• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

Dmi 2.0

Bearmann

Member
In an article on the differences in the various Skylake chipsets, Matt Bach at Puget Systems writes:

In addition to having fewer and slower PCIe lanes, H110 also still uses the older DMI 2.0 revision which means the connection between the chipset and the CPU is a bit slower than it is on the other chipsets.

The accompanying table lists Z170, H170, and B150 all as DMI3 (8GT/S) and H110 as DMI2 (5GT/S).

What is the practical significance of this? Would an H110 board always run slower than the other boards noted above? Should you avoid an H110 board if you want the best performance?
 
What is the practical significance of this? Would an H110 board always run slower than the other boards noted above? Should you avoid an H110 board if you want the best performance?

It just means that unlike other Skylake chipsets, it only has 2000MB/s (practical limit ~1600MB/s) of bandwidth between CPU and PCH, instead of ~4000MB/s (practical limit ~3200MB/s).

So if you want maximum performance with your shiny new PCIe SSD, you should avoid H110 boards. But that really shouldn't be an issue when on a budget. Otherwise performance is completely identical.
 
It just means that unlike other Skylake chipsets, it only has 2000MB/s (practical limit ~1600MB/s) of bandwidth between CPU and PCH, instead of ~4000MB/s (practical limit ~3200MB/s).

So if you want maximum performance with your shiny new PCIe SSD, you should avoid H110 boards. But that really shouldn't be an issue when on a budget. Otherwise performance is completely identical.

Would it also affect something like putting a GTX 980ti in an H110 MB?
 
So the PCIe SSD has to go through the PCH, but the GPU does not?

Yup. Since H110 doesn't allow splitting of the lanes coming from the CPU, you can't even hook up the SSD to them either.

If you're certain you'll get a PCIe SSD at some point, I'd suggest looking for a B150-based mainboard. They're not really that much more expensive, if at all. If you want I can suggest a few boards... 🙂

Would it also affect M.2 and U.2 SSD's?

M.2 and U.2 (SFF-8639) are just the physical interfaces used for PCIe SSDs. They work the exact same whichever interface you use. This is also why you can use a simple passive adaptor for converting one for the other.
 
I'm not in the market for a MB. I'm just trying to have a better understanding. I appreciate your patience, knowledge, and helpfulness.
 
Another question. Is DMI 2.0 referring to the same thing as PCIe 2.0, and DMI 3.0 alligned with PCIe 3.0?

Both yes and no. The DMI physical interface is heavily based on PCIe technology. They're not identical, but very close to it. The main reason you're not seeing PCIe 3.0 on a DMI 2.0 PCH, is simply that there isn't enough bandwidth to handle more then 2x PCIe 3.0 x1 ports.
 
Back
Top