If you're serious about working on RDMA in Corundum, how about we
set up a zoom call or something so we can get everybody on the
same page? And not just for RDMA, but for any potential feature.
Would something like this interest anyone else on the list?
Alex Forencich
Hi,
Is the support for RMDA on the roadmap for Corundum?
Which parts do you suggest we can dive into to modify?--
You received this message because you are subscribed to the Google Groups "corundum-nic" group.
To unsubscribe from this group and stop receiving emails from it, send an email to corundum-nic...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/corundum-nic/f7039d15-028a-4ab7-a249-c92618d3faffn%40googlegroups.com.
I'll need to read over that paper more carefully, but it seems
like worst-case you could basically run that in parallel with the
corundum datapath - peel off the incoming RDMA traffic, hand it
off to the StRoM stack, which would then have an extra port on the
DMA engine to access host memory. In this case, you would
basically be bypassing the core corundum datapath. But, I can
definitely see how having a "standard" RDMA implementation of some
sort as part of corundum would be useful here at least to
implement the standard part of the RDMA stack.
Alex Forencich
To view this discussion on the web visit https://groups.google.com/d/msgid/corundum-nic/af72486c-0338-48e3-9ed5-21cb4f9b0095n%40googlegroups.com.
This sort of thing is exactly why this kind of discussion makes
sense...get everyone on the same page to minimize duplication of
effort.
One thing to keep in mind: I am in the process of porting
Corundum to Intel Stratix 10 MX and DX, so whatever code goes in
to the core logic also has to run on Stratix 10. This may rule
out the use of code written for Xilinx HLS in Corundum, not to
mention using HLS will likely cause serious complications with the
current CI setup on github actions.
Alex Forencich
To view this discussion on the web visit https://groups.google.com/d/msgid/corundum-nic/8aca982a-2ca4-49d7-b89b-c8bf4c2fd61fn%40googlegroups.com.