MSI Z68A-GD65 (G3) PCI-Express 3.0 LGA1155

[cadaveca]

yeah, I can try higher multis, no problem. I don't really have sufficient cooling to do, say 2 hours of IBT, but I got enough to take a look, for sure. I still have a 55 chip that I can use as well. It's just that, as you know, 5500 MHz drops a tonne of heat.

what BIOS is on your board?

 

[EarthDog]

I dont recall what I had (original shipping bios is my guess), but what the MSI rep sent me was E7681IMS.N25 and that didnt help. they are looking in to it however.

It just takes a 'blip' to see the throttling so no need for anything but good air.

This is the Juen bios that supposedly works (have not tried yet): E7681IMS.M30

These was a not go: E7681IMS.M56 / E7681IMS.1D0

 

[leftyman]

Heavy throttling

Hello:


I have this mb with a 2500k and a noctua NH-d14 and the multi lowers from
48x to 41x with every bios except the initial release one (N00). I hope I dont ever
need a newer bios beacause I wood loose half my overclocking potential....


best regards

 

[cadaveca]

Hello:


I have this mb with a 2500k and a noctua NH-d14 and the multi lowers from
48x to 41x with every bios except the initial release one (N00). I hope I dont ever
need a newer bios beacause I wood loose half my overclocking potential....


best regards

That is due to Turbo Current and Wattage limits. If they are too low, the CPU will automatically lower multi to stay within those limits. Same applies to all boards when overclocking using Turbo mode, and you'd more than likely have the problem no matter what product you used. i did not use the release BIOS for testing, as I always try to use the most recent for any product, and I did not have any issues with throttling, or this would have reflected in the scores.

 

[EarthDog]

He can try setting the wattage/current to 250w and see if that helps. Or try non turbo overclocking. It didnt help me, or many others though. MSI is working on a solution.

Cad, your 45x multi with the nice low voltage (compared to say the 50x and 1.48v I use) is at least part of the reason why the issue doesn't show up (proper bios that doesnt have these limits is another). NOte, this user was trying 48x multi. I would imagine once he makes those adjustments to the wattage he may run into the same thing others are (47x max multi under load).

Lefty... can you email me the bios? I will pm you my info. Thanks!

 

[cadaveca]

He can try setting the wattage/current to 250w and see if that helps. Or try non turbo overclocking. It didnt help me, or many others though. MSI is working on a solution.

Cad, your 45x multi with the nice low voltage (compared to say the 50x and 1.48v I use) is at least part of the reason why the issue doesn't show up (proper bios that doesnt have these limits is another). NOte, this user was trying 48x multi. I would imagine once he makes those adjustments to the wattage he may run into the same thing others are (47x max multi under load).

Lefty... can you email me the bios? I will pm you my info. Thanks!

I totally forgot about you mentioning this issue. Funny thing is, I haven't seen it on my board. Interesting. I did have this board and did the review before public release though, so who knows whats up. I hope it's not the same issue keeping him @ 41x...that'd really suck.

 

[EarthDog]

It was your low multi and voltage that does it (or doesnt do it )...and MAYBE a 'good' bios. I can rock and roll all day long at 45x. Its when Im above 47x that it throttles back to 47... IF I set the wattage/current to their max of 250/255.

 

[cadaveca]

It was your low multi and voltage that does it (or doesnt do it )...and MAYBE a 'good' bios. I can rock and roll all day long at 45x. Its when Im above 47x that it throttles back to 47... IF I set the wattage/current to their max of 250/255.

I do not see the same, and have tried my two CPUs. My 55x chip is not so "low voltage", needing 1.35v for 4.5.

could be BIOS, I got mine direct from MSI with information that the same BIOS would be flashed to all shipping boards.

Low and behold, this BIOS is what works. My version number is not the same, however.



I was hoping he'd try the limit increases, and report back before just up and saying it was the BIOS...he could potentially get 48x without issue with how you've described the problem.

