Gigabyte Z87X-UD3H LGA 1150 Motherboard Review @ [H]

FrgMstr

FrgMstr

Just Plain Mean
Staff member
Joined
May 18, 1997
Messages
55,634
Gigabyte Z87X-UD3H LGA 1150 Motherboard Review - The motherboard we are checking out today is the GIGABYTE Z87X-UD3H which is on the lower end of the pricing spectrum but offers a solid feature set and looks to have all the ingredients of a winner. This motherboard is still very much focused on the enthusiast and overclocking crowd offering a very solid list of overclocking features.
 
Very interesting review. Especially the comments at the end about the electromigration. Is there a potential issue with the Haswell processors having some durability problems? Time will tell. I have been sitting on this i7 930 for a couple of years with no problems and have a new i7 4770k in the mail. (went with an Asus board tho). I am just not impressed with the Gigabyte boards.
 
Last edited by a moderator:
Thanks for reviewing this board. I currently have this board in my Amazon cart but after reading about the DPC latency and NIC issues I will have to wait on buying it. Also, the issue with the 4770k is very interesting and would like to definitely see a follow up on that.
 
techie81 said:
Thanks for reviewing this board. I currently have this board in my Amazon cart but after reading about the DPC latency and NIC issues I will have to wait on buying it. Also, the issue with the 4770k is very interesting and would like to definitely see a follow up on that.
Click to expand...


Yeah, I promised a retail 4770K OC article a month ago, but have been holding off on this as things have been developing and I am trying to get my ducks in line. Still going to be little objective proof and more of a subjective analysis of a very small sample of CPUs, basically just sharing our experience and thoughts. But we can pick up on this discussion in another thread. I would like to keep this one focused on the mobo.
 
Nice review.

Kyle, how long did you have this board up and running overclocked? I had a Z87X-D3H and I noticed my once Prime/OCCT/IBT stable 4770k @ 4.2ghz became unstable after a week, I had to increase the Vcore from 1.15v all the way to 1.20v to become stable again. My temps increased 8-10c while stress testing.
 
Thanks for the review. I hope you guys can do a review with the UD4H which has double the number of VRM phases for the CPU with the same part.


I have a Z87X-OC which has the same part for VRM IR3553M and same number of them (8), this is the same as the UD3H.

I'm not happy with the over clocking performance on my Z87X-OC either.

I think 16 phases of the same IR3553M will lead to better results on the UD4H.


I have read many bad reviews on Haswell gigabyte mother boards that use the Inter National rectifier power control packs to power the CPU. Its all digital and should be able to be controlled but gigabyte needs to spend more time on getting it right.
 
SonDa5 said:
Thanks for the review. I hope you guys can do a review with the UD4H which has double the number of VRM phases for the CPU with the same part.


I have a Z87X-OC which has the same part for VRM IR3553M and same number of them (8), this is the same as the UD3H.

I'm not happy with the over clocking performance on my Z87X-OC either.

I think 16 phases of the same IR3553M will lead to better results on the UD4H.


I have read many bad reviews on Haswell gigabyte mother boards that use the Inter National rectifier power control packs to power the CPU. Its all digital and should be able to be controlled but gigabyte needs to spend more time on getting it right.
Click to expand...

In my case, I'm waiting for the UD5H (which, at least, at MicroCenter, replaced the UD4H in their midrange bundle - same price in the bundle, even).

While the dual-gigabit LAN is absolutely overkill, all my OSes (Windows 8.1, Server 2012R2, and OS X ML/Mavericks) can take advantage for dual-gigabit teaming.
 
Thank you for the DPC test and quick LAN text summary! Appreciate it.

I can't believe Gigabyte still hasn't improved. I have a feeling they have one thing wrong or another with their boards since the very good P45 days.
 
This was a good review, however I feel that something is amiss here. The gaming benchmark graphs did not have the Z87X-UD3H, I only see a UD5H:

1375169267ac4bgSJ7kS_6_2.png


1375169267ac4bgSJ7kS_6_2.png


1375169267ac4bgSJ7kS_6_3.png
Click to expand...
 
