MSI H110M Pro-D Motherboard Review

[Guest]

Can low-cost motherboards support the full performance of enthusiast CPUs and graphics cards? We tested MSI's fabulously inexpensive H110M Pro-D to find out.

MSI H110M Pro-D Motherboard Review : Read more

 

[jist02]

I am curious about this comment:
"I discovered a couple of things where defaults might not have worked. For example, on one of the screens, Win8.1/10 WHQL Support was disabled, so I turned it on."

Why do you believe this should be enabled? What does it do so that Windows 10 might not have worked without enabling it?

 

[DarkSable]

I'd love some more research into why that performance gap exists - my best theory is that the bios is more rudimentary, and so is perhaps missing power-saving features like Intel speedstep?

 

[joex444]

I've been saying the same thing. The H110 has a PCIe 3.0 x16 slot which is connected to the CPU just like Z170, so why would you expect the H110 to perform any differently? The memory controller is also on the CPU, and the same Intel SATA ports are provided just without the RAID BIOS.

For the large portion of users which build a system with a locked CPU, two memory sticks, a single GPU, an SSD, and a HDD... the H110 meets those requirements. The only downside is the relative lack of USB 3.0 ports, though for most of these users their requirements are 1 keyboard, 1 mouse, and the ability to connect a flash drive or external drive. The former are handled via USB 2.0, leaving two USB 3.0 ports for devices occasionally used.

 

[turkey3_scratch]

I thought it was a GTX 970 that was going to be used, not a GT 740?

 

[Onus]

It was a GTX970; that's an error on the chart.

I'm going to be looking a little more into that performance gap when I look at a couple more H110 boards in the very near future (they're on their way); in fact that's specifically why I wanted at least one more. ASRock has historically offered slightly less performance, but uses less power, so their implementation could be the opposite of what this MSI sample showed (better performance, more power used). A small clock difference might have been there too, and unfortunately I don't think I noted it.
As to the WHQL, I might be concerned that certain drivers might not load or might not be found. If I see something similar to that on another incoming board, I will try to remember to reload it twice, with the settings each way. There's a niche test I'd like to do if I can find a specific product I ordered; a programming cable used with handheld ham radios, of a type known to be problematic due to driver issues.

 

[turkey3_scratch]

I'll look forward to those also! Great review Onus

 

[Valantar]

This was very interesting. I'm looking forward to the follow-up!

On one hand, it's not really surprising that a motherboard with no functional limitations for the relevant use cases and with fewer onboard components should perform as well, if not better than a more fully loaded alternative. After all, more onboard stuff means more work for the system, busier interconnects, and so on.

On the other hand, I wonder if there's a link between the increased power consumption and the increased performance. Might the H110 board somehow be ignoring the power limits of the CPU? Otherwise, shouldn't the power deltas between idle and Prime95 be virtually identical with identical CPUs? Prime95 stresses neither IO, RAM or PCIe, so the list of reasons for increased power consumption outside of the CPU itself is short ...

Skylake ditched the FIVR, right? So, might this board somehow be overvolting the CPU? Not that that should matter for performance unless it's running above stock clocks as well, but I can't make sense of this otherwise.

Did you by any chance log CPU frequencies during benchmark runs?

 

[littleleo]

I am curious about this comment:
"I discovered a couple of things where defaults might not have worked. For example, on one of the screens, Win8.1/10 WHQL Support was disabled, so I turned it on."

Why do you believe this should be enabled? What does it do so that Windows 10 might not have worked without enabling it?

WHQL is the Acronym for Windows Hardware Quality Labs. Where hardware and drivers are tested for Windows compatibility. For this motherboard I believe that is to enable full UEFI support. Why it is listed like that on MSI motherboards I have no idea.

 

[littleleo]

I love this, this was a great idea, and I'm really looking forward to more comparisons based on the various chipsets. I think it could be eye opening. I wonder if the additional features the feature rich chipsets like the H170 and Z170 slows down the performance on the board? Next time perhaps instead of DX9 you can use DX12.

Ideally when I am choosing a motherboard I 1st find out the CPU the customer wants to run. Then I get the list of the features the customer has to have, then the additional features he would like to have on top of that. Then I would choose the chipset has those features and look at that class of motherboard with that chipset. Then I would find the models closest to the price target of the customer. Then if there is a brand preference that should reduce the options to 1-3 models. Then the customer can choose the final model and go from there. Of course in the real world what usually happens is the guy pops in and says I have this much what can I get, lol.

 

[Onus]

I'll take that as a complement, so thank you.
Iirc, the Heaven Benchmark, even on Extreme, only went up to DX11, although might that also be a Windows 8 limitation? The whole team is moving to Windows 10 though, and adding some other updates to our testing. We each add our own flavor of course, for example using our own equipment for certain tests (e.g. the GTX970, and my LAN throughput using Igor as my target), but there is commonality to make some comparisons possible across our articles.

 

[bit_user]

I find the degree of difference to be weird and a bit disturbing. 27% better DX9 performance? That's news! How did the Prime95 performance compare?

As others have speculated, the power & performance discrepancies might be related. Many BIOS I've seen have a setting to control how fast or efficient the system is (I think this might control things like how quickly the CPU turbo boosts and how long it holds). In others, you can directly disable Speed Step. Maybe the power/efficiency defaults of these boards have something to do with these results?

