Official GIGABYTE Forum
Questions about GIGABYTE products => Motherboards with Intel processors => Topic started by: No457 Snowy on October 04, 2013, 08:18:41 am
-
I am another of the multitude of people who have built a system based on the Z87-D3HP motherboard to then find their new PC is very unstable or won't even boot when using the 4 DIMM slots. I have read many examples of this problem all over the net and there appears to be no fix coming from Gigabyte as yet. Every thread on this issue in various forums is filled with reports of different BIOS revisions that have not helped and none of those revisions apparently address this 4 DIMM issue as a documented target for fixing. As something to potentially affect so many customers you would think Gigabyte would be a little more pro-active in addressing the issue. Not only that but once the issue is revealed to the system builder It is extremely inconvenient to address by the poor retailer. I have read a large number of reports where the boards have been RMA'd by the retailer and the builder has chosen a different brand of motherboard to get around the issue and having to rebuild the PC again. This amounts to a massive waste of people's time.
So what's the word from Gigabyte on this issue? Will it even be fixed?
-
I'd suggest you change the title, as this problem is not exclusive to the D3HP. Many people with other Z87 boards are also having problems when using all 4 ram slots.
I have a Z87X-UD5H (BIOS F7). When I use 4 DIMM, I have guaranteed BSODs. If I use less than 4 sticks, the system is stable, no matter what sticks I choose or the slots I populate. No memtest86 errors or any other stress test issues at all. Just random BSODs.
All the solutions suggested might or might not work for some people but, honestly, lowering ram frequency, relaxing timings, increasing voltage or disabling C-States are not acceptable solutions, as we paid for a product we're not being able to use as advertised.
Gigabyte, please comment ASAP.
-
Hi thanks for your post, the more people that post about this ridiculous issue the more chance Gigabyte will get off their hands and fix it. In Australia there are laws regarding products that do not function as advertised and if I don't get satisfaction very soon I will be packing this product into it's box and sending it back for a refund. I was hoping for a fix from Gigabyte but I guess they don't care enough.
Unfortunately I tried to change the heading of this thread but I can't change the heading as apparently once some time has passed the option to edit posts is removed.
-
This morning the guys at Gigabyte recommended me to use their beta BIOS F8a. I flashed it this morning and mounted all my RAM in the mobo. That is a G.Skill TridentX 32Gb 4x8Gb kit at 2400Mhz and 1,65v CL10.
I have to say, today has been the first day in the history of this computer that I haven't had a BSOD will all 4 RAM slots in use. I've been working with it for about 12 hours non-stop.
EDIT: FALSE ALARM!!! Right after writing that, I got a bsod. Oh well... BSODS seem to be getting less frequent. I used to have about 4 per day and this one was the first one today. I'll give them one more week before calling this my last Gigabyte purchase. I'm going back ASUS!! The Maximus VI Formula is just... perfect!!
-
Thanks for the update.
It's good to see that they do actually recognise the issue, although in my opinion a bit of transparency and openly acknowledging the issue would go a lot further towards preserving their reputation than trying to keep it hush-hush and pretending it's a couple of isolated cases.
-
I don't see where they recognise the issue. They recommended me to change my BIOS version to see if that helped... don't they know what they add to each bios?. If BIOS F8a just adds support for the 4771, then why would it fix RAM compatibility.
What I want from them is "Yes, we messed up. The DRAM VRMs are rubbish. Here's your fix... or, here's a new MOBO, you don't have to dismantle your current computer until you get the new one... or, here's your money, buy an ASUS"
-
Hi, I specifically registered here to confirm that GA-Z87-D3HP is extremely unstable with 4 memory slots used, which is evident from frequent fatal system crashes, also known as Blue Screen of Death (BSOD).
Here's a discussion with many users experiencing this problem on the Corsair forums:
http://forum.corsair.com/FORUMS/showthread.php?t=119923&page=4
I shared my experience there. In short, I ordered a new PC 3 months ago, but got the aforementioned GA motherboard instead of the Asus I actually paid for. Wouldn't be a problem, if it was working. BSODs were occurring within 20 minutes of booting. After many days of driver and bios updates, OS reinstallation, component swapping and etc. I realized one simple thing - no crashes with 2 memory sticks (tested for 7 days). But as soon as all 4 Kingston modules were installed, BSOD within 20 minutes. The memory was thoroughly tested and replaced to rule it out. Finally, when I came across many other people describing similar experiences with the same motherboard, I had no more doubts - GA-Z87-D3HP rev 1.0 motherboard is faulty!
As soon as I requested the seller to swap it to the one I actually ordered in the first place (Asus Z87A), there were no more crashes with the same 4 memory sticks in 2 months of daily operation! Now I'm a happy owner of 32GBs of memory without any stability problems.
Searching Google reveals that too many people are having the same problem. The ones that are content would be those that don't use all 4 memory slots (the majority). But when they decide to upgrade in the future, they'll be out of luck. Gigabyte, do the right thing and issue a recall.
-
Hi all, I work for an IT distribution company here in the UK. I work there as a system builder, I build PC's all day long. Some recent builds have been created using Gigabyte Z87-HD3 motherboards. They are built to a very high standard, I've even been criticized for being a perfectionist. Many of those builds are being returned due to random BSoD's. We have a returns department who are confused by this strange issue, but reading all your posts it seems the problem has been identified... using 4x sticks of RAM. The spec of the PC is below...
Coolermaster Silencio 550 case
Gigabyte Z87-HD3
4x 8Gb Kingston RAM
Intel i7-4770k
Samsung 840 250Gb SSD
Seagate 2Tb HDD
Coolermaster 600w PSU
Unless a reasonable response is forthcoming from Gigabyte with an explanation and intend fix for this issue, all builds using Gigabyte boards will be halted. In fact, I cannot allow these builds to continue. Any return is considered a fail for me, it's a money loser. A big thanks to you guys posting here for highlighting this fundamental issue with these Gigabyte motherboards.
-
Hi All
Please note this forum is a place where users help each other or share their experience, etc.
If you want to submit a question directly to GIGABYTE then please use the following link:
http://www.gigabyte.com/support-downloads/technical-support.aspx
Thanks
-
I have a GA-Z87X-UD4H and haven't yet seen this happen; I haven't had it up and running a long time, tho', and have installed Windows several times for some experiments I'm doing (so Windows hasn't been up for longer than a few hours so far).
I have 4x4G 1333Mhz Crucial sticks in it, and I haven't seen any BSODs yet.
However, I have 4x8G 1666Mhz Crucial sticks on order - I wonder if I'll start seeing this when I try that. Ugh.
