ASUS Maximus VIII Impact LGA 1151 Motherboard Review @ [H]

FrgMstr

Just Plain Mean
Staff member
Joined
May 18, 1997
Messages
55,635
ASUS Maximus VIII Impact LGA 1151 Motherboard Review - Let’s face it, most sequels are never as good as the original films. When it comes to motherboards the opposite usually rings true. ASUS’ Republic of Gamers brand has had several iterations of the mini-ITX Maximus Impact, today we have version VIII. Hopefully ASUS continues its track record of great sequels.
 
Not one word about the non existing M.2 port anywhere in the review. How can you guys ignore such an essential feature missing on this mini-itx board without even touching on it is beyond me.
 
Two points:

1) Some believe U.2 is the way forward, not M.2. Asus are clearly one such entity, based on the Impact, Extreme, WS, and Hero/Alpha boards. All have U.2. Some have two.

2) The Asus Z170i Pro Gaming has M.2 and lacks the nonsensical audio section of the Impact. It's also about $85 cheaper.
 
Last edited:
personally I think having the U.2 port is a benefit over having a M.2
Having a U.2 port means you can use a Intel 750 1.2TB drive. With M.2 you will have to settle for just 512GB.

I could be mistaken. I haven't been keeping up on bleeding edge like I should.
 
Every other Skylake mini-itx board have m.2 slot, yet no mention of this missing in the "review". There is no way around a Samsung 950 NVMe at the moment and with mini-itx pushing size limits, small m.2 drives are currently is the way to go. Asus bet on the wrong port for this mini-itx board and every other review but this, at least touch on the topic. You can convert m.2 into u.2 with a small converter card, but not the other way around because of the missing power. Leaving this huge drawback out of the discussion really makes you wonder.
 
Last edited:
Maybe this is just new tech to HardOCP so it wasn't touched on, tho I doubt it. I am not sure why it was not mentioned. I've never heard of U.2. I am assuming it's a replacement for M.2? This could such a new tech that's it's also new to HardOCP ( myself included ) they were unfamiliar and they didn't feel comfortable talking about it in a review. Maybe an update is coming?

I think some people may mistake the absence of this not being mentioned as those people assuming it does have the M.2 port. I could see that easily happening.

Normally the reviews here are very in-depth and all inclusive with no stone being un-turned. I am sure this is just a simple mistake. During the drafting process and re-write, final review, it was accidentally omitted / over-looked.

Also, remember, times are tough and corners get cut. It just happens these days. They had a dead-line to meet possibly.

At any-rate, very thankful for the review, finished or not, errors or not. Thank you.
 
Last edited:
Maybe this is just new tech to HardOCP so it wasn't touched on, tho I doubt it. I am not sure why it was not mentioned. I've never heard of U.2. I am assuming it's a replacement for M.2?

U.2 is the current name for what was previously SFF-8639, it is a confusing story. Personally I thought U.2 was going to be the replacement for SATA Express since it failed to be useful. M.2 seemed like it was targeted for small devices, but then it started showing up on full size motherboards.

This is all new stuff still so i think manufacturers are not clear on what they should make or what people want. I am sure we could have a very long discussion about M.2 vs U.2, but the fact is they just ignored something that jumped out to many of us.

That said I have noticed some reviews here where [H] is only reporting on what a device does and ignoring the what it should do , but doesn't.

I want more than that, I want them to tell me that while this board might do everything right it s lack of M.2 is a huge mistake or if they think the lack of M.2 is no big deal then say that and tell us why you think that.
 
Lovely review. I wonder why they use regular DIMMs? Surely SODIMMs would be a better physical fit? Or would there be a performance penalty?
 
MY biggest concern is that no one has mentioned the sound issues the asus rog line is having. there are literally hundreds of threads on problem with the sound. I currently have a maximus hero VIII and I have swapped it three times and rebuilt the whole os twice and still get the same issue after the motherboard has been left to run for any amount of time. They market that the sound on these boards is supposed to be outstanding but it is anything but. A little sad that Kyle and hardocp haven't really touched on this issue. I made a choice to buy the 170 based asus line due to reviews here and I find myself pretty let down by Asus.
 
