XIII Hero - PCIE16_1 in 8x

[Zeroed85]

My GPU in the PCIE16_1 slot is running at 8x. I thought perhaps my PCIE 1x sound card occupying PCIE16_2 was causing this but after removing the sound card the PCIE16_1 slot is still in 8x. I have just one NVMe occupying the M.2_4 slot but no where in the manual does it say M.2_4 will share bandwidth with the PCIE16_1 slot.

Can someone from Asus clarify? Why is my PCIE16_1 slot in 8x with just one NVMe plugged in? Please note I am using a 10th gen 10900K but the manual clearly states that 10th gen will run the PCIE16_1 slot in 16x mode.

[Saltgrass]

Zeroed85 wrote:
Well the spec page clearly states M.2_2 shares bandwidth with PCIX16_1/2 so I goofed up. I don't understand though, why would anyone want to use M.2_2 when it half's the bandwidth to the GPU?

It's clear that RKL + PCIE 4.0 has been the focus here but with CML (which the platform supports) the storage and PCIE lane options are severely limited.

If I am reading the specs correctly, the 11th Gen processor has 4 more lanes available than the 10th Gen. Since the Motherboard was designed around that processor, it would seem logical a 10th Gen processor system would have some limitations.

I ran the 11th Gen processor on the Z490. I was not happy with that combination either.

[EncodeGR]

Zeroed85 wrote:
Well the spec page clearly states M.2_2 shares bandwidth with PCIX16_1/2 so I goofed up. I don't understand though, why would anyone want to use M.2_2 when it half's the bandwidth to the GPU?

It's clear that RKL + PCIE 4.0 has been the focus here but with CML (which the platform supports) the storage and PCIE lane options are severely limited.

Both M.2_1 and M.2_2 are CPU bound. So, if your CPU has only 16 lanes and want your PCIE16_1 to run at x16, you shouldn't use M.2_1 or M.2_2.
RKL on the other hand has 20 lanes, which lets you use both PCIE_1 and M.2_1 (or M.2_2), again bound to the CPU.

This board is build for RKL. It does support CML, but it wasn't designed with that in mind, it's just backwards compatible. You shouldn't buy such a board to use it with non-RKL CPUs, because you would be wasting a lot of its features while paying premium price.

Zeroed85 wrote:
I don't understand though, why would anyone want to use M.2_2 when it half's the bandwidth to the GPU?

Because there is minimal difference in performance ( ~ 1-2%) when going from x16 to x8 in GPU, especially when you consider the PCIe 4.0 speeds. So, some people may preffer having two M.2 Nvme drives bound to the CPU, and x8 GPU, all of these with PCIE 4.0 speeds. It's just preferrence.