| 00:09.16 | CIA-50 | BRL-CAD: 03crdueck * r51436 10/brlcad/trunk/src/librt/primitives/part/part.c: added rt_part_volume() and rt_part_surf_area() |
| 00:10.07 | CIA-50 | BRL-CAD: 03crdueck * r51437 10/brlcad/trunk/src/librt/primitives/table.c: updated rt_functab entry for part to include the new callbacks |
| 00:11.11 | CIA-50 | BRL-CAD: 03crdueck * r51438 10/brlcad/trunk/src/libged/analyze.c: added new analyze_part function |
| 00:30.57 | CIA-50 | BRL-CAD: 03starseeker * r51439 10/brlcad/trunk/src/librt/test_botpatches.cpp: Add a little more flexibility to loop pl output. |
| 03:21.28 | CIA-50 | BRL-CAD: 03Crdueck 07http://brlcad.org * r4057 10/wiki/User:Crdueck/log: |
| 04:04.05 | crdueck | brlcad: could you please take a look at the email i sent to the mailing list? |
| 04:37.34 | *** join/#brlcad andrei_ (~andrei@188.25.172.146) | |
| 04:41.40 | andrei_ | hey |
| 04:43.08 | andrei_ | how should I generate files so that it does not take so much. |
| 05:16.02 | andrei_ | http://slexy.org/view/s20oJzwhYi I m stuck with my script, it has some errors and I can t find the cause. |
| 05:16.12 | andrei_ | (standard_in) 1: parse error |
| 05:16.17 | andrei_ | awk: cmd. line:1: fatal: division by zero attempted |
| 05:39.10 | andrei_ | trying to figure out on myself meanwhile |
| 07:40.16 | andrei_ | With the current status of the script ( ran it 4 times * on two different set of data ) the variation is +/- 0.001. |
| 07:42.11 | andrei_ | I ll proceed to recording data. Hopefully won t need so much time. |
| 07:51.39 | *** join/#brlcad PrezKennedy (~DarkCalf@173.231.40.99) | |
| 09:22.51 | andrei_ | I m currently stuck with my script http://slexy.org/view/s20wmX3vSK . The log_file seems to not exist. tho it should be created when server starts |
| 09:22.57 | andrei_ | there are no tpkg running in background. |
| 09:23.26 | andrei_ | I ll be working on it myself meanwhile. |
| 09:27.55 | *** join/#brlcad andrei_ (andrei@188.25.172.146) | |
| 09:36.30 | *** join/#brlcad stas_ (~stas@82.208.133.12) | |
| 10:31.25 | *** join/#brlcad Jak_o_Shadows (~Fake@unaffiliated/jak-o-shadows/x-0479135) | |
| 10:55.19 | CIA-50 | BRL-CAD: 03phoenixyjll * r51440 10/brlcad/trunk/src/librt/primitives/dsp/dsp_brep.cpp: Implement degree reduction to Bezier curves to simplify the representation of the dsp top surface. But still a long way to go. |
| 11:04.39 | CIA-50 | BRL-CAD: 03Phoenix 07http://brlcad.org * r4058 10/wiki/User:Phoenix/GSoc2012/Reports: /* Week 7 */ |
| 11:05.17 | CIA-50 | BRL-CAD: 03Phoenix 07http://brlcad.org * r4059 10/wiki/User:Phoenix/GSoc2012/Reports: /* Week 7 */ |
| 11:07.47 | CIA-50 | BRL-CAD: 03Phoenix 07http://brlcad.org * r4060 10/wiki/User:Phoenix/GSoc2012/Reports: /* Week 7 */ |
| 11:08.27 | CIA-50 | BRL-CAD: 03Phoenix 07http://brlcad.org * r4061 10/wiki/User:Phoenix/GSoc2012/Reports: /* Week 7 */ |
| 11:08.49 | CIA-50 | BRL-CAD: 03Phoenix 07http://brlcad.org * r4062 10/wiki/User:Phoenix/GSoc2012/Reports: /* Week 6 */ |
| 12:02.56 | *** join/#brlcad Jak_o_Shadows1 (~Fake@CPE-144-136-211-241.sa.bigpond.net.au) | |
| 12:03.03 | *** join/#brlcad Jak_o_Shadows1 (~Fake@unaffiliated/jak-o-shadows/x-0479135) | |
| 12:35.32 | *** join/#brlcad ksuzee (~ksu@46.149.82.166) | |
| 12:36.15 | CIA-50 | BRL-CAD: 03Ksuzee 07http://brlcad.org * r4063 10/wiki/User:Ksuzee/Reports: |
