March 14, 2019 
< | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 | 27 | 28 | 29 | 30 | 31 | >
March 14, 2019 [01:08:43] *** drbobbeaty <drbobbeaty!~drbob@73.110.42.56> has quit IRC (Ping timeout: 246 seconds)[02:53:40] *** Shio_ <Shio_!~Thunderbi@151.100.wtt.p2wt.com> has joined #zeromq[02:56:37] *** Shio_1 <Shio_1!~Thunderbi@151.100.wtt.p2wt.com> has joined #zeromq[02:58:22] *** Shio_ <Shio_!~Thunderbi@151.100.wtt.p2wt.com> has quit IRC (Ping timeout: 255 seconds)[02:58:23] *** Shio_1 is now known as Shio_[05:56:12] *** jud <jud!~jud@cpe-70-113-106-222.austin.res.rr.com> has joined #zeromq[05:56:12] *** jud <jud!~jud@cpe-70-113-106-222.austin.res.rr.com> has quit IRC (Changing host)[05:56:12] *** jud <jud!~jud@unaffiliated/jud> has joined #zeromq[05:57:28] *** jud^ <jud^!~jud@cpe-70-113-106-222.austin.res.rr.com> has quit IRC (Ping timeout: 255 seconds)[06:05:00] *** jud <jud!~jud@unaffiliated/jud> has quit IRC (Quit: Leaving)[07:58:35] *** guido_g <guido_g!~guido_g@195.49.156.9> has joined #zeromq[08:57:15] *** FabriceB <FabriceB!~FabriceB@lfbn-1-1047-122.w86-247.abo.wanadoo.fr> has quit IRC (Ping timeout: 246 seconds)[09:48:29] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has joined #zeromq[09:53:22] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has quit IRC (Read error: Connection reset by peer)[09:53:25] *** jimklimov1 <jimklimov1!~jimklimov@31.7.243.238> has joined #zeromq[10:07:25] *** jimklimov1 <jimklimov1!~jimklimov@31.7.243.238> has quit IRC (Read error: Connection reset by peer)[10:07:43] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has joined #zeromq[10:23:08] *** bluca <bluca!~bluca@2a01:4b00:f419:6f00:b00c:66c8:99df:336> has joined #zeromq[11:00:22] *** FabriceB <FabriceB!~FabriceB@mhx-outbound.3ds.com> has joined #zeromq[11:05:38] *** FabriceB_ <FabriceB_!~FabriceB@mhx-outbound.3ds.com> has joined #zeromq[11:08:23] *** FabriceB <FabriceB!~FabriceB@mhx-outbound.3ds.com> has quit IRC (Ping timeout: 245 seconds)[11:08:23] *** FabriceB_ is now known as FabriceB[11:14:21] *** Shio_ <Shio_!~Thunderbi@151.100.wtt.p2wt.com> has quit IRC (Ping timeout: 252 seconds)[11:21:55] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has quit IRC (Read error: Connection reset by peer)[11:22:25] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has joined #zeromq[13:05:41] *** kamil4 <kamil4!~kamyk@user-5-173-248-46.play-internet.pl> has joined #zeromq[13:07:34] *** kamil4 <kamil4!~kamyk@user-5-173-248-46.play-internet.pl> has left #zeromq[13:10:20] *** guido_g <guido_g!~guido_g@195.49.156.9> has quit IRC (Quit: Leaving.)[13:19:37] *** Maxxed <Maxxed!~Maxxed@ec2-107-20-252-119.compute-1.amazonaws.com> has quit IRC (Ping timeout: 245 seconds)[15:50:54] *** cognifloyd <cognifloyd!~cognifloy@072-191-249-006.res.spectrum.com> has joined #zeromq[16:50:48] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has quit IRC (Ping timeout: 250 seconds)[17:30:27] *** Maxxed <Maxxed!~Maxxed@ec2-107-20-252-119.compute-1.amazonaws.com> has joined #zeromq[18:06:29] *** drbobbeaty <drbobbeaty!~drbob@2601:249:1000:71d0:d04d:a142:6e09:606e> has joined #zeromq[18:28:07] *** FabriceB <FabriceB!~FabriceB@mhx-outbound.3ds.com> has quit IRC (Ping timeout: 240 seconds)[19:27:41] *** guido_g <guido_g!~guido_g@ip1f103f9c.dynamic.kabel-deutschland.de> has joined #zeromq[20:37:27] *** FabriceB <FabriceB!~FabriceB@lfbn-1-1047-122.w86-247.abo.wanadoo.fr> has joined #zeromq[20:39:11] *** FabriceB_ <FabriceB_!~FabriceB@lfbn-1-1047-122.w86-247.abo.wanadoo.fr> has joined #zeromq[20:42:11] *** FabriceB <FabriceB!~FabriceB@lfbn-1-1047-122.w86-247.abo.wanadoo.fr> has quit IRC (Ping timeout: 255 seconds)[20:42:11] *** FabriceB_ is now known as FabriceB[21:58:37] *** bluca <bluca!~bluca@2a01:4b00:f419:6f00:b00c:66c8:99df:336> has quit IRC (Quit: Leaving.)[23:34:48] *** SebasC <SebasC!~Sebas@D93EF1AE.cm-28.dynamic.ziggo.nl> has joined #zeromq[23:35:16] <SebasC> Zeromq message routing drops about 20% when moving from a 1byte id, to a 2 byte id. (using router to router)..[23:35:38] <SebasC> Is there something wrong with the indexing?[23:36:15] <SebasC> correction: 1 byte to 7 bytes[23:41:19] <bv> Is this at high rate and without ZMQ_ROUTER_MANDATORY?[23:42:55] <SebasC> high rate with ZMQ_ROUTER_MANDATORY[23:43:04] <SebasC> 800k msgs/s[23:44:16] <SebasC> drops to 500-600k when moving from 1 byte to 7 byte id, found this out by accident when I messed this macro for finding byte size of binary id: #define NODE_ID_SIZE( id ) (( (64-__builtin_clzll( (uint64_t)id )) / 8) + (((64-__builtin_clzll( (uint64_t)id )) % 8) ? 1 : 0))[23:46:32] <bv> oh boy[23:49:40] <SebasC> hmmm.. cannot seem to replicate it... when I retry it.[23:49:49] <SebasC> might be another glitch.. sorry for the mix up[23:51:14] <bv> I'm curious to hear how it goes.[23:52:35] <bv> If I parse it right, I think that NODE_ID_SIZE is correct. It returns the integral number of bytes just big enough to hold the id.[23:53:00] <SebasC> Yes :) that is the fixed macro :) in the previous one I forget the 64- parts[23:54:20] <bv> ah. I have never looked into using __builtin's. They bind the code to using GCC as the compiler, right?[23:54:33] <SebasC> yes, they are asm instructions[23:54:40] <SebasC> I love those :) use them all the time.[23:55:11] <SebasC> I got some sweet code for you :)[23:55:12] <SebasC> https://pastebin.com/XNJ3RPn5[23:55:37] <SebasC> it's a thread safe bit list :) I use it to find free worker threads (128 threads)[23:56:14] <SebasC> if the first set if !bits[0] I check bits[1].. (set these to volatile) you can add as many as you want :)[23:57:02] <SebasC> I am writing cluster software right now :) so there is some proprietary source for free ;)[23:57:52] <SebasC> its how I set the bits on :) the second function is something else.. how I turn them off? you can deduct it from the first function ;)