Am I correct that if there is no announcement of PCIe 4.0 compatible components (CPU, GPU, motherboard) at CES 2018, there is a very very low chance that they will be available for consumers to make high end PC this year?
By "ready by the end of the year", do you mean available to purchase by the end of 2018?
I mean ready as in finalized by the PCIe-SIG so device manufacturers can finalize their own 5.0 devices' design and get ready to launch stuff.
Put yourself in the device manufacturer's shoes: PCIe 3.0 has been around for five years, PCIe 4.0 don't have anything to plug into yet even if they existed and by the time the first servers with PCIe 4.0 launch, PCIe 5.0 may be less than a year away. Would you bother spending tens of millions in compliance testing equipment and R&D for a standard that still fails to deliver the bandwidth current high-bandwidth devices require (mainly NVMe SSDs) and will be obsolete after perhaps a year of production...
PCIe 5.0 may be ready by the end of the year, I wouldn't be too surprised if some manufacturers end up skipping straight to 5.0 next year instead of launching 4.0 in 2018.
Yea especially the only reason reason for a faster PCIe is the M.2 slots for NVMes. If they were PCIe Slot SSD's they can utilize all 16 lanes if they wanted to
PCIe 5.0 may be ready by the end of the year, I wouldn't be too surprised if some manufacturers end up skipping straight to 5.0 next year instead of launching 4.0 in 2018.
By "ready by the end of the year", do you mean available to purchase by the end of 2018?
I want to build a new system with possible multiple GPUs running at at least x16x16x16x16 for CUDA computations.
By "ready by the end of the year", do you mean available to purchase by the end of 2018?
I mean ready as in finalized by the PCIe-SIG so device manufacturers can finalize their own 5.0 devices' design and get ready to launch stuff.
Put yourself in the device manufacturer's shoes: PCIe 3.0 has been around for five years, PCIe 4.0 don't have anything to plug into yet even if they existed and by the time the first servers with PCIe 4.0 launch, PCIe 5.0 may be less than a year away. Would you bother spending tens of millions in compliance testing equipment and R&D for a standard that still fails to deliver the bandwidth current high-bandwidth devices require (mainly NVMe SSDs) and will be obsolete after perhaps a year of production?
For many companies, it will make more sense to skip PCIe 4.0 altogether and focus on 5.0 instead.