I see on the Nvidia driver website that there is now an aarch64 Linux .run file available for all modern Quadro cards that I’m assuming was ported up from the Jetson platforms. Seems starting in July with versions 450 and later.
Some of our projects feature Xilinx FPGA SoCs in house (Zynq platform) which have a Quad Core ARM Cortex A53 and a 16x PCI-e Gen3 connection for peripherals.
With this newly available aarch64 nvidia driver, are there any concerns with integrating a Quadro grade card to this PCI-e slot to act as the GPU for this ARM SoC?
I see the README in the .run file calls out Tegra processors, X-Gene and ThunderX. Are these hardware requirements a strict requirement or just known working platforms?
The Zynq platforms use Yocto and PetaLinux to create the embedded linux build used on their systems. Currently we can run a 3.10 Linux kernel w/ an Xserver of 1.20.1. So I believe we meet all the Minimum Software Requirements listed in the README
Also is there CUDA support? I see CUDA on ARM available, but seems to expect an SBSA architecture. But then there seem to be Debian repos that can pull a CUDA Toolkit. Are those available on an embedded linux system?
Thanks in advance!