- Tim
-
People have short memory's. This was there with the Z77 boards too. Some people can use 4 sticks and some people can only run two.
Is it the hardware people are using. I don't think anybody really knows at this point even Gigabyte.
Plus I have see this on other brands of boards too. I have been using Gigabyte boards for over 7 yrs.
I have never really had a board that did boot from using 2, 4, 6, & 8 slots of memory. I have had memory brands that didn't work.
I have see BSOD when using the XMP profile. Is this one of the problems. The memory list I have never used for ea board.
How do they do the test. Just slap some memory in and if boots OK that memory works next.
I mostly use G-Skill and a little Kingston. Does it really matter what speed of memory you use NO.
Do really see any performance change when going from 1333-mhz to 2400-mhz about 95% of the people will NOT.
Plus people come here start a thread tell us that they are having a problem. But how many of the them put up their SPEC"s too. "ONE"
http://forum.giga-byte.co.uk/index.php/topic,2195.msg10772.html#msg10772
-
... However, I have 4x8G 1666Mhz Crucial sticks on order - I wonder if I'll start seeing this when I try that. Ugh.
- Tim
Sorry, of course I meant 1600Mhz :-)
- Tim
-
People have short memory's. This was there with the Z77 boards too. Some people can use 4 sticks and some people can only run two.
Not everybody tests all systems. Most people skip chipsets and CPU generations. This ud5h is the first board that I've had problems with. It's ridiculous that a 4 dimm memory kit listed in the compatibility list cause problems with a computer. I mean, using 4 DIMMs isn't an outrageous thing to do and memory should be so standard that it should always work, unless there was a hardware fault. It's not like we're trying to fill our ram slots with pretzels and expecting things to work.
We're using RAM -> 4 dimms -> 4 slots... should make sense.
Here I am, still using only 16GB of the 32GB that I bought. I don't care that it's happened before. I don't care that this is a common thing. All I know is that this isn't working as advertised, Gigabyte didn't help me and only offered assistance like "did you make sure the computer is turned on?" (they offered a beta bios, as if they didn't know what the enigneers changed in that BIOS version).
I can't afford the downtime of having to RMA the mobo and wait for a new one to arrive. I could buy an ASUS right now, but then I'd have to rebuild the system from scratch and, again, I can't afford downtime. At least they could be honest and admit that this isn't an isolated problem and that there's nothing they can do from the BIOS side.
Well, at least the build is 100% stable when I use 2 dimms. Making a living here...
-
Just an update - I received my Crucial 4x8g sticks last night and installed them (32g total).
I ran a couple of video benchmarking programs for a few hours to see if I could trigger SOMETHING (since I figured video benchmarking would get memory and processor used quite a bit), and everything worked fine (tho' I noticed the Intel 4600 GPU isn't the fastest in the world, but I don't game so there you go).
No BSODs as yet.
- Tim
-
The only person not having a problem here is thessie. He is run the memory that the CPU really supports without OC.
But the thing is anything over 1600-mhz has to be OC. If there is anybody that is having a problem with 1333 / 1600 memory post it.
I'll say it again it is the hardware people are using. You just can't drop in 2400 memory set the speed or the profile and think it is going to work.
Yes Gigabyte advertised that you can run higher speed memory. But Gigabyte is telling you that anything over 1600 has to be Overclock too.
Support for DDR3 3000(O.C.) / 2933(O.C.) / 2800(O.C.) / 2666(O.C.) / 2600(O.C.) / 2500(O.C.) / 2400(O.C.) / 2200(O.C.) / 2133(O.C.) / 2000(O.C.) / 1866(O.C.) / 1800(O.C.) / 1600 / 1333 MHz
The OP never posted any of his spec's. The more memory you put in and the faster you run that memory the more it is going to stress the board.
The more you run the more you have to tweak the bios to support the faster memory.
-
The only person not having a problem here is thessie. He is run the memory that the CPU really supports without OC.
But the thing is anything over 1600-mhz has to be OC. If there is anybody that is having a problem with 1333 / 1600 memory post it.
I'll say it again it is the hardware people are using. You just can't drop in 2400 memory set the speed or the profile and think it is going to work.
Yes Gigabyte advertised that you can run higher speed memory. But Gigabyte is telling you that anything over 1600 has to be Overclock too.
Support for DDR3 3000(O.C.) / 2933(O.C.) / 2800(O.C.) / 2666(O.C.) / 2600(O.C.) / 2500(O.C.) / 2400(O.C.) / 2200(O.C.) / 2133(O.C.) / 2000(O.C.) / 1866(O.C.) / 1800(O.C.) / 1600 / 1333 MHz
The OP never posted any of his spec's. The more memory you put in and the faster you run that memory the more it is going to stress the board.
The more you run the more you have to tweak the bios to support the faster memory.
I had problems even when running the dimms at 1333Mhz. I spent weeks rulling out potential problems (GPU, BIOS settings, general voltages, system agent voltage, XMP, airflow, clocks... These settings were not a problem when running 2 sticks. The problem is as simple as it sounds: Populate more than 2 ram slots = BSOD sooner or later.
-
The only person not having a problem here is thessie. He is run the memory that the CPU really supports without OC.
But the thing is anything over 1600-mhz has to be OC. If there is anybody that is having a problem with 1333 / 1600 memory post it.
I'll say it again it is the hardware people are using. You just can't drop in 2400 memory set the speed or the profile and think it is going to work.
Yes Gigabyte advertised that you can run higher speed memory. But Gigabyte is telling you that anything over 1600 has to be Overclock too.
Support for DDR3 3000(O.C.) / 2933(O.C.) / 2800(O.C.) / 2666(O.C.) / 2600(O.C.) / 2500(O.C.) / 2400(O.C.) / 2200(O.C.) / 2133(O.C.) / 2000(O.C.) / 1866(O.C.) / 1800(O.C.) / 1600 / 1333 MHz
The OP never posted any of his spec's. The more memory you put in and the faster you run that memory the more it is going to stress the board.
The more you run the more you have to tweak the bios to support the faster memory.
That would be me. I'm running 4 sticks at their stock 1600 MHz clocks.
Motherboard: Gigabyte Z87X-UD3H
RAM: 8GB G.Skill F3-12800CL6-2GBXH (4x2GB)
CPU: Intel i7 4770K
OS: Windows 7 x64
I am on a 100% clean system, and have reinstalled Windows a couple of times thinking it was IRST issues. I decided to Google and ended up here. Liike all the others, it only happens with all 4 DIMM slots filled. So you can forget the "it must be your overclocked memory" theory -- mine are at stock clocks.
