--- Log opened Sat Jan 23 00:00:42 2021 |
01:15 | | mac [macdjord@Nightstar-re5.7if.45.45.IP] has joined #code |
01:15 | | mode/#code [+o mac] by ChanServ |
01:18 | | macdjord|slep [macdjord@Nightstar-re5.7if.45.45.IP] has quit [Ping timeout: 121 seconds] |
01:21 | | macdjord|slep [macdjord@Nightstar-re5.7if.45.45.IP] has joined #code |
01:21 | | mode/#code [+o macdjord|slep] by ChanServ |
01:24 | | mac [macdjord@Nightstar-re5.7if.45.45.IP] has quit [Ping timeout: 121 seconds] |
03:00 | | catalyst_ [catalyst@Nightstar-ejd4sd.cable.virginm.net] has quit [Connection closed] |
03:54 | | Degi [Degi@Nightstar-gm4ka0.dyn.telefonica.de] has quit [Operation timed out] |
03:54 | | Degi [Degi@Nightstar-omun51.dyn.telefonica.de] has joined #code |
04:24 | | VirusJTG [VirusJTG@Nightstar-42s.jso.104.208.IP] has quit [Connection closed] |
04:24 | | VirusJTG [VirusJTG@Nightstar-42s.jso.104.208.IP] has joined #code |
04:24 | | mode/#code [+ao VirusJTG VirusJTG] by ChanServ |
05:03 | | Vornicus [Vorn@ServerAdministrator.Nightstar.Net] has quit [Connection closed] |
05:09 | | Vorntastic [uid293981@Nightstar-h2b233.irccloud.com] has joined #code |
05:09 | | mode/#code [+qo Vorntastic Vorntastic] by ChanServ |
06:13 | | SmithKurosaki [uid215460@Nightstar-ed0oqj.irccloud.com] has quit [[NS] Quit: Connection closed for inactivity] |
06:17 | | celticminstrel [celticminst@Nightstar-n1gkap.dsl.bell.ca] has quit [[NS] Quit: And lo! The computer falls into a deep sleep, to awake again some other day!] |
06:44 | | catalyst [catalyst@Nightstar-ejd4sd.cable.virginm.net] has joined #code |
07:12 | | abudhabi [abudhabi@Nightstar-msdsqr.adsl.tpnet.pl] has joined #code |
09:41 | | Kindamoody[zZz] is now known as Kindamoody |
09:58 | | Emmy [Emmy@Nightstar-l49opt.fixed.kpn.net] has joined #code |
10:03 | | mode/#code [+o ErikMesoy] by ChanServ |
10:34 | | Kindamoody is now known as Kindamoody|afk |
13:35 | | john_cephalopoda [john@Nightstar-p87qss.dip0.t-ipconnect.de] has joined #code |
14:48 | | Vornicus [Vorn@ServerAdministrator.Nightstar.Net] has joined #code |
14:48 | | mode/#code [+qo Vornicus Vornicus] by ChanServ |
15:59 | | Vorntastic [uid293981@Nightstar-h2b233.irccloud.com] has quit [[NS] Quit: Connection closed for inactivity] |
16:19 | | Kizor [moryok@Nightstar-e0a4sm.utu.fi] has quit [Ping timeout: 121 seconds] |
16:29 | | Kizor [moryok@Nightstar-e0a4sm.utu.fi] has joined #code |
16:30 | | Kizor is now known as NSGuest42457 |
16:33 | | NSGuest42457 is now known as Kizor |
16:35 | | Kindamoody|afk is now known as Kindamoody |
17:20 | | Kizor [moryok@Nightstar-e0a4sm.utu.fi] has quit [Ping timeout: 121 seconds] |
17:40 | | Kizor [moryok@Nightstar-e0a4sm.utu.fi] has joined #code |
17:41 | | Kizor is now known as NSGuest56414 |
19:20 | | celticminstrel [celticminst@Nightstar-n1gkap.dsl.bell.ca] has joined #code |
19:20 | | mode/#code [+o celticminstrel] by ChanServ |
19:49 | | NSGuest56414 is now known as Kizor |
20:54 | | ToxicFrog [ToxicFrog@ServerAdministrator.Nightstar.Net] has quit [The TLS connection was non-properly terminated.] |
21:06 | <&[R]> | https://i.imgur.com/EFgu8xQ.png |
21:09 | <@Tamber> | ............ |
21:09 | < john_cephalopoda> | It is fast. |
21:10 | | Reiver is now known as Joe |
21:10 | < john_cephalopoda> | If it was done with a loop and bit shifting, it would take <number of bits> loop rounds. With that approach, the compiler will likely build a pretty jump table with O(1) access. |
21:11 | | Kizor is now known as Shirley |
21:11 | | Vornicus is now known as Misty |
21:12 | < john_cephalopoda> | The jump table is definitely the less memory-efficient solution though. |
21:13 | <&McMartin> | I am also going to assume that that function's text was mechanically generated. |
21:13 | <~Misty> | not actually faster though because the dereferences will also cost actual time |
21:13 | <&McMartin> | You're replacing 32 branches with one computed GOTO. |
21:14 | <&McMartin> | Er. |
21:14 | <&McMartin> | 16 branches. |
21:14 | <&McMartin> | These days whether that is *faster* will depend on how good your branch prediction is compared to the cost of a miss in the instruction cache |
21:16 | < john_cephalopoda> | InstructionTable[ip](); Should be pretty fast. |
21:21 | < john_cephalopoda> | My naive solution for the other way would be https://bpa.st/WKNIA . |
21:21 | < john_cephalopoda> | Ah, damn, ip & 15 is wrong. |
21:22 | < john_cephalopoda> | Should be ip & 0x8000, if I am not mistaken. |
21:30 | | abudhabi [abudhabi@Nightstar-msdsqr.adsl.tpnet.pl] has quit [NickServ (RECOVER command used by abudhabi_)] |
21:30 | | abudhabi [abudhabi@Nightstar-cgcbhm.adsl.tpnet.pl] has joined #code |
21:59 | | ToxicFrog [ToxicFrog@ServerAdministrator.Nightstar.Net] has joined #code |
21:59 | | mode/#code [+ao ToxicFrog ToxicFrog] by ChanServ |
22:04 | | Shirley is now known as Allison |
22:06 | | Allison is now known as Shirley |
22:29 | <~Misty> | 16 branches, the instructions for each thing is 16 dereferences |
22:44 | < john_cephalopoda> | 16 branches for which one? |
22:44 | < john_cephalopoda> | Ah, the loop one? |
22:45 | < john_cephalopoda> | Wouldn't it be 32 branches? 16 for the if/else, 16 for the condition of the for loop? (Though it should be easy to unroll.) |
22:50 | | Pink_ [uid208117@Nightstar-h2b233.irccloud.com] has joined #code |
22:50 | <~Misty> | with that small a thing, yeah, I'd say unroll is reasonably likely |
22:55 | | Emmy [Emmy@Nightstar-l49opt.fixed.kpn.net] has quit [Ping timeout: 121 seconds] |
23:56 | < john_cephalopoda> | Hm, I wonder if the look-up table of the other approach could be generated using a compiler macro. |
--- Log closed Sun Jan 24 00:00:44 2021 |