Not one word about the non existing M.2 port anywhere in the review. How can you guys ignore such an essential feature missing on this mini-itx board without even touching on it is beyond me.

It's an oversight on my part. It's one of the things I had in my notes to go over, but for whatever reason I didn't include that.

Two points:

1) Some believe U.2 is the way forward, not U.2. Asus are clearly one such entity, based on the Impact, Extreme, WS, and Hero/Alpha boards. All have U.2. Some have two.

2) The Asus Z170i Pro Gaming has M.2 and lacks the nonsensical audio section of the Impact. It's also about $85 cheaper.

U.2 is great, there is no doubt about it. It should be the way forward but it's too early to tell. The problem is that the only consumer drives I know of that use it are the Intel SSD 750's. This forces the user into using the Intel SSD 750 if they want a PCIe based SSD drive and don't want to give up their PCIe x16 slot for it. Sure you can do a riser or something like that to split the slot if your case has room for it and you can get both devices to play nice doing that.

personally I think having the U.2 port is a benefit over having a M.2
Having a U.2 port means you can use a Intel 750 1.2TB drive. With M.2 you will have to settle for just 512GB.

I could be mistaken. I haven't been keeping up on bleeding edge like I should.

In terms of capacity it seems like the right choice. But again you are limited to the Intel SSD 750 at this point. The Intel 750 is going to cost a lot more than the Samsung 950 Pro in all capacities but the 400GB model. 400GB is a little anemic capacity wise. There is no doubt its fast but the Samsung 950 Pro is faster in some workloads or usage scenarios.

Every other Skylake mini-itx board have m.2 slot, yet no mention of this missing in the "review". There is no way around a Samsung 950 NVMe at the moment and with mini-itx pushing size limits, small m.2 drives are currently is the way to go. Asus bet on the wrong port for this mini-itx board and every other review but this, at least touch on the topic. You can convert m.2 into u.2 with a small converter card, but not the other way around because of the missing power. Leaving this huge drawback out of the discussion really makes you wonder.

ASUS made a trade off. With mini-ITX that's what engineers have to do. The Maximus VIII Impact currently has the most capable VRM design of any mITX motherboard and is easily one of the best overclockers. And I don't just mean today, but in the long run it will probably prove to be the case. Additionally you have a better sound solution than the standard Realtek ALC1150 options out there. ASUS included U.2 thinking that should be enough and that this solution was overall the best combination of features for their target market.

There is a solution to this which just occurred to me. ASUS could have gone for a vertical M.2 slot the way they did on the X99 Deluxe. That would seem to be the best way to go here.

Maybe this is just new tech to HardOCP so it wasn't touched on, tho I doubt it. I am not sure why it was not mentioned. I've never heard of U.2. I am assuming it's a replacement for M.2? This could such a new tech that's it's also new to HardOCP ( myself included ) they were unfamiliar and they didn't feel comfortable talking about it in a review. Maybe an update is coming?

I think some people may mistake the absence of this not being mentioned as those people assuming it does have the M.2 port. I could see that easily happening.

Normally the reviews here are very in-depth and all inclusive with no stone being un-turned. I am sure this is just a simple mistake. During the drafting process and re-write, final review, it was accidentally omitted / over-looked.

Also, remember, times are tough and corners get cut. It just happens these days. They had a dead-line to meet possibly.

At any-rate, very thankful for the review, finished or not, errors or not. Thank you.

It's something I intended to touch on and not including that is a mistake on my part. U.2 isn't a replacement for M.2, but rather an alternative to it. It's the consumer name for the SFF-8639 connector which has been used in the server market for some time now. With server OEMs like HP and Dell flexibility and scalability are the name of the game. OEMs basically use the same chassis / case for as many systems as possible. OEMs can offer any combination of 2.5" drives in SATA, SAS or PCIe, or PCIe via SFF-8639 connectors. The SFF-8639 was renamed "U.2" by Intel and it's an interface that ASUS has embraced with open arms. In this case at the expense of M.2.

ASUS may know something we don't and more manufacturers are planning to use this connector. I hope they do as M.2 is limiting. Depending on where the M.2 slots are your SSD can get baked by GPUs and M.2 takes up a lot of real estate on the PCB.

