My question might be confusing a bit. TL;DR: can you test how to get max storage performance on Ryzen 3000 + x570 ? At least two prosumer 3000+/3000+ NVMe SSDs.
I am a code creator - do not confuse me with buzzwordy content creator. I do require a lot of CPU and RAM but unlike content creator I prefer storage performance. I am waiting for x570 (and maybe x590? ?550?). You will be among first to test new motherboards. I have read that PCI split on Ryzen 3000 is following: 16 for GPU, 4 for NVMe storage directly from CPU and 4 to chipset to 24 in total. I would like to build workstation in following configuration: GPU (current rx580, maybe something faster in the future), 2 x Samsung pro/WD 750/other 3000/3000+ NVMe (so PCI 3.0 x4). Will there be a MB that splits 4 lane PCI 4.0 directly from CPU into 2 slots 2 lane 4.0 - which as I understand is backwards compatible and should equal to 4 lane PCI 3.0? Or maybe split GPU slot into 2 x 8 lane PCI 4.0 so that I can put GPU into one slot and additional card for NVMe ? 8 lane 4.0 is the same bandwidth as 16 lane 3.0? so with bifurcation I could use ASUS Hyper M.2 X16 Card and run even more NVMe drives? I would prefer to leave PCIe slots available for future - 10GBps NIC, maybe some FPGA or TensorFlow accelerator etc.
Thank you and have a nice day.