 

[heky]

It is not a problem of setting the power limits. It is a problem with the latest bios versions from MSI. I have the same problem, throttling under linx, ibt, etc. with anything higher than 4.5, and i can set the power limits to 250W(which btw this chip can never achieve). On my board, with bioses up to 1.C, there are no problems, but 1.D,1.E, and all the betas after 1.C have this issue. MSI really should get their act together. I am using a 6 months old bios becouse of that.

 

[EarthDog]

Its several bios versions. You guys *may* have got lucky with the bios and the fact that you never pushed past 45x. Again, it doesnt happen at that 'low' multi is why you didnt see it. With any bios I can sit at 45x under load with power limits set. Im betting good money says that if you tried 50x, cad, it would throttle...unless you have a good bios. But again, we know you didnt see it because you havent pushed to where it begins to show this behavior.

Heky, can you email me the bios version you have? YHPM.

 

[cadaveca]

we know you didnt see it because you havent pushed to where it begins to show this behavior.

That's a pretty large assumption. And I'll leave it at that. You asked me to test @ 50x, and I did, and I did not have the problem. I am also using a BIOS from prior to release, that you may not have.

It is not a problem of setting the power limits.

There is more to power settings that just current and wattage limits, there is also a setting in BIOS that probably isn't open to the end user that can fix this. It should be automatically enabled once power limits are increased, not sure why it isn't.

MSI simply needs to make this option open to the end user, as ECS has.

You all are talking like i don't know what's up...do you even know what setting I am refering to?

 

[EarthDog]

Sure its an assumption. You didnt tell me via PM, nor mention it in this thread that you tested that multi! Not to mention in this thread you said, "your 55x chip needs 1.37 @ 45x", with nary a mention that you went past that. How am I supposed to know you did that?? Im a reviewer, not a mind reader!

With that, I will PM you my info and if its no trouble, please send me your bios, or LMK which one it is on their website and I will get it.

Thanks!

You all are talking like i don't know what's up...do you even know what setting I am refering to?

Overspeed Protection? But that needs disabled, so thats not it...?

Note we have mentioned that information 'like you dont know whats up' becuase we had no idea, since you didnt tell anyone, that you tested it. Please accept my apologies for that, but know you brought that on yourself for not sharing your testing.

Please be more helpful to the public you voluntarily serve instead of being cryptic about this bios setting that you apparently have that we may not? I can also share this information with MSI to hopefully speed up the problem resolution.

 

[cadaveca]

you can rest assurred that from me personally, when I say I'm gonna do something, it gets done.

I did say in my post up a few that i tried BOTH CPUs without the problem. Since we discussed the problem only happens at higher multis, and you requested I try 50x, that is EXACTLY what I was referring to in the post above.


I mention the other CPU so let you know it's not really something CPU specifc, as neither my high VID nor my low VID chip have the problem.


Clearly, since there may be an issue, and my board does not have it currently, I'm not about to flash a BIOS that might cause the problem.

In BIOS is a setting for MAXIMUM TDP. This is NOT overspeed protection...it's a literal switch that raises the power consumption limits, which seem to be something not specific to Turbo, like the wattage and current limits are.

Second option in this screenshot. This option is DIFFERENT than the adjustable Current setting(default 97a, up to 200a). Adjusting the variable current limit that some BIOSes have(and this board has as well, but only adjustable via XTU) will not increase the max multi available under high load without this setting being adjusted as well.


This is pretty basic stuff, IMHO, so if MSI disabled the option, it must have been on purpose. I cannot aruge against MSI doing so, as it may save a few CPUs.


Things like this are why i tell users trying to OC to the limit, say under sub-zero, to always start with the original release BIOS, no matter what board you are using. Typically BIOS updates adjust things because stability problems or such have been noticed, so it's only natural that at times, BIOSes will limit clocking abilities.

 

[EarthDog]

Yeah sorry, couldnt pull that out of any of your posts. Thanks for taking the time to test and show your results.