Excellent review guys! Well done on looking further into the electromigration rather than just blaming it on the motherboard. I felt something was amiss on the Sabertooth motherboard review when it couldn't hold a 4.5 over your 2 day torture test cause I consider the Sabertooth line to be the pinnacle of stability and overclocking. I had never heard of electromigration until reading this article but it does make more sense now.

On the UD3 though, I am a little disappointed that the recent Gigabyte Z87 boards are looking a little quirky. Their Ivy boards were very solid. Looks like MSI and Gigabyte have swapped places with Haswell with MSI racking up a couple Gold awards right out of the gate.
 
MacLeod said:
Excellent review guys! Well done on looking further into the electromigration rather than just blaming it on the motherboard. I felt something was amiss on the Sabertooth motherboard review when it couldn't hold a 4.5 over your 2 day torture test cause I consider the Sabertooth line to be the pinnacle of stability and overclocking. I had never heard of electromigration until reading this article but it does make more sense now.

On the UD3 though, I am a little disappointed that the recent Gigabyte Z87 boards are looking a little quirky. Their Ivy boards were very solid. Looks like MSI and Gigabyte have swapped places with Haswell with MSI racking up a couple Gold awards right out of the gate.
Click to expand...

With MSI they've always impressed out of the gates more than say 6mnths later. I've had many MSI boards that I loved the first month I had them
 
Cmustang87 said:
This was a good review, however I feel that something is amiss here. The gaming benchmark graphs did not have the Z87X-UD3H, I only see a UD5H:
Click to expand...

We upped the wrong set of graphs on that page. Will have Dan up the right ones. My bad on not catching that when we went live.

Here are the raw scores for comparison.

Games:

Lost Planet NoAA - / 640x480 low

4770K - 343 fps


Heaven Full Screen, DX11, NoAA - / Low,640x480

4770K - 367 fps


BioShock Infinite - / Very Low 640x480

4770K - 440 fps


Metro Last Light - DX11- NoSSAA No PhysX - 1024x768 - Low - AF4X - MoBlur Low - Tess Off

4770K - 158 fps


As I have talked about in the past. Mobo benchmarks now days are much more to see if something is wrong, rather than "right."
 
I've sent Kyle the updated graphs. You should see them shortly.
 
I've been hoping for some rapid fixes with this board, but for being the first GB board I've tweaked (my wife has the p45-ud5h, but that's a stock build) this has been a let down for sure. The fan controls aren't as nice as my ancient Abit board I used with my q6600 that I upgraded from, I have odd bugs (wake on mouse when it's turned off in bios), and my OC hasn't been what I was hoping for.

Here's hoping we'll get something soon to make this better.
 
The place where I see the most difference in terms of benchmarks is in regard to network testing. Those can literally be all over the map using the same NIC on different boards. For whatever reason this board showed unusually low transfer speeds compared to other boards with the same Intel i217v NIC.
 
Jeremy C said:
I've been hoping for some rapid fixes with this board, but for being the first GB board I've tweaked (my wife has the p45-ud5h, but that's a stock build) this has been a let down for sure. The fan controls aren't as nice as my ancient Abit board I used with my q6600 that I upgraded from, I have odd bugs (wake on mouse when it's turned off in bios), and my OC hasn't been what I was hoping for.

Here's hoping we'll get something soon to make this better.
Click to expand...
Do you confirm the DPC latency problems?
 
Warl0rd said:
Do you confirm the DPC latency problems?
Click to expand...

I'll take a look tonight and let you know.

What I can confirm for sure is that there is still issues with the RAID speeds on this board, and it affects more than just the member disks. I had 2 1TB VR's in R0 on this board, and read/write never got close to what it should have. What I didn't realize was that I was also taking about a 10% speed hit on my 840 Pro. After removing the VR's and switching back to AHCI, the 840 is running at it's rates speeds (both transfer and IOP). The VR's are on an ancient p45 GB board now, and they are hitting the numbers I expected with this machine.
 
I have one of these boards. Had a few odd problems with it. When I set primary display output to PCIE 1, it doesn't work. Since I 1st built the system I had this prob when setting it up. Last night the pc crashed whilst in Crysis 3 single player campaign. Now it wont display through my GTX 690 at all. It's like it just doesn't see it at all. It'll display through the inegrated gpu but not descreet. When I go into Nvidia control panel it says 'no Nvidia hardware detected'.
Has anyone else come accross anything like this?
 
