**[Official]** GIGABYTE Z77X Owners/Discussion/Information/Support Thread and Club (UD5H, UD3H,... - Page 132

Originally Posted by Sin0822 

The deal with some users reporting issues with random BSODs after loading optimized defaults. It is odd issues b/c it doesn't' happen during OCing, I just tell people to RMA their boards if they have odd issues like that. Also some users have said changing the BCLK to 100.00 gets rid of it, so GB is changing the default BCLK to a bit lower than it is now to make up for the 0.1% of CPUs that can't deal with a 0.7% increase in BCLK.

Originally Posted by bobn4burton 

The default/spec for BCLK is 100.0 isn't it? So are you saying that the default BCLK setting that GB has been using is higher than 100.0? If so, what do they set the default to? And are they changing that default value with the different BIOS releases?
If that's the case, I know some CPU's can be very sensitive to a higher BCLK, while others will go to 103+ with no problems. So this would actually make sense if they had bumped this default value trying to get higher default scores...and yet have a decent amount of people with freeze/lock-ups because their particular cpu is sensitive to higher BCLK.

Originally Posted by Snakes 

Well, I believe my Prime95 large FFT test has been failing because my vcore was too low. What I don't understand is that I'd manually set vcore to 1.140 in BIOS and it would run at only 1.056v during testing. I set it back to auto and it now runs right around the default of 1.155 as you'd expect. Maybe adjusting the vcore by offset instead of manually typing in the value would be the way to go, I'll have to look into how to accomplish that. At auto the temps are a little bit high but still reasonable.

Originally Posted by Sin0822 

The deal with some users reporting issues with random BSODs after loading optimized defaults. It is odd issues b/c it doesn't' happen during OCing, I just tell people to RMA their boards if they have odd issues like that. Also some users have said changing the BCLK to 100.00 gets rid of it, so GB is changing the default BCLK to a bit lower than it is now to make up for the 0.1% of CPUs that can't deal with a 0.7% increase in BCLK. I just tell people to manually set their vcore, change their BCLK, or OC. OC is the odd one, the issue doesn't' happen when OCed it seems. Thus GB is lowering the default BCLK back to closer to 100. GB did it that way to help stock benchmark scores, they only did it because some boards, i will not list the boards, will take all the CPU Core ratios to top instead of just a single core, thus making the other boards look bad. Now you can't see this or set this, but it will happen on a stock default load up, so GB isn't going to do that so they increase the BCLK slightly so the top clocks are a bit higher. The issues is that no one states what they can and should do, and they will do what they can as to not hurt stability. Some board makers will have the top turbo multiplier be one greater than Intel spec as well. All games, and caused by reviewers who don't want to compare correctly. Running a benchmark with turbo in itself wont give you very consistent results either.
However freezing in the BIOS is a result of a unstable OC.
I would not worry about the BSOD issue, a lot of users are not reporting it.
I would get the X.