Im not asking you to flash to a different bios, Im asking you to send me yours. You have a PM with my email.

I *think* I have IA Core limit in there and tried that..... I will have to check. This board has been collecting dust for almost 2 weeks now.

 

[cadaveca]

This option isn't available to the end user in the MSI BIOS that I have seen, unfortunately. MSI merely needs to expose it.


I couldn't find anything else that may lead to this issue, and the cause and behavior are pretty specific, which helps eliminate the probable causes. It can ONLY be that option, IMHO, causing the problem.

Myself, I'm more interested in why you need to update the BIOS in the first place.

 

[EarthDog]

Me personally? I tried this on the bios that came with it (E7681IMS v23.2B1) first and experienced the throttling. And the one he gave me (E7681IMS.N25) did not work as well. I will try yours (E7681IMS N21) I just received and report back. If that resolves it, I believe we can confidently say that something in that bios that isnt in the others did it.

You will have to give me some time though as I am moving Friday and packing...

 

[heky]

@EarthDog
The latest beta bios fixes the throttling issue for your board.

http://forum-en.msi.com/index.php?topic=151148.50

 

[EarthDog]

Confirmed that N34 and N35 fixed this issue. Thanks for the PM Heky!!!

 

[orangec]

Hi,

I have an ASUS P8Z68 V-Pro board, and as many owners, I have a problem with awakening from sleep mode , and booting up , if Internal PCC overvoltage is set to Auto or Enabled, so disabling is the "cure" for this issue ,

so my question is the "Internal PLL overvoltage" present in Msi z68a-gd65 (g3) BIOS, and has any of you experienced, issue mentioned above,

tnx

 

[cadaveca]

Most boards have this issue.


Personally, I do not think anyone should be really using PLL overvoltage, so no, I have not checked. Do you really require it?

 

[orangec]

Do you really require it?

Actually, no , I was just curios, why something strongly related to the overclocking needs to be disabled, on a board which is ment to be a fine oveclocker. Contrary to my initial belief, this issue si present with Z 68 boards , regardless of vendor. Anyway, one of the solutions is setting Ai Overclocker Tuner to [X.M.P] under Ai Tweaker Menu, and BCLK to 100 . More complex way to fix multiple boot issue : http://blog.tabinda.net/microhard/p67-dualdouble-triple-boot-issue/

I hope someone will find this usefull, cheers

 

[cadaveca]

Actually, no , I was just curios, why something strongly related to the overclocking needs to be disabled, on a board which is ment to be a fine oveclocker. Contrary to my initial belief, this issue si present with Z 68 boards , regardless of vendor. Anyway, one of the solutions is setting Ai Overclocker Tuner to [X.M.P] under Ai Tweaker Menu, and BCLK to 100 . More complex way to fix multiple boot issue : http://blog.tabinda.net/microhard/p67-dualdouble-triple-boot-issue/

I hope someone will find this usefull, cheers

I understand your opinion on that, however, that is kinda of how it works when you overclock to the level that requires PLL Overvoltage - some caveats must be made. There are many boards that do not even have that option at all, perhaps due to the issues it causes.

 

[orangec]

Well, I will have to get into all this. Next step is to see how my Kingston can do at 1600 MHz, at lower voltages, then 1, 65 V, which is automatically set by the [X.M.P] profile. Great forum you got there, cadaveca, glad I found it

 

[ColdPlay]

amd + pcie 3.0?

NIce review!
I would def bee looking for pcie 3.0 on my next board/build!
Would be nice to find an amd board that supports it.
Does anyone know when that will be?
I think the fx 8150 would do nice paired with pcie 3.0!

Thx Cold

 

[SonDa5]

I'm on my second Z68 GD65 G3. First one was defective and I'm getting ready to install OS on 2nd.

First I want to make sure I have a good BIOS.

What BIOS is the best stable BIOS for full over clocking features?