Normal
Running the slots @ PCIe3 at x8 electrically was my implication, as we know that there is no meaningful performance difference between x16 and x8 at gen3 speeds on current GPUs.So, a mATX board with simultaneous support for 3 x8 slots (x16 physical) and 1 (or more) x4 M.2/U.2 devices supporting NVMe devices, and mITX/mDTX with simultaneous support for at least 2 PCIe devices (possibly via bifurcation) at x8 electrically and 1 (or more) x4 M.2/U.2 devices is entirely doable (on paper).Even if consumers wouldn't initially snatch these up (assuming they come to market and aren't borked in some other way, desigh-wise), system integrators and boutiques will, and will rapidly school the mainstream about how to build truly amazing systems with "only" x8 gen3 support.
Running the slots @ PCIe3 at x8 electrically was my implication, as we know that there is no meaningful performance difference between x16 and x8 at gen3 speeds on current GPUs.
So, a mATX board with simultaneous support for 3 x8 slots (x16 physical) and 1 (or more) x4 M.2/U.2 devices supporting NVMe devices, and mITX/mDTX with simultaneous support for at least 2 PCIe devices (possibly via bifurcation) at x8 electrically and 1 (or more) x4 M.2/U.2 devices is entirely doable (on paper).
Even if consumers wouldn't initially snatch these up (assuming they come to market and aren't borked in some other way, desigh-wise), system integrators and boutiques will, and will rapidly school the mainstream about how to build truly amazing systems with "only" x8 gen3 support.