| 12:36.35 | CIA-50 | BRL-CAD: 03Ksuzee 07http://brlcad.org * r4064 10/wiki/User:Ksuzee/Reports: |
| 12:40.38 | *** join/#brlcad d_rossberg (~rossberg@BZ.BZFLAG.BZ) | |
| 12:56.07 | *** join/#brlcad andrei_ (andrei@188.25.172.146) | |
| 13:26.44 | *** join/#brlcad Bariton (~kvirc@p5090598C.dip.t-dialin.net) | |
| 14:04.35 | brlcad | crdueck: replied, but haven't dug in too deep |
| 14:07.17 | brlcad | andrei_: stuck how so? |
| 14:09.58 | ``Erik | starseeker: http://bofh.ntk.net/OtherStuff/TheCorner/striped.php |
| 14:10.29 | brlcad | andrei_: does it work if you ssh to localhost instead? |
| 14:27.43 | brlcad | andrei_: since you have stable variation on one file and one packet size, next up is to graph one file for variable packet sizes |
| 14:34.23 | brlcad | run every packet size 1 through 2048, then 2048 to 4194304 in increments of 2048 |
| 14:34.49 | brlcad | that's 4096 samples at two different scales (two different graphs) |
| 14:37.57 | brlcad | probably should use the same file for all, 4194304 bytes big or some multiple bigger |
| 15:29.02 | crdueck | brlcad: thanks for your reply on the part primitive topic, but I meant the email I'd sent earlier titled "next steps for my project", which hasnt gotten any replies yet. |
| 15:29.45 | brlcad | crdueck: okay, I'll get to it hopefully this weekend -- I have a massive e-mail backlock and lots of pending deadlines to deal with |
| 15:29.57 | brlcad | crdueck: what was the gist? |
| 15:30.22 | brlcad | if you're done with the analyze functions, that's great ;) |
| 15:32.05 | crdueck | yes, there arent too many more primitives that can be analyzed in terms of surface area and volume, so i'm running out of things to do in terms of what i proposed in the GSoC application |
| 15:32.24 | crdueck | i was wondering if there was an obvious extension to the work i'd already done |
| 15:33.36 | brlcad | crdueck: there's always more work to be done |
| 15:35.07 | brlcad | so I find it hard to believe that you've actually correctly implemented several of the harder primitives... dsp, extrude, sketch, and brep come to mind, maybe pnts too |
| 15:35.23 | brlcad | if any aren't done, that's what's next |
| 15:35.34 | brlcad | if you have TODO items on any of the ones that are done, that's what's next |
| 15:35.53 | crdueck | i've looked at them, but how would one analyze sketch? |
| 15:36.13 | brlcad | if you really do believe everything is done, then next up would be to prove it |
| 15:36.32 | brlcad | sketch certainly has area |
| 15:37.35 | brlcad | solving area for a sketch surface is similar to solving it for a brep/nurbs surface |
| 15:38.58 | brlcad | i believe you implemented dsp, but do know that dsp 1) can be smooth or faceted and 2) can have holes |
| 15:39.22 | crdueck | okay, well i'll look into analyzing brep, and then try sketch from there. |
| 15:39.35 | crdueck | i havent done anything with dsp though. |
| 15:40.18 | brlcad | sketch is probably easier |
| 15:40.41 | brlcad | piecewise simple, might be able to use some sort of painter's algorithm |
| 15:42.12 | brlcad | break it down into a tessellation, then just have to deal with the arc/circle/bezier pieces |