Lovely review. I wonder why they use regular DIMMs? Surely SODIMMs would be a better physical fit? Or would there be a performance penalty?

SODIMMS would fit better but they don't make SODIMMS that match the speeds of the desktop parts. Desktop DIMMs are used to give this motherboard the ability to match motherboards like the Maximus VIII Extreme's performance.

MY biggest concern is that no one has mentioned the sound issues the asus rog line is having. there are literally hundreds of threads on problem with the sound. I currently have a maximus hero VIII and I have swapped it three times and rebuilt the whole os twice and still get the same issue after the motherboard has been left to run for any amount of time. They market that the sound on these boards is supposed to be outstanding but it is anything but. A little sad that Kyle and hardocp haven't really touched on this issue. I made a choice to buy the 170 based asus line due to reviews here and I find myself pretty let down by Asus.

I haven't touched on the issue because I haven't experienced it. There are tons of problem threads posted on forums specifically about ASUS motherboards because ASUS has sells far more motherboards than most manufacturers do. I see the same thing about GIGABYTE, ASRock and MSI motherboards. Forum posters, regardless of how many never are representative of the whole. I'm sure there are several people without issues for every one having problems. I have 3x ROG boards running in my systems at home without any audio issues or problems of any kind. I've tested nearly the entire ROG line since it's inception. All of the problems that I experienced with the Maximus VIII Impact were mentioned in the review. These all boiled down to DPC latency due to the Sonic Studio software. Once that was disabled everything worked fine.

I haven't experienced poor customer service from ASUS either despite tons of forum members claiming otherwise. I acknowledge that these problems may exist, but until I experience them first hand I don't have anything to add to the topic. It's the same with audio. Until I experience such problems there isn't anything for me to "touch on."

I'm human, I make mistakes. For whatever reason I didn't touch on the M.2 slot being absent on this motherboard. In part it's something that I don't specifically test as I don't have an M.2 drive for that as of yet. As a result I glossed over it. I will make sure this is addressed every time in the future. One final thought on the subject: M.2 may seem ubiquitous to a lot of you but remember, last generation there were several motherboards that didn't have it. Many of them traded SATA Express for it or vice versa.
 
Last edited:
Thanks for the response Dan I have been trying everything I can think of to resolve the issue with my maximus heroViii. This is the first time I have heard anyone mention dpc latency. I may pm you later about this as honestly with three boards swapped out it has to be a driver implementation issue. I agree they are the biggest seller of higher end boards so yes it makes sense that there would be more issues due to volume alone. I will say the vast majority of the sound complaints I have read seem to be on the hero line so not sure why that is the case.
Thanks over all for you response in regards to U.2 port. Honestly I don't see what the fuss is about as it hasn't really become a requirement yet.
 
Great review, and I'm considering a big upgrade for my mitx, Corsair 250D system. This might be my go-to.

Question: What is the weird port in between the riser sound card and the wifi antenna anchors in the top of this image
14489387580Cy8IdlYi8_1_11_l.jpg

?
 
If I were a betting man......I would bet on U.2/SFF-8639 coming out on top here.
 
Great article as always. Also kudos to Dan for quickly addressing the M.2/U.2 questions.

I'm not sure if this has been covered before: wired network numbers look pretty slow to me, is it being limited by the test "server" machine? Would it be feasible to setup a RAM disk or SSD on that side to discard a bottleneck that could be distorting those results?
 
There may be some limitations there but for the most part we are following the testing methods recommended by the motherboard manufacturer. The other system is still PCIe 2.0 but I'm using an actual server adapter and not the onboard controller. I could go with a PCIe 3.0 system and a new server controller which is an eventuality for me but I don't know when. As for the software side that system runs Windows Server 2008 R2 Enterprise. I'd have to switch to Windows 10 Pro or Windows 2012 R2 for a better driver / TCP/IP stack.

A CAT6 Crossover cable is used to eliminate the bottleneck of a switch. I can run some tests and see what if anything I can do about that. If say the test is faster going to my personal workstation then I'll look at upgrading that box. If it isn't I'll stick with what I've got. I'll post the results here and let you know how it goes.