-
The only person not having a problem here is thessie. He is run the memory that the CPU really supports without OC.
But the thing is anything over 1600-mhz has to be OC. If there is anybody that is having a problem with 1333 / 1600 memory post it.
I'll say it again it is the hardware people are using. You just can't drop in 2400 memory set the speed or the profile and think it is going to work.
Yes Gigabyte advertised that you can run higher speed memory. But Gigabyte is telling you that anything over 1600 has to be Overclock too.
Support for DDR3 3000(O.C.) / 2933(O.C.) / 2800(O.C.) / 2666(O.C.) / 2600(O.C.) / 2500(O.C.) / 2400(O.C.) / 2200(O.C.) / 2133(O.C.) / 2000(O.C.) / 1866(O.C.) / 1800(O.C.) / 1600 / 1333 MHz
The OP never posted any of his spec's. The more memory you put in and the faster you run that memory the more it is going to stress the board.
The more you run the more you have to tweak the bios to support the faster memory.
You are assuming a lot in your post, I'm not overclocking anything, and the RAM is listed by Gigabyte as approved for use in the Z87-D3HP motherboard. It is high quality Corsair RAM and XMP is not enabled in the BIOS so it is running at the safe stock default speed of 1333MHz, I have tried it at 1600MHz and get the same results. I do not agree with you that the issue is the hardware I am using. Here are my specs.
Z87-D3HP motherboard
Corsair Memory CMX 1600C9 DDR3 (X2 sets) Total = 4X4GB sticks
CPU = Intel i5 4670
Windows 7 x64
2 DIMMS are fine but 4 DIMMS is a fail.
-
@No457 Snowy
Have you tried another brand of memory. Because it is on their list that doesn't mean it will run.
@OUT FOX EM
You are assuming a lot in your post, I'm not overclocking anything
@No457 Snowy
So you can forget the "it must be your overclocked memory" theory -- mine are at stock clocks.
Where in my post that it said you are OC your memory. This is what I posted: "But the thing is anything over 1600-mhz has to be OC."
Yes Gigabyte advertised that you can run higher speed memory. But Gigabyte is telling you that anything over 1600 has to be Overclock too.
So where is my OC theory. I was only Quoting what Gigabyte is saying. I never said that your were OC your memory.
That is why I ask if any body was having problems with 1333 & 1600. Plus I'll say it again that thessie isn't having a problem.
You should ask him what he is doing to get it to run.
-
@No457 Snowy
Have you tried another brand of memory. Because it is on their list that doesn't mean it will run.
@OUT FOX EM
You are assuming a lot in your post, I'm not overclocking anything
@No457 Snowy
So you can forget the "it must be your overclocked memory" theory -- mine are at stock clocks.
Where in my post that it said you are OC your memory. This is what I posted: "But the thing is anything over 1600-mhz has to be OC."
Yes Gigabyte advertised that you can run higher speed memory. But Gigabyte is telling you that anything over 1600 has to be Overclock too.
So where is my OC theory. I was only Quoting what Gigabyte is saying. I never said that your were OC your memory.
That is why I ask if any body was having problems with 1333 & 1600. Plus I'll say it again that thessie isn't having a problem.
You should ask him what he is doing to get it to run.
I read what you said. Did you read what I said? My memory runs at 1600 MHz stock. That ends your theory. Try again.
-
That ends your theory
So what theory is that. Show me where in any of my post that I said you were OC your memory.
This is my theory ask: thessie what is he doing that you guys are not. He has run 4 sticks of 1333mhz 16gbs & 4 sticks of 1600mhz 32gbs
thessie
I have 4x4G 1333Mhz Crucial sticks in it, and I haven't seen any BSODs yet.
However, I have 4x8G 1600Mhz Crucial sticks on order - I wonder if I'll start seeing this when I try that. Ugh.
Just an update - I received my Crucial 4x8g sticks last night and installed them (32g total).
I ran a couple of video benchmarking programs for a few hours to see if I could trigger SOMETHING (since I figured video benchmarking would get memory and processor used quite a bit), and everything worked fine (tho' I noticed the Intel 4600 GPU isn't the fastest in the world, but I don't game so there you go).
No BSODs as yet.
What bios is he running? How does he have his bios setup? What other hardware is he running that is different than yours?
-
That ends your theory
So what theory is that. Show me where in any of my post that I said you were OC your memory.
This is my theory ask: thessie what is he doing that you guys are not. He has run 4 sticks of 1333mhz 16gbs & 4 sticks of 1600mhz 32gbs
thessie
I have 4x4G 1333Mhz Crucial sticks in it, and I haven't seen any BSODs yet.
However, I have 4x8G 1600Mhz Crucial sticks on order - I wonder if I'll start seeing this when I try that. Ugh.
Just an update - I received my Crucial 4x8g sticks last night and installed them (32g total).
I ran a couple of video benchmarking programs for a few hours to see if I could trigger SOMETHING (since I figured video benchmarking would get memory and processor used quite a bit), and everything worked fine (tho' I noticed the Intel 4600 GPU isn't the fastest in the world, but I don't game so there you go).
No BSODs as yet.
What bios is he running? How does he have his bios setup? What other hardware is he running that is different than yours?
If it is all so specific to run these boards why does Gigabyte not publish this one very specific combination of hardware and config specs then? Instead we are to assume that even the published recommended RAM list they DO publish not worth consulting, according to you, and then we need to consult that one guy on the internet that managed to hit the jackpot of that one magical combination of hardware/config settings that Gigabyte themselves seem unable to enlighten us with? and unless we all follow that one particular recipe we should not expect our systems to operate as advertised?
That's just ridiculous.
If the boards only run 4 DIMMS with Crucial RAM and a whole bunch of other very specific settings like this one guy has, then Gigabyte needs to be up front about that and say so, don't publish a Corsair RAM part number on a list as compatible if it isn't.
Gigabyte need to own this and tell us, the paying customers EXACTLY what we need to use/do to run and fully utilise the product we paid for. The product does not operate as described, this has been reported by a LOT of people and it appears that Gigabyte do not know what specifics to recommend or how to fix it.
-
I'm sorry you guys are having a problem. But the thing is thessie is not why? I run two x79 boards and I have some problems with those boards too.
But have found a work around on both boards and they run great. My His 7850 4gb card only will run in one of the boards. Why I don't know don't care.
Both boards will BSOD if I run a Profile in the bios. So I set the speed and that is it. Like I said at the start some of the z77 boards had the same problems.
It looks like Gigabyte never fix it from the z77 boards. The only thing I can say is run two sticks of memory till Gigabyte might or can fix this problem.
