Yangfisher1 opened a new issue, #3056: URL: https://github.com/apache/brpc/issues/3056
Hi, I noticed that RDMA performance on my machines sometimes fluctuate. I suspect this might be related to NUMA nodes, as the machine has a multi-NUMA architecture. I want to implement NUMA-aware memory allocation instead of directly using posix_memalign for allocation. To achieve this, It needs to first select a socket, then serialize the request to the corresponding NUMA block buffer based on the RDMA device associated with that socket. However, the current implementation serializes first and then selects the socket. What potential issues could arise from reversing the order of these two operations? -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
