Posts: 2
Threads: 1
Joined: Jul 2024
Hello! I'm a student studying RVV, and I'm very interested in spatz. I have obtained some materials about spatz, including papers, materials in the github repository, PPT on the official website, etc. however, I found that the documentation provided in the github repository is not comprehensive. The introduction in the papers also has some differences from the current hardware design.
I want to know the hardware design details of spatz and how spatz works with snitch. Could you please provide me with more detailed documentation? It will be very helpful to me.
Thank you very much!
Posts: 147
Threads: 0
Joined: Oct 2018
Hello,
We get these types of questions quite a lot (wanting more and better documentation). It is not for the lack of trying, but please understand that these are active research projects conducted by Ph.D. students in our group and the main goal is to figure out better and more efficient ways of implementing these functionalities. Peer reviewed papers take some time to get published, in the meantime, we could have already made improvements and/or adapted the design to work in other environments. It is not that we are proud of it, but documentation ends up behind developments as a result.
Everything we have on these designs are actually on the GitHub pages, internal repositories only contain technology specific implementation details (which we are not allowed to make public) and not better documentation or more up to date code.
As an open source project, we would actually be very happy to get contributions, and as you can see the most we need hep with is documentation.
Cheers,
KGF
Visit pulp-platform.org and follow us on twitter @pulp_platform
Posts: 2
Threads: 1
Joined: Jul 2024
(10-16-2024, 06:28 AM)kgf Wrote: Hello,
We get these types of questions quite a lot (wanting more and better documentation). It is not for the lack of trying, but please understand that these are active research projects conducted by Ph.D. students in our group and the main goal is to figure out better and more efficient ways of implementing these functionalities. Peer reviewed papers take some time to get published, in the meantime, we could have already made improvements and/or adapted the design to work in other environments. It is not that we are proud of it, but documentation ends up behind developments as a result.
Everything we have on these designs are actually on the GitHub pages, internal repositories only contain technology specific implementation details (which we are not allowed to make public) and not better documentation or more up to date code.
As an open source project, we would actually be very happy to get contributions, and as you can see the most we need hep with is documentation.
Cheers,
KGF
Thank you very much for your reply and wish you all the best in your research.