From reading all the reviews that really more people are having problems with the USB connection than memory problem.
Plus go over to Tweaktown to see if they have a beta bios that will help.
http://forums.tweaktown.com/gigabyte/
-
Brand new Z87X-UD4H here, i bought it yesterday afternoon(24/12) im using Corsair Dominator Platinum 4x4gb 2800 C11 here without issues, also using xmp profile. They are running at 2800 11-13-13-36 all dimms filled. Using i7 4770k without oc (gotta buy a mew cooler for that) paired with a titan a dedicated asus sound card, one samsung ssd and a wd mechanical hd. What should i do to trigger the issue? Mine is Rev 1.0 amd came with F5 BIOS, updated yesterday to F7. I Noticed some mouse hangs in Windows 8 using it in a usb3.0 port then switched to a front usb2.0 and the problem is gone.
-
have you guys checked with gigabyte support?
you may email them to get this solved
-
Brand new Z87X-UD4H here, i bought it yesterday afternoon(24/12) im using Corsair Dominator Platinum 4x4gb 2800 C11 here without issues, also using xmp profile. They are running at 2800 11-13-13-36 all dimms filled. Using i7 4770k without oc (gotta buy a mew cooler for that) paired with a titan a dedicated asus sound card, one samsung ssd and a wd mechanical hd. What should i do to trigger the issue? Mine is Rev 1.0 amd came with F5 BIOS, updated yesterday to F7. I Noticed some mouse hangs in Windows 8 using it in a usb3.0 port then switched to a front usb2.0 and the problem is gone.
If you are not having any problems don't mess with to cause the problem. Then you might be screwed leave well enough a lone.
-
tweak town has the latest bios.
http://forums.tweaktown.com/gigabyte/28441-gigabyte-latest-beta-bios.html
Z87-d3hp --f6f
z87x-ud5h-f8e
you may try it.
-
Latest Beta BIOS (F8b) does not help for my UD4H. I guess gigabyte is just incapable to fix this. Better anyone gets another board. Shame on you, Gigabyte! Never thought that you would suck like this! This was my last board from you and thats for sure! Why? Because you haven't fixed this issue even after 6 months!
-
update:
Horrible voltage requierement for OC (In AUTO at 4.2 voltage JUMPS to 1.48, before in mi last sabertooth z87 jumped itself only at 1.28 at 4.3 in AUTO) Of course temps reaches 100ºc and system shut down.
Some kind of incompatibility, stability issues with my usb 2.0 pheripheras, laaags eveywhere in the mouse. tried turning off XHCI and enable it with no luck.
I had many prime95 errors when i manually oced the Memory .
I had by now up to 4 BSODs..
WHY THE HECK DID I LEFT ASUS TO TRY GIGABYTE, WHAT AN ASS HOLE. Maximus VI Hero.. here i come !
-
The C state could be the problem.
We had numerous issues with C states on IBM servers running x5000 series Xeon's where the return from idle state of a CPU / core would throw memory errors. If your in windows then it could be turning down or idling a core, and when that core returns from idle it then throws memory errors and BSOD's.
You can read IBM's (rather cryptic) resolve for it here :
http://www-947.ibm.com/support/entry/portal/docdisplay?lndocid=migr-5085841
Can someone try changing the C state back down to ACPI C2 and see if this resolves the issue ?
(note that the 'Intel' C state is a higher number, ie Intel C6 is ACPI C3)
I'm interested in this motherboard but a core requirement is being able to run 4x memory modules.
Good luck :)
-
Just confirming the setting ..
In the BIOS under Advanced CPU core features, change the C3/C6 state support setting to effectively disable C3/C6.
ie. change it to C2 or lower rather then 'Auto'.
(I don't have one of the machines in front of me so I'm not sure what options you actually get)
Kinda has a logical path as this C3/C6 state would only be initiated by an advanced OS like Windows.
Ergo, Windows BSOD's because it tells one or more cores to power down, then when they powers on again it gets a memory error and BSOD's.
You then fire up Memtest and check the RAM and do not get any errors because Memtest is a very low level OS which does not ever tell the cores to power down.
When I had this on our servers it only effected a particular bank of memory (the system has 18 memory slots with three banks of three channels for each CPU). It may be that in the desktop world this only affects the 4th bank due something like the distance from bank to the chipset.
Lemme know how you go ;)
-
In the BIOS under Advanced CPU core features, change the C3/C6 state support setting to effectively disable C3/C6.
ie. change it to C2 or lower rather then 'Auto'.
(I don't have one of the machines in front of me so I'm not sure what options you actually get)
Kinda has a logical path as this C3/C6 state would only be initiated by an advanced OS like Windows.
Ergo, Windows BSOD's because it tells one or more cores to power down, then when they powers on again it gets a memory error and BSOD's.
You then fire up Memtest and check the RAM and do not get any errors because Memtest is a very low level OS which does not ever tell the cores to power down.
Like you, Sh0ck-wave, I was quite interested in the UD4H until I read about these problems, but I backed off the purchase when I did a little more research and found out that this was going on. I'd still like to get one if there's ever an official fix, though I can't wait much longer and am beginning to lean strongly toward another manufacturer.
Your proposed solution sounded interesting, but I knew nothing about C states so I did a little research. What you're suggesting makes sense, and I wouldn't be surprised if it helped. However, if I understand C states correctly, I'm not sure I would consider this an acceptable solution. Again, I might not know what I'm talking about, but I got the impression that dropping to C2 or lower would cause inactive cores to draw significantly more power and thus create more heat and increase the rate of thermal degradation of the CPU. If that's the case, and I don't absolutely know that it is, I'd still have to choose a different motherboard.
So, I was hoping you or anyone else here would know if my impression is anywhere near the reality.
-
One day I gave up and decided to stop the "research" and continue just using 16Gb out of the 32Gb that I bought. I used this computer with 16Gb for some months and didn't even stop to think about it. The computer was so stable that I even went and overclocked it slightly (right now at 4.3Ghz on aftermarket air).
Two days ago I decided to give my 32Gb of RAM another go, but this time I wanted to try something I had read somewhere. That is increasing these 3 settings by +0,200v
-System agent voltage
-CPU IO Analog
-CPU IO Digital
I can't tell if this finally solved the problem, but I haven't had a BSOD in about 40 hours of use, which used to be impossible the first time I used all my 4 DIMMs at the same time.
I have to say this is a Hackintosh build.
-
Well well, this seems interesting. Would you believe me if I say I've been having BSODs while all dimm slots in use with my z87x-d3h board?
