How does Xavier's Volta implementation compare to the GV100's Volta implementation? (Solved)

Hi;

This seemed to be the only relevant VOLTA architecture documentation that I could find:

Is there an equivalent document for the Xavier implementation?

How does Xavier’s implementation compare to the GV100? What things that can be done on the GV100 are not possible on Xavier?

Thankyou in advance for any guidance you can provide.

Hi John, recommend that you tune into our webinar tomorrow, as we will be covering the Xavier GPU in detail.

They both share a similar Volta SM with Tensor Cores, with Xavier having less FP64 cores. GV100 has NVLINK whereas Jetson AGX Xavier does not.

Thankyou Dusty for that information and i will indeed watch the webinar tomorrow.

argh. my dentist appt is at the same time! will this be saved and viewable later in the day?

No problem, yes it will be available for streaming On-Demand within 24 hours, and I will be posting the slides to the announcement threads here.

TYVM!

Slides have now been posted here — https://github.com/dusty-nv/jetson-presentations/raw/master/20181004_Jetson_AGX_Xavier_New_Era_Autonomous_Machines.pdf

VG! Tnx!