Seeing how careful is your LAN setup I'm probably asking something you have already regarded. Anyway, my point was more towards the storage setup in the server machine. If you're using a single (non-RAID) HDD that could be the limiting factor. Anything else would easily saturate a Gb-E connection. Aside from that it seems you have a pretty robust NIC/OS combination, well above the typical user so that shouldn't be a concern :)

Again, thanks for addressing our questions, really appreciated!
 
Seeing how careful is your LAN setup I'm probably asking something you have already regarded. Anyway, my point was more towards the storage setup in the server machine. If you're using a single (non-RAID) HDD that could be the limiting factor. Anything else would easily saturate a Gb-E connection. Aside from that it seems you have a pretty robust NIC/OS combination, well above the typical user so that shouldn't be a concern :)

Again, thanks for addressing our questions, really appreciated!

I hadn't given much thought to the drive configuration. The test files reside on the H drive which I believe is a RAID 10 setup on an LSI PCIe SATA/SAS controller. Its an 8308ELP if I recall correctly. As for the specific drives, I do not recall their models off hand. I believe they are 1TB, but I do not recall what they are. They should be 7200 RPM drives at the very least.

That box is typically where my old hardware goes.
 
I hadn't given much thought to the drive configuration. The test files reside on the H drive which I believe is a RAID 10 setup on an LSI PCIe SATA/SAS controller. Its an 8308ELP if I recall correctly. As for the specific drives, I do not recall their models off hand. I believe they are 1TB, but I do not recall what they are. They should be 7200 RPM drives at the very least.

That box is typically where my old hardware goes.

That certainly eliminates any possibility of the storage side limiting performance. Even the oldest 1TB drives should each easily manage 100MB/s sequential reads. I guess it's all about the specific benchmark test.

And yeah, old hardware either goes to the server or relatives' systems :D
 
I don't buy into the argument that M.2 needs much more space than U.2. Almost all other boards have the M.2 port in i otherwise unused position on the back of the PCB.


U.2 most likely winning you say? A quick look on Newegg shows 192 SSD's with m.2 connector and 3 with U.2. Not winning any time soon. There is currently no way around Samsung 950 Pro over Intel 750 on price if you want a NVMe drive.


Not addressed in the review. Should the review not be updated to address the one major weakness of the board (the "oversight" as you put it) as the review is now being linked to from all the usual tech websites? Would make the review seem less biased.

The M.2 slot doesn't take up much more space on the PCB but the overall slot sure as hell does. When the drive is actually mounted it can take up quite a bit of space. This is fine as you said in most cases as it usually goes over unused portions of the PCB. However, on some motherboards that's not possible. The X99 Deluxe is one example. One of its slots is conventional while the other pretty much had to be vertical. In the case of the Maximus VIII Impact there wasn't room for it. A vertical slot would have been the only way to get one on board. I don't know why ASUS didn't do that, but they should have.

As for the article, that's up to the boss man.
 
I am starting to pay attention more to these smaller boards. So this review is important to me.

My goal for 2016 is to build an i5 Skylake with ddr4, 2 channel memory is fine, of course with the new Nvidia pascal architecture coming out I want that card. Also, I am totally and forever, completely done with SLI / Xfire. My experience with this single 980 ti card has been beautiful. I mean, I love the consistent daily headache free experience of owning one video card.

So I want a small form factor system, no optical, no onboard audio ( outside DAC ) and I finally want to use a M.2 / U.2 drive. Something bare metal, compact and fast.

My days of huge D900 / Lian Li cases are gone. My D900 literally weighed around 65lbs fully loaded.
 
I am starting to pay attention more to these smaller boards. So this review is important to me.

My goal for 2016 is to build an i5 Skylake with ddr4, 2 channel memory is fine, of course with the new Nvidia pascal architecture coming out I want that card. Also, I am totally and forever, completely done with SLI / Xfire. My experience with this single 980 ti card has been beautiful. I mean, I love the consistent daily headache free experience of owning one video card.

So I want a small form factor system, no optical, no onboard audio ( outside DAC ) and I finally want to use a M.2 / U.2 drive. Something bare metal, compact and fast.

My days of huge D900 / Lian Li cases are gone. My D900 literally weighed around 65lbs fully loaded.