Yes, it's true and I've been scratching my head a while now. The newes F8 bios is no help on this regard, since I updated it yesterday and i had yet another BSOD while watching movie. It could be my aging GEIL 2133mhz Ram also, but since I've read so much about these issues with only gigabyte boards, I start to believe that the ram is not to blame.
I'm going to get new 4x set of ddr3 in day or two and I'll keep you posted. If BSODs stop, the ram is to blame, if not... well I guess I'll have to contact gigabyte.
-
I can confirm: NO CRASHES IN 4 DAYS (that's a lot!).
All I did was add 0,2v to:
-System agent voltage
-CPU IO Analog
-CPU IO Digital
Build:
z87x-ud5h
4770k
32Gb Gskill Tridentx 2400
Can someone else try this? I hope gigabyte take note as well.
-
I hope gigabyte take note as well.
They already have and right from the day the board came out. You have to OC your CPU a little just like flexmanta did. ;D
Support for DDR3 3000(O.C.) / 2933(O.C.) / 2800(O.C.) / 2666(O.C.) / 2600(O.C.) / 2500(O.C.) / 2400(O.C.) / 2200(O.C.) / 2133(O.C.) / 2000(O.C.) / 1866(O.C.) / 1800(O.C.) / 1600 / 1333 MHz memory modules.
By the book it says that you really only can run 1600 / 1333 MHz memory modules. At standard settings.
But it does say that you should check on their site for a updated list.
Yes I have read that people that are running lower speed memory are having problems too.
Plus what OS are you running too. I have heard that people running Windows 8 are having more problems than Windows 7.
-
I had crashes both in Windows 8 and in OSX. Now, with the 0,2v voltage increase in system agent and CPU IO Analog and Digital, no crashes in windows 8 or in OSX. Yes, I've done a very conservative multiplier overclock and memory is using XMP running at 2400Mhz (as advertised).
-
I have two sets of DIMMS and two computers.
The memory sets I use are as follows :
Set 1 - 4 x 8 GB Corsair Vengeance DDR3 1600 MHz 10-10-10-27
Set 2 - 4 x 4 GB Corsair Vengeance DDR3 1600 MHz 9-9-9-24
I've tested both of these sets with my new Gigabyte Z87-D3HP Motherboard
When I use the 8GB DIMMS I get constant crashes
When using the 4GB DIMMS it never crashes.
I might add that I placed the problematic 8GB DIMMS into another PC with the Gigabyte motherboard (GA-Z68XP-UD5) and ran a memory test on them for 5 days and it worked perfectly without any errors.
The issue for me at least appears to be related only to the 8GB DIMMS and the Z87-D3HP Motherboard.
I installed various linux distributions on this PC at first and it crashed several times, then I installed Windows 7 and started getting the blue screen issues.
I'm using the 4GB DIMMS in the Z87-D3HP right now and it's working fine.
The 8GB DIMMs are also working fine in the older PC on the GA-Z68XP-UD5
So the memory is confirmed good for me but this motherboard (Z87-D3HP) has issues with the 8GB DIMMS that cause it to randomly crash.
-
@ nmc
How are you setting the memory in the bios? Plus till Gigabyte gets this fix you might have to run the 16gb till then.
One other thing too is you are running 4 sticks where other people can't.
-
@ nmc
How are you setting the memory in the bios? Plus till Gigabyte gets this fix you might have to run the 16gb till then.
One other thing too is you are running 4 sticks where other people can't.
Everything in my BIOS is set to the default settings as the board came. I don't know what most of those settings are so I never touch them.
-
They will never fix this. We are already on F8... they are inapt
-
They will never fix this. We are already on F8... they are inapt
I agree - this motherboard of mine will go on the scrapheap and I will remember the name 'Gigabyte' as something to avoid in the future due to this problem.
Now I have a decision to make, what brand to try next....
-
Sorry to say that I've gotten cold feet over this whole situation.
I was after the micro-ATX version of the Gigabyte Z87 board, but opted for the Asus Z87M-Plus instead as there were write-ups from people who had made exactly the type of system I wanted to make using it (an ESXi white-box running linux and win servers as well as a windows 7 gaming VM with passthrough GFX card).
RAM is a big requirement, and the last thing I wanted was a situation like this.
I did install a Gigabyte 7750 OC 2GB card though for the passthrough and that's gone sweet, so I guess the 'Big G' still got it's foot in there.
Good luck everyone.
-
Yep, I can confirm, no 4 sticks of ddr3 ram with z87-d3h without issues. I finally got my Team Group Vulcan CL11 2400Mhz memory 4x4GB and it won't cope with gigabyte when all dimm slots in use. However the new F8 bios has been able to stop the random bsods which were frequent with F7 bios. I ran HCI memory test over night and I got two night in a row memory errors with team group memory and my geil memory. Then I took two dimms out and no erors. I tested again both team group and geil.
I have to say, I'm really baffled about how gigabyte handles things. Haswell was already old when it came out and the motherboards were ready long before haswell was released, so how come gigabyte hasn't still been able to get rid of this problem? As if they havn't done any proper testing on their boards.
Why do I say so, well here is a list of problems even so long after haswell release:
1. 4 dimm memory problem
2. Bios dashboard cannot give you fullhd picture unless you hook up something to your hdmi port alongside with dvi port
3. Bios dashboard doesn't save and load properly your settings and you need to boot sometimes several times before you get the right setting which you supposedly setup. Yes, when you hit aplly, you will see that bios has set the clocks and voltage and s***, but then you boot to windows and cpu-z says different. Then you need to boot once or twise more and voila, you get what you want. But not always, sometimes you have to again bios and set some more. Really annoying.
4. Don't know you guys, but I had some serious GFX card not been able to give signal to monitor issue as well. For some reason I got black screen with my radeon 7850 and different brand 7870 cards. Then I hooked yp asus r9 290 and voila, picture straight away. Same with xfx 7970 card. I tried both my 78xx series cards in every pci-e slot and no result. Setting were the same all the time, pci-e is set to primary output.
5. Then there is the software package which is nice looking and work pretty ok, but I cannot fathom how the coders have been able to make so simple software so sluggish and hog so much memory?
Anyway, I'm done with gigabyte, on paper really good, but in practice, not so much. I've already placed an order for asrock z87 killer board. It's the same price but better hardware and better software.
-
Appears to be fixed with F9a. Haven't crashed yet @ 1600mhz xmp enabled but still testing. It would've crashed by now multiple times on F8 so keeping my fingers crossed.
-
Anybody that is going to throw the board in a scrapheap or trash throw them my way I'll make them work. ;D
-
I have two sets of DIMMS and two computers.
