So Nvidia can do 25 users per GPU at the moment with their "hypervisor aware" VDI cards. That's first gen.
The real trick will be doing texture swaps in conjunction with compression scaling to account for changes in network latency. That shit will be fucking magic.
But even if you get this perfectly scaled, overall latency/contention will make remote gaming over WAN suck for a mass roll-out with today's broadband and thus limits the addressable market.
So Nvidia can do 25 users per GPU at the moment with their "hypervisor aware" VDI cards. That's first gen.
The real trick will be doing texture swaps in conjunction with compression scaling to account for changes in network latency. That shit will be fucking magic.
But even if you get this perfectly scaled, overall latency/contention will make remote gaming over WAN suck for a mass roll-out with today's broadband and thus limits the addressable market.
edit for tenses, i haven't slept in 36 hours.