| 15:44.51 | crdueck | alright, i'll work these more difficult primitives. I may have questions about how they're used since they arent common geometric shapes |
| 15:58.26 | brlcad | crdueck: they're crazy-common, just not basic euclidean :) |
| 18:16.24 | *** join/#brlcad stas_ (~stas@188.24.47.249) | |
| 18:21.34 | *** join/#brlcad andrei_ (~andrei@188.25.172.146) | |
| 18:21.38 | andrei_ | hei |
| 18:21.50 | andrei_ | wrong lol. |
| 18:22.18 | brlcad | hei lol. |
| 18:23.05 | andrei_ | brlcad, why would I use 4194304 ? |
| 18:23.14 | andrei_ | I used 65000 because as far as I know that is the tcp payload |
| 18:23.47 | brlcad | what happens if you tell libpkg to go bigger? |
| 18:24.29 | andrei_ | I haven't tried. |
| 18:24.36 | brlcad | exactly :) |
| 18:24.59 | andrei_ | I thought I said what the issue was |
| 18:25.24 | andrei_ | in order to determine the average I erase the log_file after 10 iterrations |
| 18:25.32 | andrei_ | 10 iterations are used for one average |
| 18:25.51 | andrei_ | if I wouldn't erase it the first 10 values would influence the second 10 values. |
| 18:25.57 | brlcad | right, that make sense |
| 18:26.21 | andrei_ | I ll paste the two lines that cause trouble |
| 18:26.28 | brlcad | and no, you didn't say that was the issue |
| 18:26.40 | andrei_ | sorry then ... |
| 18:26.45 | andrei_ | ssh andrei@192.168.1.100 'cat log_file' | awk '/Elapsed_time/{sum+=$2;n++} END {printf "Elapsed_time %.5f\n", sum/n}' |
| 18:26.52 | andrei_ | ssh andrei@192.168.1.100 'rm log_file' |
| 18:26.59 | andrei_ | there is a " wait " after first ssh |
| 18:27.08 | andrei_ | if I run 1 iterations it is all good. |
| 18:27.14 | andrei_ | ( 1 set of 10 ) |
| 18:27.26 | andrei_ | if I run more, ssh / awk complains about file not existing |
| 18:27.31 | andrei_ | I guess I m timing something wrong |
| 18:27.37 | brlcad | sure, no I get the issue -- so you've either got to change how you sum elapsed time, or you need to restart the log file |
| 18:27.56 | andrei_ | yes |
| 18:28.14 | brlcad | I'd personally suggest shutting down the tpkg server after every use if you're not |
| 18:28.32 | brlcad | I think you are, but should be sure |
| 18:28.39 | andrei_ | shouldn't it shut down by itself after transmision is finished? |
| 18:28.48 | brlcad | you tell me |
| 18:28.57 | andrei_ | well I read that in tpkg.c |
| 18:29.04 | andrei_ | and I haven't changed that |
| 18:29.06 | andrei_ | so yes |
| 18:29.07 | brlcad | okay |
| 18:29.32 | andrei_ | if I would shut down tpkg, that wouldn't solve the issue |
| 18:29.37 | brlcad | easiest is probably to just change how you're dealing with the log |
| 18:29.38 | brlcad | don't delete it |
| 18:30.13 | andrei_ | could I echo " null " |
| 18:30.14 | andrei_ | ? |
| 18:30.20 | andrei_ | like this echo " null " > log_file ? |
| 18:30.59 | brlcad | not sure what that would accomplish |
| 18:31.49 | brlcad | if the server always starts before the client by making the client wait for the log file to exist, it's not clear why the log file wouldn't always exist |
| 18:33.06 | andrei_ | http://slexy.org/view/s21dKkqphK this is the current script |
| 18:34.25 | andrei_ | perhaps you see something that I don't. I don't understand why it doesn't work. |
| 18:37.51 | brlcad | yeah, you have a bug ;) |
| 18:37.57 | brlcad | sees it |
| 18:38.39 | andrei_ | where? |