The memory sets I use are as follows :
Set 1 - 4 x 8 GB Corsair Vengeance DDR3 1600 MHz 10-10-10-27
Set 2 - 4 x 4 GB Corsair Vengeance DDR3 1600 MHz 9-9-9-24
I've tested both of these sets with my new Gigabyte Z87-D3HP Motherboard
When I use the 8GB DIMMS I get constant crashes
When using the 4GB DIMMS it never crashes.
I might add that I placed the problematic 8GB DIMMS into another PC with the Gigabyte motherboard (GA-Z68XP-UD5) and ran a memory test on them for 5 days and it worked perfectly without any errors.
The issue for me at least appears to be related only to the 8GB DIMMS and the Z87-D3HP Motherboard.
I installed various linux distributions on this PC at first and it crashed several times, then I installed Windows 7 and started getting the blue screen issues.
I'm using the 4GB DIMMS in the Z87-D3HP right now and it's working fine.
The 8GB DIMMs are also working fine in the older PC on the GA-Z68XP-UD5
So the memory is confirmed good for me but this motherboard (Z87-D3HP) has issues with the 8GB DIMMS that cause it to randomly crash.
Hey mate I got same issue as you when I was running Z87-D3HP F6 BIOS, my memory model is CMZ32GX3M4X1600C10. Fortunately it seems to be fixed after I updating F7a BIOS from the link below.
http://forums.tweaktown.com/gigabyte/28441-gigabyte-latest-beta-bios.html (http://forums.tweaktown.com/gigabyte/28441-gigabyte-latest-beta-bios.html)
You might give it a try buddy. Good luck!
-
I use a Z87-DS3H from them. Worst board I've owned in my life, but Hackintosh. Same problem. I can't use more than 2 bays or it locks up. I contacted Gigabyte and they basically replied with a bunch of questions totally irrelevant... including asking me about the specs of the RAM I'm running... which I had given them the exact model number of...
I'm using Corsair Vengeance 1600mhz RAM. It's on their list of supported RAM. I'm using a Gigabyte NVIDA GTX760 card. Liquid cooled i7-4770k. This is ridiculous. I'm actually contemplating sending them a copy of my receipt (bought it all about a month ago at a place with a 2 week return policy) and telling them I want my money back.... but judging by the reply I've already gotten from these people, I'm guessing they won't be able to figure out what the character we commonly refer to as numbers on the receipt are.
-
Can anyone try this fix?
It seems that this guy could solve the problem using the manual profile and increasing the value of trrd, tfAW and tRFC.
Complete post:
Hi guys, just to share, I think I found the problem with 4 dimm memory instability on haswell gigabyte (Z87/H87)
I collect a lot of information about memory changes from z68/z77 to z87 and found that memory subtimings has changed, and now we have "haswell compatible memory" for some brands. I didnt find all changes but it clear to me that tRFC and tFAW have higher values now.
and comparing some memory brands who has 16GB (2 sticks) and 32GB (4 sticks) on the same model/IC, I also found tRFC and tFAW higher on the 4 sticks kit.
so, first conclusion is: HASWELL NEEDS HIGHER tRFC and tFAW and 4 DIMM needs even HIGHER values.
I dont know how other motherboard brands solve this, but gigabyte still isnt fix the 4 dimm problem on their boards, you can see a few users happy using brand new memories from QVL (gigabyte have a very small ones) but there is a lot of people still getting problems.
I have a old 16GB 4 DIMM memory kit with works fine at 2 DIMM XMP 1600Mhz 1.35v at my H87 board, but wont work at all with 4 DIMM. i try everything: lowering timing/latency, increase System Agent,, IO Digital/Analog voltage and several others tricks without success. But I didnt try on sub-timings values and when I see the values I notice that are VERY LOW tRFC and tFAW compating to other 4 DIMM kit. it seems tRRD is also related to tFAW, so i recommend higher a little (I use 8)
so, if you can try a little experiment with 4 dimm, do this:
1) note (or print/photo) all RAM TIMINGS you are using on XMP profile.
2) disable XMP profile and put memory timing on MANUAL
3) change manually all timings values to the XMP values you printed, but increase tRFC to 280˜310 and tFAW to 50˜80, also change tRRD to 8 or higher ..... it seems this could solve the problem with 4 dimm... (**)
4) adjust memory voltage to XMP values and memory multiplier to 1600MHz (depending on RAM, it could be changed...)
5) put the memory mode to NORMAL or ENHANCE STABILITY. verify if command rate is 2T
Reboot and check in BIOS if the values are ACTUALLY set (in my H87, they only changed after a cold boot: remove power cord - or turn of power switch if you have - for 20 seconds and turn on again).
do a Cold boot and check if the values still there.
then go to OSX/Windows and run some benchmarks. check the timing values inside windows (AIDA, Everest, etc. I don't know how to check in OSX) to see if it still there even if you reboot or cold boot.
You don't need to change voltages or other settings with works well with 2 DIMM, just these two timing values.
(**) these values almost double from memory kits with use double side ram or has 2 and 4 dimm configurations (same model/chip). So it seems theses values have direct relation with memory density. Also, older memory modules (launched before haswell) has 120˜160 tRFC and 28˜40 tFAW, but the new ones (advertising as haswell compatible) has higher values. tRRD is related to tFAW, the minimum possible value is 4 and some 4 dimm kits on gigabyte QVL has tRRD 6, 7 or 8. I recommend 8 for stabilty.
gigabyte bios has bugs, in my case it require a cold boot to change values, but i think it is possible to do manually solve this issue if we understand exactly where the problems is.
share your comments here about this experiment. please.
at least with my memory kit, it seems to be solved. I'm now running 4 DIMM 16GB 1.35v 1600MHz replicating XMP profile at manual mode (but increase tRFC and tFAW) without changing any other voltage parameters for 30 hours free of crashes.
(english is not my natural language)
Source: http://www.tonymacx86.com/general-help/125717-fix-gigabyte-4-dimm-memory-instability.html
-
The only way I "fixed" this was by enabling the onboard intel graphics, whatever the BIOS version and fixes.
-
I also had the dreaded issues with 4 sticks of RAM.
