Sunday November 14, 2010

[Time] NameMessage
[00:42] Guthur umm, is there some reason why streamer and forwarder are identical
[00:42] Guthur just to remind the user to use the appropriate end point types?
[10:24] Guthur Is it just me or are the z_msg decode and encode functions miss-named?
[10:25] Guthur encode to me seems like it should be taking a string and encoding it and decode taking a byte stream and returning the string
[10:25] mato sustrik: are you there?
[10:26] Guthur not to mention that the comment block is the same for each function which just adds to the confusion
[10:27] mato Guthur: raise an issue somewhere (issue tracker? zmq ML) with Pieter Hintjens, ZFL is his project
[10:27] Guthur it's part of the guide as well
[10:27] mato that's his project too :-)
[10:27] Guthur hehe, ok
[10:27] mato I would have preferred that the guide use plain ZMQ and not push any other layers, but it's up to Pieter...
[10:28] mato I suggest you take that up with him, on the ZMQ mailing list
[10:28] Guthur Most of the guide is plain ZMQ to be fair
[10:28] Guthur it's just the last advanced example section
[10:28] Guthur which is fair enough IMHO
[10:29] Guthur Even then it doesn't use anything that is not shown in the example
[10:29] Guthur ZFL is not required as far as I am aware
[10:29] Guthur though I have not actually completed the last few examples
[10:30] Guthur I personally think the guide is a great introduction to ZMQ
[10:37] Guthur There could be more examples from the bindings though.
[10:37] Guthur I think Java is probably quite a glaring omission, considering how popular it is
[10:45] keffo I like the new site, much better!
[14:25] keffo ./local_thr.exe tcp:// 10000 10000
[14:25] keffo Your average throughput is 28732 [msg/s]
[14:25] keffo Your average throughput is 2298 [Mb/s]
[14:25] keffo vrom vrom
[14:28] keffo 10 byte msg size: Your average throughput is 3179966 [msg/s]
[14:28] keffo Your average throughput is 254 [Mb/s]
[16:00] Guthur keffo...
[16:00] Guthur Your average throughput is 28854 [msg/s]
[16:00] Guthur Your average throughput is 2308 [Mb/s]
[16:00] Guthur that's the CLR binding
[16:00] Guthur though totally meaningless of course, hehe
[16:12] keffo it is yeah
[16:12] keffo meaningless?
[16:16] Guthur yep
[16:17] Guthur I should run C version though
[16:17] keffo this was debug build in an ubuntu VM
[16:17] Guthur I am personally interested in comparing the performance of the clrzmq binding
[16:17] keffo it was the clr binding
[16:17] Guthur yep
[16:18] keffo oh you mean the fact that it's a binding rather than pure C that is meaningless? Yeah, I agree
[16:18] Guthur umm nope not really, I meant comparing mine to your figures
[16:18] keffo (debug/release was equally meaningless)
[16:19] Guthur different computers obviously to not make for a good comparison
[16:19] Guthur to/do
[16:20] Guthur it would be nice to see if there was much overhead in going through the CLR interop
[16:21] Guthur Wouldn't make me want to develop a significantly complex application in C even if there was but it would be nice to knpw
[16:21] Guthur know*
[16:24] keffo probably insignificant