I hear you there. I've had many full-size cases for about 10years, and I just recently went with a mini-itx case, and I won't go back. I'm also not swapping/testing parts like I used to, but since I change parts less frequently. It's been very nice to keep my desktop on my desktop, without losing a ton of desk real estate.
 
I am an owner of this impact board, and its a piece of arts really. Any mini itx to me is piece of art.

Well regarding m2/u2 Asus has addresses this issue for people who wants M.2 by making a converter from U.2 to M.2
source here in the comment section of this video
https://www.youtube.com/watch?v=oKqybmuZK4I

Why ASUS didnt implement M.2 you say?
well its because M.2 generates alot of heat which is not aplicable on a mini ITX

here is what ASUS Said:
Asus speaking to bit-tech on the Impact's use of U.2 over M.2:

Besides the PCB space and potential thermal issues of M.2 on ITX (on a high-end OC board), and that U.2 drives are easy to install / upgrade in an ITX case (especially once built), U.2 is also not constrained by the 25W limit of M.2.

For a high-end, future-looking platform, M.2 is also limiting in terms of drive capacities. If you look at the 950 Pro, for example, the backside of the PCB is empty. And while you could potentially create M.2s with NAND on the backside, you are then compounding the thermal issues (again, especially relevant for ITX). As such M.2s will need to wait for 3rd Gen V-NAND to prevent having to populate both sides, and that will be expensive.

The U.2 format, since it connects to a regular SSD format, allows for faster and larger capacity drives much sooner. It’s likely Samsung may even have to adopt the format for these reasons. And since ROG is always about providing the fastest and most future-looking formats, all these reasons combined pointed to using U.2 and not M.2.

Source:
http://www.bit-tech.net/hardware/2015/10/23/asus-maximus-viii-impact-review/8

as for O.C. its a beast motherboard with so low voltages and so many energy saving potential up to C8 state. also including C7s state which lower USB connected devices voltages
9sCrGpA.jpg
 
As for the article, that's up to the boss man.

I think DrRetina's request is fair. Not everyone reading the article is coming here to discuss the issue. Am I wrong or is it common policy at [H] not to update published articles? IMHO fixing errors/omissions strengthens the site's image, not the contrary. Personally I find Dan's humble and transparent approach to the M.2/U.2 thing (in the forum) very assuring.

(It's late and I'm not 100% all the above was correct english, I hope it made sense)
 
I would agree, but at the same time it really does seem like they are just throwing technology out there randomly. M.2 makes perfect sense for a laptop or smaller device, but for any size desktop computer you will always have room for an SSD.

The major advantage of M.2 is that it requires no additional cabling for power or data. This can be excellent in SFF rigs either to tidy things up, or for actual thermal concerns that can be caused by cable clutter in incredibly small enclosures. In cases like the NCASE M1, NFC S4 or Dan A4-SFX, it can provide varying benefit.

U.2 most likely winning you say? A quick look on Newegg shows 192 SSD's with m.2 connector and 3 with U.2. Not winning any time soon. There is currently no way around Samsung 950 Pro over Intel 750 on price if you want a NVMe drive.

Those 192 options would be varying length, socket and key types. The specific M.2 length (2280), Socket type (3), and Key type (M) has far less options available, in either AHCI or NVMe. This combination is the only possible contender to U.2 for performance, but has the caveat of thermal concerns in sustained performance metrics.
 
U2 already won in the server market. Who cares what Newegg shows. That is the stupidest thing I have ever read. In fact the M2 standard was always stupid to me. Give me U2 all day everyday.
M2 is subject to high heat, problematic placement and more. Its just a stupid part of a mainboard to put something as sensitive as data on.
 
U2 already won in the server market. Who cares what Newegg shows. That is the stupidest thing I have ever read. In fact the M2 standard was always stupid to me. Give me U2 all day everyday.
M2 is subject to high heat, problematic placement and more. Its just a stupid part of a mainboard to put something as sensitive as data on.

While M.2 may be dominant in the consumer space now, I'm thinking U.2 may win out eventually because it allows server oriented drives to be reconfigured and sold as consumer products ala the Intel SSD 750 series.
 
Back
Top