Does number of CPU PCI-e lanes affect IRAY performance with 2x gpu's?

I recently stumbled upon an interesting article proving that in most gpu rendering engines it doesn't even matter, if your CPU can handle more than 16 direct PCI-e 3.0 lanes when you have more than one video card
It got me thinking, does Iray work that way as well?
Technically speaking, if you would install 2 identical GPU's on a single motherboard with 2 PCI-e x16 slots, your CPU would have to support 32 direct PCI-e lanes for these two cards to operate at full bandwidth. If this is not the case, and you have for example an i7 8700 (16 lanes), both cards are practically 'degraded' to operate at 8x speed.
So my question is - does any of you ever used a 2x gpu configuration with a 16-lane cpu, and then switched to 32-lane or better and actually seen any performance boost during rendering?
Cheers
Comments
The short answer in no.
The long answer is, pcie lanes do matter for out of core rendering or game engines, as well as real-time pbr engines. So if you plan to use other engines then more lanes may come handy. Also a future version of iray may implement ooc or real-time pbr. But for the actual version of iray it doesn't matter.
PCIE lanes are needed for transfering data between the GPU and the CPU/RAM. Iray loads the whole scene at the start of the render so having a card on a x1 slot will load the data at 1/16th the speed of the same card on a x16 slot. However PCIE gen 3 has a lot of bandwidth, nearly 1 Gb/sec, per lane so even transfering all the data over one lane won't take that long. After that the GPU doesn't "talk" to the CPU that much and PCIE bandwidth is even less relevant.
Okay, thanks for explanation :)