My system ran flawlessly with 2x4GB G.Skill RAM Ripjaws by themselves
I first ran this PC3-2400 Ripjaws X 8GB set by itself for 24hrs using XMP profile and adjusted the voltage to 1.66Volts
http://www.newegg.com/Product/Product.aspx?Item=N82E16820231666
When I added my second set of G.Skill PC3-2400 Sniper with the same wait states but different model
http://www.newegg.com/Product/Product.aspx?Item=N82E16820231667
The system would power on, do some post codes, shut off, power back on, do post codes and then off again forever until I killed the power
I did finally get all 4 to work but I had to start with the pair of 4GB G.Skill Sniper PC3-2400 first in the A slots
I turned off XMP profile and did a manual setup with the wait states as specified at 11-13-13-31 and left the others on auto
I turned off the CPU energy saving features so that my i3-4330 is locked at a constant 3.5GHz
I used the standard memory speed adjustments then dialed in at PC3-1600 with the wait states at 11-13-13-31, the rest on Auto
I went into the CPU voltage settings and added 0.20 volts to the following and hit apply:
-System agent voltage
-CPU IO Analog
-CPU IO Digital
I saved and turned the power off, waited for the P/S to discharge and added the Ripjaws 2x4GB modules in the B channel
The system booted this time finally!!!!
I kept checking my voltage monitoring and wait states to make sure I was reading 1.66volts and my wait states were sticking.
I raised the memory multiplier to first get 1866 saved and rebooted.
To my surprise it started and I went into the BIOS to make sure voltages and wait states were sticking again.
I raised all 4 sticks to 2133MHz and at 11-13-13-31 times, 1.66 volts (rated 1.65)
Saved and restarted, it went into the BIOS again. Everything looked good.
Right now I am stability testing. I am going to run all 4 modules (16GB) all night on Memtest86+ 4.20 all night at 2133MHz
Its 1:30AM and I don't have any errors yet after only 35minutes.
I will check this in the morning.
I may just leave them at 2133 and call it quits at this point. The difference between 2133 and their default rates of 2400 is pretty much bragging rights.
OH, I forgot to mention I am using the F8 BIOS. I cannot seem to get the F9 to install with either @BIOS in Win 8.1 or off the USB stick flashing from inside the BIOS GUI
-
I also had the dreaded issues with 4 sticks of RAM.
My system ran flawlessly with 2x4GB G.Skill RAM Ripjaws by themselves
I first ran this PC3-2400 Ripjaws X 8GB set by itself for 24hrs using XMP profile and adjusted the voltage to 1.66Volts
http://www.newegg.com/Product/Product.aspx?Item=N82E16820231666
When I added my second set of G.Skill PC3-2400 Sniper with the same wait states but different model
http://www.newegg.com/Product/Product.aspx?Item=N82E16820231667
The system would power on, do some post codes, shut off, power back on, do post codes and then off again forever until I killed the power
I did finally get all 4 to work but I had to start with the pair of 4GB G.Skill Sniper PC3-2400 first in the A slots
I turned off XMP profile and did a manual setup with the wait states as specified at 11-13-13-31 and left the others on auto
I turned off the CPU energy saving features so that my i3-4330 is locked at a constant 3.5GHz
I used the standard memory speed adjustments then dialed in at PC3-1600 with the wait states at 11-13-13-31, the rest on Auto
I went into the CPU voltage settings and added 0.20 volts to the following and hit apply:
-System agent voltage
-CPU IO Analog
-CPU IO Digital
I saved and turned the power off, waited for the P/S to discharge and added the Ripjaws 2x4GB modules in the B channel
The system booted this time finally!!!!
I kept checking my voltage monitoring and wait states to make sure I was reading 1.66volts and my wait states were sticking.
I raised the memory multiplier to first get 1866 saved and rebooted.
To my surprise it started and I went into the BIOS to make sure voltages and wait states were sticking again.
I raised all 4 sticks to 2133MHz and at 11-13-13-31 times, 1.66 volts (rated 1.65)
Saved and restarted, it went into the BIOS again. Everything looked good.
Right now I am stability testing. I am going to run all 4 modules (16GB) all night on Memtest86+ 4.20 all night at 2133MHz
Its 1:30AM and I don't have any errors yet after only 35minutes.
I will check this in the morning.
I may just leave them at 2133 and call it quits at this point. The difference between 2133 and their default rates of 2400 is pretty much bragging rights.
OH, I forgot to mention I am using the F8 BIOS. I cannot seem to get the F9 to install with either @BIOS in Win 8.1 or off the USB stick flashing from inside the BIOS GUI
Next day after 9 hours of MemTest86+, 8 passes, no errors with 4 sticks
Win 8.1 Pro, no issues, installed several apps
-
wow I've been pulling my hair out from the bsods tried so many things, just found this thread thanks !
Z87-D3HP motherboard with 4x 1600mhz corsair RAM
Always same BSODs "PAGE_FAULT_IN_NONPAGED_AREA", is this it ?
Really random, memtest runs fine but afaik it's since it's not a memory issue, it won't show on memtest anyway.
Is there any way to stresstest the system to reproduce the bsods ?
I've got little time left, if I can't fix it soon I'll have to RMA it and go with some other brand cheers
-
Today i returned a z87mx-d3h.
First I got instant freezes and after I manually set the voltages (raised them) and set the timing the system ran for 2 days non stop.
After reboot I got a freeze again.
I am tired of debugging Gigabyte Mainboards.
Anyone else experience the freezes with other brands (z87) Boards or are these freezes a Gigabyte z87 issue ?
-
I'm going back ASUS!! The Maximus VI Formula is just... perfect!!
I returned a GA z87MX-D3H today.
Do you meanwhile have tried a ASUS Board ? They are also freezing ?
-
@psycho is right, enabling on-board video fixes it.
Z87X-UD4H
i7-4770k
4x8gb Mushkin 997110 (1600mhz 9-9-9 1.35v)
With the i7's onboard video disabled and all four 8gb sticks in the board I got app crashes and BSOD's about every hour or two regardless of what timings or voltage I used. With onboard video enabled the crashes have gone away completely. Its 100% stable using the RAM's XMP profile with no additional BIOS changes.
-
The only way I "fixed" this was by enabling the onboard intel graphics, whatever the BIOS version and fixes.
I do also work.
I thought I was returning crazy.
at least 10 or 15 daily BSOD
thanks
PD: My English is not very good, sorry.
Z87X-UD3H -- BIOS F8
i7 - 4770k
4x4GB g.skill ripjaws z 1866
Win 8.1 update.
-
Lowering memory to 1333hz fixed it for me (1 week and no bsods), but I find it inadequate solution
Cool, I'm going to try enabling onboard intel graphics as suggested (and get memory back to 1600hz)
-
I bought a Z87X-UD5H.
