In what way does the increase in PCI lanes stop holding gpu manufacturers back? To the best of my knowledge a single GPU doesn't use more than 16 lanes.
Only multi-gpu systems are held back by the PCI lanes. So I can see threadripper being great for supercomputers, but I don't see how it affects Nvidia or AMDs GPU section.
Currently it seems 8 lanes of PCIe 3.0 is not quite enough, but 16 lanes is plenty. So we'll need a few years to get up to utilizing 16 lanes completely.
At which point 4.0 may be ready.
You mean like having a GPU and one or more PCIE NVMe storage-devices?
Typically you get something like 3 full length PCIe slotson a xeon e3 board (c236 chipset). BUT they can't all run full speed at 16x (16/16/16). You have to then choose between 16/8/8 and 16/16/0 due to the lack of PCI lanes.
https://www.youtube.com/watch?time_continue=2&v=8KOTiee5RqA
on top of that, many people want to stream their game video when they play their games on full speed, then a couple NVMe SSDs eat those lanes as well.
More is definitely better here.
If they are screwed it's by their own choice to abandon desktop and server CPUs to bet all on mobile.
Also, while Intel maybe winning the mobile (really only laptop) market now, ARM is slowly catching up. It's really only a matter of years before say someone like Apple chooses to ditch Intel in favor of using their own A processors in their computers.
Intel's strategy of limiting the PCI throughput to hold GPU manufactures back is over, and these AMD cpu's are going to be paired super well with GPU makers, mostly Nvidia but of course their own ATI, which is really going to making Intel look sad soon. Boat loads of major players have been irked by Intel holding back PCI throughput, AMD let it rip, and "thread ripped" too!