Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Does infinity support roce card? #9

Open
fqhuang opened this issue Mar 16, 2021 · 6 comments
Open

Does infinity support roce card? #9

fqhuang opened this issue Mar 16, 2021 · 6 comments

Comments

@fqhuang
Copy link

fqhuang commented Mar 16, 2021

I got the same issue like this: #5.

my card is mlx5 with roce mode:
CA 'mlx5_0'
CA type: MT4119
Number of ports: 1
Firmware version: 16.29.1016
Hardware version: 0
Node GUID: 0x0c42a103007555a4
System image GUID: 0x0c42a103007555a4
Port 1:
State: Active
Physical state: LinkUp
Rate: 100
Base lid: 0
LMC: 0
SM lid: 0
Capability mask: 0x00010000
Port GUID: 0x0e42a1fffe7555a4
Link layer: Ethernet

Thanks!
Toby

@claudebarthels
Copy link
Owner

Hello Toby,

no, Infinity does not support RoCE cards. However, the changes that would be required have already been factored out into single class: infinity/src/infinity/queues/QueuePairFactory. This class is responsible for establishing the connections, and returning connected and fully functional queue pairs. As far as I understand, the connection setup is the primary difference between InfiniBand and RoCE - the rest of the code is agnostic to the network technology.

Unfortunately, I never had access to RoCE cards - hence there is no QueuePairFactory for RoCE networks.

Hope this helps,
Claude

@fqhuang
Copy link
Author

fqhuang commented Mar 17, 2021

Hello Claude,
Thanks, I will try to modify it.

Toby

@ShaneKirkbride
Copy link

@fqhuang were you successful? at adding RoCE functionality?

@fqhuang
Copy link
Author

fqhuang commented Sep 23, 2021

@ShaneKirkbride no, I don't use this lib for RoCE finally.

@sarthak-moorjani
Copy link

Hi @fqhuang - What did you finally use for RoCE?

@claudebarthels - Thanks for the great library. I am also facing the same issue since I have RoCE cards, could you help me in explaining that what exactly might need to change for this to work for RoCE cards? And where can I find the relevant documentation to make these changes?
Also, do you think that the changes be bulky or small changes in connection establishment?

Thank you!

@sarthak-moorjani
Copy link

Basically the issue I am facing is that the programs hangs after I run a read write test.. Exactly like in the first comment of the issue here -
#5

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

4 participants