I am still using the onboard so I can't say for certain. I plan on getting a card, hope this is not a issue I am not aware of yet.
Have you tried setting it to iGPU? if a card is present it will bypass and simply go to PCIE1
Any chance you can try your card on another computer to determine if it's working or failed?
 
Dan_D said:
Yours has consistently higher latency than it should. WTF?
Click to expand...

I would love to know that myself. I also get random drop outs of the rear output jack (with the "audio device unplugged" pop-up), so I'm thinking I might RMA this board and try again. I wonder if the C2's are making their way out yet...
 
Jeremy C said:
I would love to know that myself. I also get random drop outs of the rear output jack (with the "audio device unplugged" pop-up), so I'm thinking I might RMA this board and try again. I wonder if the C2's are making their way out yet...
Click to expand...

Not that I know of, but an RMA sounds like a good idea. The latency your seeing shouldn't be that bad. Nor should you have audio-device unplugged messages.
 
Jeremy C said:
I would love to know that myself. I also get random drop outs of the rear output jack (with the "audio device unplugged" pop-up), so I'm thinking I might RMA this board and try again. I wonder if the C2's are making their way out yet...
Click to expand...

I have this exact same pop up from the Realtek software. Does it regularly.
My card turned out to be faulty, had to RMA it, new card in now, being seen by OS and Nvidia driver, but can't yet get it to display through it, just keeps displaying through the iGPU instead. There are def some problems with these boards. With it basically being an updated Z77 in many ways, I can't see what possible excuses Gigabyte could have. I'm finding a few things across the web now of different problems with these boards, including the prob I'm having with the display. Really not what I'd expect from a company such as Gigabyte.
 
Dan_D said:
Not that I know of, but an RMA sounds like a good idea. The latency your seeing shouldn't be that bad. Nor should you have audio-device unplugged messages.
Click to expand...

This is ancient, but figured I would make an update since I just finally figured out the DPC latency concern. I RMA's my board, did a fresh W8 install, and was still seeing the same latency. After doing some reading apparently DPC Latency checker is not a good program to use on W8 machines. When I first tried there was nothing on the website about it, but now on the download page there is this message:

Windows 8 Compatibility: The DPC latency utility runs on Windows 8 but does not show correct values. The output suggests that the Windows 8 kernel performs badly and introduces a constant latency of one millisecond which is not the case in practice. DPCs in the Windows 8 kernel behave identical to Windows 7. The utility produces incorrect results because the implementation of kernel timers has changed in Windows 8 which causes a side effect with the measuring algorithm used by the utility. Thesycon is working on a new version of the DPC latency utility and will make it available on this site as soon as it is finished.
Click to expand...

So I downloaded latencymon and get much more accurate numbers. Though I don't have a graph to visualize it, I'm showing about 50µs with spikes to about 100µs. Highest spike from running 10 minutes was 600µs. Not the greatest, but a far site better than what I was reporting previously. I also have since disabled the onboard sound and with an Sound Blaster Zx, and the numbers above include that change.
 
PGHammer said:
In my case, I'm waiting for the UD5H (which, at least, at MicroCenter, replaced the UD4H in their midrange bundle - same price in the bundle, even).

While the dual-gigabit LAN is absolutely overkill, all my OSes (Windows 8.1, Server 2012R2, and OS X ML/Mavericks) can take advantage for dual-gigabit teaming.
Click to expand...

I don't think you can do dual teaming with different Lan chipset.
You can check this with Gigabyte tech support.
 
Both NIC's on the UD5H are Intel and can be teamed to a certain degree.
I have mine set to load balance, team #0.
I've done benches and I don't see any gains with teaming.
I suppose you could use one for fault tolerance but in that case the NIC would simply fail and it's unlikely Intel's will fail.

As to this thread, the UD3H sure seems to be a problematic board if you go by what NE has to say. Most of the issues appear to be the early bios which
appear to be ironed out with F8.

I have a chance to get this board for free but I have no use for it, already have the UD5H plus the Z97 boards will probably be out in less then 6 months.
 
I'm trying to reconcile the widely different DPC latency results that I've noted on the internet. (Other than it's on the internet)

In the Hard Forum test the results were horrible:
1375169267ac4bgSJ7kS_4_2_l.gif

Totally unacceptable.

