I use epgm with zeromq and the c++ api all the time. I don't see the rate issues you see.
Are you: - Setting ZMQ_RATE before you call connect ? - Are you setting ZMQ_RATE on both ends (pub/sub) Other things to try: - test using c/c++ rather than python to verify Other things pgm related but do not affect bandwidth: - I have had issues with ZMQ_RECOVERY_IVL being too large in the past but that's with hundreds of subscribers. - There is a random crash using zeromq/pgm with more than one context thread, I have never ever managed to recreate on demand though. On Sun, Jun 28, 2020 at 3:41 PM Osiris Pedroso <opedr...@gmail.com> wrote: > Have you measured the same scenario without zeromq? > By the way, does zeromq have a mock implementation that simply passes all > requests to the underlying socket? > That would be useful to measure zeromq impact on transmission in > performance evaluations providing a baseline. > > On Sun, Jun 28, 2020, 11:33 董晓滨 <dongxiaobin0...@163.com> wrote: > >> yes, that is right. But I just want to know why there are so big >> difference between the performance of TCP and EPGM? >> >> 董晓滨 >> 邮箱:dongxiaobin0...@163.com >> >> <https://maas.mail.163.com/dashi-web-extend/html/proSignature.html?ftlId=1&name=%E8%91%A3%E6%99%93%E6%BB%A8&uid=dongxiaobin0614%40163.com&iconUrl=https%3A%2F%2Fmail-online.nosdn.127.net%2Fqiyelogo%2FdefaultAvatar.png&items=%5B%22%E9%82%AE%E7%AE%B1%EF%BC%9Adongxiaobin0614%40163.com%22%5D> >> >> 签名由 网易邮箱大师 <https://mail.163.com/dashi/dlpro.html?from=mail88> 定制 >> >> On 06/28/2020 15:56, Doron Somech <somdo...@gmail.com> wrote: >> Don't you want to multiply the rate by 8 as well? As it is bits? >> >> On Mon, Jun 22, 2020 at 8:39 PM 董晓滨 <dongxiaobin0...@163.com> wrote: >> >>> Hi, everyone ! >>> I want to use zeromq pub/sub to transport large size msgs (e.g. images >>> data). >>> I did the below test in python environment: >>> 1. send different sizes of msg (from 2kb - 16mb) with random data >>> 2. one publisher in one docker container, and one subscriber in one >>> other seperate container >>> 3. calculate the total data size and time, to estimate the throughput >>> (total size / time) in MB/s >>> >>> tcp have good performance, but epgm I only got 20M/s throughput even >>> after I set the socket.rate, hwm ... Just want to know the reason. I hope I >>> can use epgm because it is multicast protocol which should have advantage >>> than tcp in multi subscriber cases.. test result have been shown below and >>> test code has been attached. >>> >>> >>> >>> >>> >>> _______________________________________________ >>> zeromq-dev mailing list >>> zeromq-dev@lists.zeromq.org >>> https://lists.zeromq.org/mailman/listinfo/zeromq-dev >>> >> _______________________________________________ >> zeromq-dev mailing list >> zeromq-dev@lists.zeromq.org >> https://lists.zeromq.org/mailman/listinfo/zeromq-dev >> > _______________________________________________ > zeromq-dev mailing list > zeromq-dev@lists.zeromq.org > https://lists.zeromq.org/mailman/listinfo/zeromq-dev >
_______________________________________________ zeromq-dev mailing list zeromq-dev@lists.zeromq.org https://lists.zeromq.org/mailman/listinfo/zeromq-dev