 

[gondor]

WHQL is the Acronym for Windows Hardware Quality Labs. Where hardware and drivers are tested for Windows compatibility. For this motherboard I believe that is to enable full UEFI support. Why it is listed like that on MSI motherboards I have no idea.

My Gigabyte (Z170 based) board has similar option and it doesn't appear to have anything to do with UEFI (there are different options for that).

I'm running Windows 7, although I did upgarde to Windows 10 just to claim the license and then immediately revert back (thanks Clonezilla!).

 

[kancaras]

How can CPU temperature differ so much??????? It looks unbelievable.

 

[turkey3_scratch]

How can CPU temperature differ so much??????? It looks unbelievable.

I think that's motherboard temp (something on it), but if not, it would be because the CPU was processing more stuff more rigorously, which could also relate to the better performance in some areas.

 

[bit_user]

How can CPU temperature differ so much??????? It looks unbelievable.

I think that's motherboard temp (something on it), but if not, it would be because the CPU was processing more stuff more rigorously, which could also relate to the better performance in some areas.

The chart is labelled CPU temperature.

Assuming the cooler is the same, then the difference could depend on how the motherboard's BIOS handles power management (or does Windows override that?) and how much airflow the physical layout of the motherboard (and placement of the GFX card) allows. I don't know if its VRM could also affect CPU temps. There's also some room for testing error, in the sense that it's hard to install a heatsink exactly the same, in both machines.

This needs to be investigated.

 

[turkey3_scratch]

How can CPU temperature differ so much??????? It looks unbelievable.

I think that's motherboard temp (something on it), but if not, it would be because the CPU was processing more stuff more rigorously, which could also relate to the better performance in some areas.

The chart is labelled CPU temperature.

Assuming the cooler is the same, then the difference could depend on how the motherboard's BIOS handles power management (or does Windows override that?) and how much airflow the physical layout of the motherboard (and placement of the GFX card) allows. I don't know if its VRM could also affect CPU temps. There's also some room for testing error, in the sense that it's hard to install a heatsink exactly the same, in both machines.

This needs to be investigated.

I can't see the motherboard layout having any more than a negligible effect on CPU temps. Motherboards also use very little power so it's highly doubtful that waste heat from the VRMs and chipset could be heating up the CPU. The only reason the CPU would be that much hotter is if the CPU is doing more work on this motherboard, thereby making it hotter.

In the power consumption tests also, you could see the MSI measured from the wall used a lot more power. I'm questioning if this is actually from the motherboard having higher power consumption or if it's from the CPU working harder, which also would explain the higher performance somewhat.

 

[Onus]

I've used the same cooler, and the same thermal paste (that came with the cooler), on all my reviews, placing a small dab of paste in the center of the CPU and letting the cooler spread it. My methods have been consistent, but a sample-size of one is not going to be entirely reliable.
It is worth noting that I am now out of that thermal paste (after the latest two, yet to appear, but in the queue), so future results won't be directly comparable. I'll be using the same cooler, but will have to switch to AS5, as I have a large tube of that now.

 

[bit_user]

I've used the same cooler, and the same thermal paste (that came with the cooler), on all my reviews, placing a small dab of paste in the center of the CPU and letting the cooler spread it. My methods have been consistent, but a sample-size of one is not going to be entirely reliable.

Thanks for the info. It's good to know the cooler is the same unit. That eliminates one source of differences.

I wish we could know whether the BIOS' power management settings were respected by Windows, or whether Windows' power management schemes override the BIOS settings. To me, that's the best potential explanation.

 

[jrhansen]

Nice test.
IMO, besides what others noted about performance/temperature readings, it actually seems very likely that the simpler 110 Board would run a small amount faster if the 2 chipsets have about the same theoretically performance. Having more features, like raid, Multiple LANs, m2, multiple USB and raids wich is derived from their own controller chips WILL affect performance in a bad way if not used with caution. They are ALL connected to the same busses, wich will create more traffic and will cause some kind of performance degradation of the combined motherboard.
Having a ton of features that you don't use is both a waste of money and a much more complex MB might actually slow things down.
It's the same reason that a true game enthusiast also WILL have a boot setting in Windows where 80% of the standard stuff windows runs in background is turned off so you don't waste valuable CPU performance that might help the game run smoother

 

[bit_user]

Having more features, like raid, Multiple LANs, m2, multiple USB and raids wich is derived from their own controller chips WILL affect performance in a bad way if not used with caution. They are ALL connected to the same busses, wich will create more traffic and will cause some kind of performance degradation of the combined motherboard.

The term "bus" no longer really applies to the modern PC. The only thing I think really deserves the name "bus" is memory, where you can have more than one DIMM per channel. Everything else is point-to-point. Even USB isn't very bus-like, if you're not using hubs.

For a more detailed explanation, check out this article: http://www.tomshardware.com/reviews/intel-skylake-processors-101,4498.html#p3

For a more general example, you could read about PCIe: https://en.wikipedia.org/wiki/PCI_Express

Even with a true bus, merely having an inactive device on the bus doesn't really add any overhead. It does consume more power, since all devices are continually watching the bus to determine when they are the target of a given access.

Anyway, I guess Thunderbolt is another example of a true bus, since you can daisy chain devices. I don't know much about it, to be honest.

 

[Valentin_5]

Wattage and temp are a huge deal. I don't get it, why it has "performance" at "pros"? The performance is minimal.

 

[Onus]

Its performance matches (or slightly exceeds) that of much more expensive boards, so I consider that a "pro."