Which code?
I actually didn't do that on purpose lol, it is supposed to be enabled.
Now to Kitguru's sad attempt to get hits. LOL I mean if you don't know what kind of caps those are you really shouldn't ever comment on them like you know what is going on!!! haha it is too funny.
First of all the GIGABYTE Z77 boards they have NO ISSUEs OCing at ALL. The GB boards surprise under LN2 too in pure frequency(seriously take it for a spin, many LN2 OCers use them), IMO bring it on, these boards have no issues with CPU OC like someone stated they heard a few pages back.
As to the lack of multi-layer ceramic (MLCC) capacitors on the back of the board, there is a reason for why some board makers use are both types of capacitors, that reason is that part of their VRM needs the use of high frequency ceramics which have low ESR, while other VRMs don't because the ESR and speed of their VRMs are good enough without them. The capacitors in general act as large energy storage, in normal operation there are huge swings in current, and the rate at which that current can be provided to the CPU as it changes load states is extremely important. That is why there is an output capacitor bank. It can sink or supply power as needed. However the caps have to charge and they have to discharge, so in general two types of capacitors are used, ceramics which are extremely small are good at operating at high frequencies and have very low capacitance. Electrolytic ones are those that look like cans, they also have higher ESR as well as high capacitance ratings, they generally also don't operate at high frequencies.
ESR=effective series resistance
ESL=effective series inductance
Basically just like MOSFETs(MOSFETs can be thought of as having capacitance and resistance), capacitors have parasitics which are characteristics one wants to minimize in the capacitor. Inductance and resistance are two very big ones, which hinder a capacitors performance. Higher ESL and ESR usually isn't what you want. To get over this manufacturers use a mix of the ceramic and aluminum electrolytic. Sometimes manufacturers will use tantalum to replace some ceramics and electrolytic. It all depends on the rating of the capacitors, but really the price of the capacitors they want to use and the amount of them. If you pay more you can get lower ESR and ESL capacitors, if you pay less you will need to use other methods to prevent performance issues. The most commen method is to use a mix of caps, however GB chose to flip the bill and just use their electrolytic, why? Because simply they can as their caps have very low ESR.
Now when a VRM works each phase provides some current and all that current is provided to the CPU as a whole, each phase however operates while the others are off, and so in a 6 phase VRM each phases is on for 1/6 of the time, however they switch on and off hundreds of thousands of times per second. In normal operation the inductors and capacitors work together to provide steady power to the CPU as the power is provided to the inductors, and the inductors to the cpu and capacitors. However if there is a big load, the VRM isn't prepared and will have to depend on the energy stored in the capacitors to take over the entire load while the rest of the VRM can respond. Also after a large event the VRM has to recover and depends on the capacitors to sink the extra energy. Basically the capacitor bank is slow and thus limits the system transient performance. Thus the characteristics of the capacitors used needs to be taken into consideration. A mix of capacitors needs to be used to make sure a few things happen. #1 that the output voltage doesn't deviate from what the designers want during a large load, #2 so that the power can be provided quickly enough #3 as much power as is needed can be provided and there are more things, but those are the three important ones for this conversation. The ESL and ESR are the two characteristics of the capacitors that affect their performance. On the ceramics one must minimize their ESL so that they can properly support the output voltage and make sure it doesn't deviate, and the electrolytics needs to have their ESR low enough to minimize the voltage drop from increased current. However you don't needs to use any electrolytics, take for example some volterra VRMs, they have little to no can type capacitors. You will also have boards on the other which will will have extremely well speced aluminum electrolytic which are in fact good enough to work without ceramics, such as these boards. These GB electrolytics have better ESR than the majority of tantalum capacitors, as so someone told me.
Perhaps ASRock is angry that they don't have any digital PWm supplier and are thus trying to make it seems like GB sucks lol.
FYI ceramics are cheap as hell, GB would put them on if they are needed. However it looks like they are not.
I think fasty explained it well enough.

Originally Posted by mandrix 

Try changing the vcore LLC to different settings so you can see the effect of droop and how to counter it with the "right" LLC setting. Some droop is not a bad thing, but obviously if you can't run stable then you need to make changes. The LLC on these boards is fabulous IMO and finally on Gigabyte boards we have both DVID & LLC working together (if you want to go that route).
As an exercise I proved that you can actually run a +0.00 DVID & get a stable x45 OC just tweaking the vcore LLC. However this didn't allow me to reach my lowest possible vcore. Check out LLC with the different settings and monitor the vcore under load with each and you'll see that it's a handy tool.
Or maybe I misunderstand you completely in which case disregard.

Originally Posted by KillrBuckeye 

I've noticed that CPUID HWMonitor seems to be reporting VTT instead of VCore on the UD5H. This had me quite confused as I was monitoring this value trying to figure out what the various VCore options in the BIOS were doing. CPU-Z seems to report VCore correctly, though.
If I'm trying to reduce my temperatures, what is good starting point for PLL voltage? It's 1.8V on "Auto", so should I start with something like 1.7V? Also, is there any reason to play with the CPU Vtt LLC, or should I just stick with the VCore LLC for now?