Ask Your Question
0

Spice Vs PCoIP - Where is Spice Hardware Acceleration

asked 2014-10-09 04:09:56 -0500

willbittner gravatar image

Hi Everyone, As a diehard linux and kvm fan, I have spent a lot of time tinkering with the spice protocol (including trying new compression techniques etc) while working on a DaaS ( Desktop as a Service ) or VDI (Buzzwords..) Manaaged Solution with zeroclients using things such as android USB sticks, raspberry pi, android monitors, thin clients and so on.

As I am now launching the managed service based company, I am sad to say that Zero Clients leveraging Teradicis PCoIP protocol is really kicking Spice ass. Most notably on window dragging, and events where large pixel areas are changing AND the XY position is changing so it cannot activate a video codec, and it hurts me greatly to see VMWare grabbing this emerging market and will end up leaving KVM/OpenStack and the open source community stuck to catch up later. PCoIP has the ability to do server side 3D Rendering also... Are there any hardware accelerated Spice protocol FPGAs or ASICS or boards or possibly a port to OpenCL to be able to leverage the new OpenCL to FPGA boards? On either the client or server side?

Please someone point me to a sign of hope so I can contribute to the project! If there is nothing and you are interested in helping me build a hardware accelerated FPGA ( to later hopefully open source design so kvm/spice/openstack zeroclients can be built and compete) please email me: will@brookfield.ventures

edit retag flag offensive close merge delete

1 answer

Sort by ยป oldest newest most voted
0

answered 2014-10-09 11:51:38 -0500

mpetason gravatar image

updated 2014-10-09 11:53:29 -0500

PCI Passthrough for GPU + The Video Card that were very popular for VDI should be a starting point:

https://wiki.openstack.org/wiki/Pci_p...

http://www.teradici.com/products-and-...

Then you need to figure out how to take advantage of it in the same way VMware does.

Since most instances aren't actually accessing a GPU you could expect some limitations with Graphics, especially in a shared environment where the CPU is doing all of the work. I would go along the same lines that I looked into with VMware to do the same thing, but instead figure out how to incorporate it into the OpenStack code for Nova.

edit flag offensive delete link more

Your Answer

Please start posting anonymously - your entry will be published after you log in or create a new account.

Add Answer

Get to know Ask OpenStack

Resources for moderators

Question Tools

1 follower

Stats

Asked: 2014-10-09 04:09:56 -0500

Seen: 4,161 times

Last updated: Oct 09 '14