It's not even turning on because I'm using Haswell Refresh ::)
I'm gonna use it with 32GB RAM. Now I'm afraid I have another boom to boom when I get the BIOS updated :(
-
Update:
~4 days after switching back to 1666mhz memory and enabling internal graphics had a BSOD
Gonna finish building a backup pc, then switching main to asus motherboard
Been buying exclusive gigabyte hardware for 7 years, not touching anything from them again, lesson learnt
-
Thanks for the update.
My mobo is still on RMA due to DualBios issue. It will be a shame if after I receive it back I have RAM issue. I'll have to RMA again and buy another one. I see I won't be able to properly use my PC for like 3 months yet.
Thanks Gigabyte, we love you!
-
@svartberg I got exactly the same motherboard Z87 D3HP (rev1) (and same problem btw)
Which BIOS are you using ? On which OS ?
I'll try switching internal graphics back on but seemed I had the same problem before.
For the record :
(http://www.zupmage.eu/i/nxdMMFZpYc.png)
About my tries with Gigabyte support (french link) : http://www.wareziens.net/forum/topic-21833-le-support-gigabyte-page-1.html
-
Guys try this BIOS:
GA-Z87-D3HP - F7c (http://www.mediafire.com/download/i68us0bgiwa7r88/Z87D3HP.7c.zip)
Edit:
Other Z87 boards BIOS also updated on tt forum today.
http://forums.tweaktown.com/gigabyte/28441-gigabyte-latest-beta-bios.html
-
Updated from bio v9 to the new bios 10b with my Z87X-ud3h, to see if 4 DIMM issue fixed, but I now have no on board audio (as far as my OS X build is concerned). Has anyone else seen this?
I have yet to replace my 3rd and 4th DIMMs, so unable so far to report on that issue.
-
@svartberg I got exactly the same motherboard Z87 D3HP (rev1) (and same problem btw)
Which BIOS are you using ? On which OS ?
F2 Bios on Windows 7 64bit
-
Guys try this BIOS:
GA-Z87-D3HP - F7c (http://www.mediafire.com/download/i68us0bgiwa7r88/Z87D3HP.7c.zip)
Got f7b actually but I'll upgrade as soon as possible and keep it touch thx.
-
Other Z87 boards BIOS also updated on tt forum today.
http://forums.tweaktown.com/gigabyte/28441-gigabyte-latest-beta-bios.html
I have 4 DIMMS in my Z87X-UD3H with F10b. No crashes any more (3 days). :)
Seem to be some hidden changes to audio, but I now have that working again.
-
Everything I read points that its a hardware board limitation (failure caused by poor power transfer to the RAM or something)
That's why I gave up on it, and switching the board to different brand soon
Btw on 1666mhz I get about 1 crash a week, faulty engineering is still faulty
-
Guys try this BIOS:
GA-Z87-D3HP - F7c (http://www.mediafire.com/download/i68us0bgiwa7r88/Z87D3HP.7c.zip)
Edit:
Other Z87 boards BIOS also updated on tt forum today.
http://forums.tweaktown.com/gigabyte/28441-gigabyte-latest-beta-bios.html
Just tried, had to disable Intel virtualization in order to boot but it seems better.
Motherboard had auto-detect 1600 memory as 1333, even with only 2 slots. I've forced 1600 from BIOS (with some graphical strange bugs) and it seems to work pretty well (with 4 slots 4gb each)!
I'll get back to you in a few days.
Thansk @ex58 for the update notice.
-
Everything I read points that its a hardware board limitation (failure caused by poor power transfer to the RAM or something)
That's why I gave up on it, and switching the board to different brand soon
Btw on 1666mhz I get about 1 crash a week, faulty engineering is still faulty
1666 MHZ....?
You even don't know your memory speed,not to mention your opinion.
-
1666 MHZ....?
You even don't know your memory speed,not to mention your opinion.
It's a pun, lighten up
-
@svartberg: flash the bios with the f7c ;)
-
Hy
new entry - my GA-Z87-HD3 -brand new - have a bigger issue. I can start just whith 1 dimm memory :D.
I instal my components and my sistem not start. First sign from motherboard - continuous short beep and self restarting after 10 or 15sec. After a short break i manage to start just whith a single memory installed. I had no luck with dealer service. Sorry but the matherboard was packed improperly and "physical defect caused bent pins. not subject to warranty". No test, no conclusion. Some good advice?
https://www.youtube.com/watch?v=FdXYIFSP8yo&feature=youtu.be
-
Update, bought asus motherboard z07a, unlike the gigabyte the z97A has a crappy onboard sound so had to buy that too
Anyway screwed again, the asus z97A doesn't support the Kingston KVR16N11/8 8gb sticsk in 32gb mode :(
So lost a lot of money & time, jakes on me :(
I only post to warn you guy the real problem in my case was kingston poor compatibility not motherboards.
-
Just an update, I previously reported that enabling on-board video on my Z87X-UD4H seemed to solve my BSOD's in Win8 but I've switched to Fedora Linux and under Linux it was immediately obvious that my RAM was still unstable. So I started testing with Memtest86 and sure enough I had a bad stick. Looks like my original set was bad and so was the set Newegg sent me to replace it. Mushkin support was great. They sent me a new 4-stick 4x8GB set (#994110) and it is completely stable now. It wasn't the motherboard after all.
Don't make the mistaken I did and do your debugging based on whether you get app crashes or BSOD's. Properly test your sticks with Memtest86. I used V5.1.0 Free from www.memtest86.com. Tests #8 and #9 consistently threw errors before but now with the new 32GB quad kit I can run memtest86 for 12+ hours with no errors and Windows & Linux are completely stable.
-
I've just been on with Gigabyte support about this issue on my Z87-HD3 board.
I had 2 pairs of Corsair Vengeance LP RAM, and I bought some more - 2 new pairs of the same brand/model (cheeky bid on ebay worked!) so I put one of the new pairs in, and instant system instability. Processes would crash, occasional BSODs, etc.
Turns out that the XMP profiles on the RAM can change. My 3 pairs of RAM all looked the same, same brand, model, etc but one was rev 2.12, another rev 7.4 and my original pair rev 8.0
Turns out the rev 2 and 7 sticks shows XMP 1.2, and the rev 8 sticks showed XMP 1.3 - which had different timings. The tRC was different at 33 (whereas support showed the ram running with tRC of 41).
In my case, I swapped the sticks about and they all showed XMP 1.2 and stability has been perfect since.
So I have to say its not gigabytes fault (in my case), more Corsair's for changing the spec of their ram without any notification. It could be this is the same issue for other RAM manufacturers, though I see a few posters are using Corsair too.
To tell, get CPU-Z and run it, on the SPD tab you can see the timings and XMP version used for each RAM slot.