And yet...over at overclock.net when they tested with the F6 bios they got this:
xjkt.png


Which is the same low latency seen in the Asrock Fatality board.

Is is just a BIOS update or is it a total system/OS thing and it's a YMMV thing?
 
The BIOS and drivers can easily alter the DPC latency results. GIGABYTE often has to release a new UEFI BIOS to correct for poor DPC latency on many models. Most of the ones I test tend to be horrible unless they've had a couple of BIOS revisions so I can't say I'm surprised. Lastly installed hardware makes allot of difference. Though on the test bench they probably don't have anything more than a discreet graphics card installed in the test system and the bare minimum components needed to run the system. At least that's how we do it.
 
I have had this board since the first of the year. I found early on that it seemed to have trouble running ram (only two sticks) at 1600mhz. Verified this with two different sets of Corsair Vengeance ram (CMZ4GX3M1A1600C9 and CMY16GX3M2A1600C9)

I would get a looping sound freeze in HL2 based Day of Defeat Source, that seemed to be resolved if I would change the ram back to 1333mhz instead. Once the F8 bios came out, this seemed to improve a lot (I set the ram back to 1600mhz), but I still get an occasional looping sound freeze out of the blue.

Very curious about this DPC latency talk - I have no idea what it is but will have to research it and run that test on my machine.

FYI all specs in my sig - and nothing is overclocked

EDIT: omg I can't believe I just admitted to not overclocking here... This is the first one since the good ole' Abit BP6 days, I promise!
 
Last edited:
Basically the Deferred Procedure Call (DPC) is something in the Windows OS which allows higher priority tasks to act as interrupt to other lower priority tasks when something needs to be done quickly. Typically device drivers and A/V applications tend to do this. Badly written drivers or UEFI BIOS issues can create poor DPC latency resulting in audio or video drop outs and potential synchronization issues.

Naturally this is undesirable.
 
Thanks for chiming in on the latency call, Dan. I'm late to the Z-87 party, which maybe is a good thing. Do you ever re-test boards when the BIOS has matured? Or should I stop whining about dpc latency and just get the gaudy Killer Fatality board?

I need to build a low latency DAW/video edit system, at a bargain price. Otherwise I'd just get a 4930 x79 system and curse myself in 6 months. My old rig (sub 100us dpc) is failing and need to upgrade now. So it's a 4770K and a mid-priced Z-87 board. Happy to see you guys testing for dpc latency, except for the Asus Z-87-A board. Most sites still don't. And [H] is my fav review site. It's probably just the color scheme, but your tests look good too!
 
clockdogg said:
Thanks for chiming in on the latency call, Dan. I'm late to the Z-87 party, which maybe is a good thing. Do you ever re-test boards when the BIOS has matured? Or should I stop whining about dpc latency and just get the gaudy Killer Fatality board?

I need to build a low latency DAW/video edit system, at a bargain price. Otherwise I'd just get a 4930 x79 system and curse myself in 6 months. My old rig (sub 100us dpc) is failing and need to upgrade now. So it's a 4770K and a mid-priced Z-87 board. Happy to see you guys testing for dpc latency, except for the Asus Z-87-A board. Most sites still don't. And [H] is my fav review site. It's probably just the color scheme, but your tests look good too!
Click to expand...

The Z87-A and a few others predate us actually testing for DPC latency. It was something we started doing at the request of people here on the forums. We do not go back and retest boards after several BIOS revisions. We often find that there is little to be gained by doing so. Rarely does a BIOS update change our opinion on a motherboard. It's often best to look at the forums to see what people are actually getting with such boards in the wild later on down the line.
 
All academic now for those obsolete Z87 boards. They're so last week. ;-)

Have new Z97 boards with new dpc latency issues to look forward to.
 
clockdogg said:
All academic now for those obsolete Z87 boards. They're so last week. ;-)

Have new Z97 boards with new dpc latency issues to look forward to.
Click to expand...

I haven't seen anything from GIGABYTE or MSI as of yet. I know Kyle has but we are still in the dark about DPC latencies on those boards. Though historically MSI has been good and GIGABYTE is hit and miss with that.
 
You must log in or register to reply here.
Back
Top