than just 32m, run a variety of stuff that is bandwidth heavy......more importantly what matters is the ram that can not do XX timings on older bios and then can in later.

Also you need to compare to a diff board.....at those speeds thats like 16.234 1m on other boards.

69K effeciency is really below average.

Here's 69k slightly more effecient than yours on asus......i'm not even close to your NB speeds and horrible timings etc theres room for quite alot of improvement.

[neo mike]

So what are you saying that the MSI sucks in the efficency dept ? from your numbers and mine it sure looks like its lacking in the efficency area

[shaolin95]

I seem to have a lot of problems trying to max my memory...I can do 1333 at 6-6-6-24-28 1T but if I try to OC more it just does not go...prime errors.
Are you guys running 2T to push it harder and also disabling interleave?

[chew*]

So what are you saying that the MSI sucks in the efficency dept ? from your numbers and mine it sure looks like its lacking in the efficency area

Yep which is why i said it's clocks are artificial......sure it might set high clock records. Other boards need far less to match it in performance though.

[neo mike]

Ahhhh ok now i got you, thanks for that information, from what i have seen on here i would say the Asus Crosshair 3 is the most overall efficent, for us air guys, or darn close to it.

[FlawleZ]

So what are you saying that the MSI sucks in the efficency dept ? from your numbers and mine it sure looks like its lacking in the efficency area

Yes and no. Chew* has been praising this board from the start as well as point out that this board doesn't have the same efficiency as other boards do with the same settings. If you read through his original testing in the GD70 thread you'll see what I'm talking about. The performance will alter slightly with different BIOSes but overall won't quite match the Gigabyte or Asus.

[shaolin95]

So it is sort of what DFI used to do...usually you could get very high FSB with DFI but at looser settings.

[chew*]

Yep preety much there is a right tool for every job, MSI is great board for users and has certain strengths for benching as it does weaknesses. So do all the other boards.

As always there is a right tool for every job.

Will you actually see this diff as a 24/7 user? I doubt it. I just which they would take a diff approach to the memory issue.

Look at it this way , If i did all my memory reviews on that board and all users of other motherboards couldn't attain same results/timings......they would be ticked

[FlawleZ]

Yep which is why i said it's clocks are artificial......sure it might set high clock records. Other boards need far less to match it in performance though.

I have a feeling chew* that the differences in performance we see clock v. clock is less efficient microcode for the DDR3 RAM. Like it was previously mentioned MSI likely updated the code to better support different types of modules yet in doing so has a negative hit on performance because they altered the code for support and not performance.

[chew*]

I have a feeling chew* that the differences in performance we see clock v. clock is less efficient microcode for the DDR3 RAM. Like it was previously mentioned MSI likely updated the code to better support different types of modules yet in doing so has a negative hit on performance because they altered the code for support and not performance.

It could play one way or the other......the flip side of that is they can claim there board does rated specs and other board manufacuturers boards just suck and can't support ram when in reality thats not whats going on here. Personally I would want the most performance with the least effort.

Could also be the pressure they get from endusers with there little screw up of claiming 2133 DDR 3 right on the board.

[shaolin95]

chew, are you disabling the memory interleave to get higher mem oc? I cannot seem to get much from my RAM no matter what. I used 2T and got it stable now a 1480, it would not pass prime95 at 1T at those speeds with the current timings 6-6-6-24-28

[neo mike]

Right Chew, they did claim 2133 ddr3, that sort of got me on the MSI to begin with, and we all know 2133 is a line of xoxokdjf, total false advertising.

[chew*]

chew, are you disabling the memory interleave to get higher mem oc? I cannot seem to get much from my RAM no matter what. I used 2T and got it stable now a 1480, it would not pass prime95 at 1T at those speeds with the current timings 6-6-6-24-28

When running actual benches I never back off performance settings to achieve a result. That usually defeats the purpose....if that was the case I wouldn't be running 90 ns on 2gig sticks

[shaolin95]

So I guess my memory is just lame.
I can RMA since it is not able to do the spec 1666 7-7-7-21 but I guess is not worth the hassle.

[chew*]

So I guess my memory is just lame.
I can RMA since it is not able to do the spec 1666 7-7-7-21 but I guess is not worth the hassle.

spec for the OCZ's is not 7-7-7-21 @ 1666....... its just 7-7-7 1600 and 2t is part of spec.

They should do 1600 7-7-7-24-33 2T at rated volts.