| 18:39.45 | brlcad | how could line 30 possibly not find log_file? |
| 18:41.55 | andrei_ | if it's not there? |
| 18:42.22 | andrei_ | on remote machine the file is in /home/andrei |
| 18:42.37 | brlcad | and under what circumstance will that file not exist? |
| 18:42.51 | brlcad | what circumstance in the code logic |
| 18:44.11 | andrei_ | if the server or client do not start properly ? |
| 18:44.23 | brlcad | print a "STARTING SERVER" and "STARTING CLIENT" before both, run your script, see what happens |
| 18:48.05 | andrei_ | I get 10 sets of Starting Client / server |
| 18:48.11 | andrei_ | Elapsed_time 0.15133 |
| 18:48.12 | andrei_ | <PROTECTED> |
| 18:48.15 | andrei_ | then the valid data. |
| 18:48.27 | andrei_ | Then the error occurs |
| 18:48.53 | brlcad | read just the "STARTING SERVER" and "STARTING CLIENT" lines |
| 18:49.24 | brlcad | server -> client -> server -> client -> ... |
| 18:49.43 | andrei_ | http://slexy.org/view/s213t7Zm5X |
| 18:49.47 | andrei_ | yes, I see nothing wrong |
| 18:55.00 | brlcad | maybe pay attention to the ordering of the elapsed time line |
| 18:55.36 | brlcad | print a message before the "cat log_file" line |
| 18:57.48 | andrei_ | ah |
| 18:57.51 | andrei_ | I m an idiot. |
| 18:57.53 | andrei_ | I found it |
| 18:58.07 | andrei_ | avc never becomes 0 again , in never recycles |
| 18:58.10 | andrei_ | meh |
| 19:01.53 | andrei_ | now regarding the 3d graph |
| 19:02.08 | andrei_ | should I be interested in total packages sent / package size / speed / file size? |
| 19:02.15 | andrei_ | which one of those 3 |
| 19:02.35 | andrei_ | I m not sure what to choose from packages sent or package size |
| 19:02.41 | andrei_ | both can be determined from the other |
| 19:19.06 | brlcad | andrei_: yay :) |
| 19:19.20 | brlcad | don't do 3d next, just 2d for those two ranges |
| 19:19.50 | brlcad | if that looks good, then can define a similarly useful range for data sizes |
| 19:20.15 | andrei_ | so the file size should be fixed. |
| 19:20.27 | brlcad | right, |
| 19:20.41 | andrei_ | great, that would speed up things by aproximatively 47 hours :) |
| 19:20.48 | brlcad | I suggested 8MB earlier |
| 19:21.14 | brlcad | note your * 1000 multiplier isn't quite right/useful |
| 19:21.22 | brlcad | should be 1024 |
| 19:22.34 | brlcad | packet size 1 through 2048, then 2048 to 4194304 in increments of 2048 .. all using a randome bs=1024 count=8192 file |
| 19:27.26 | andrei_ | I have started the script for 1 -> 2048 |
| 20:03.33 | *** join/#brlcad andrei_ (~andrei@188.25.172.146) | |
| 20:04.44 | andrei_ | brlcad, there s some wierd issue |
| 20:04.52 | andrei_ | server starts on remote |
| 20:05.01 | andrei_ | when client starts it s stuck on connecting to ... |
| 20:05.15 | andrei_ | I rebooted both stations, changed sockets. |
| 20:05.59 | andrei_ | any clue what might cause this? |
| 20:06.39 | brlcad | does it hang if you replace 192.168.1.100 with localhost? (don't forget to rename log_file to server_log_file and client_log_file to differentiate) |
| 20:09.38 | andrei_ | I get connection refused |
| 20:14.13 | andrei_ | I ll find the issue. |
| 20:33.43 | andrei_ | brlcad |
| 20:33.55 | andrei_ | it ...works....apparently reading 8mb byte by byte does take a while. |
| 20:42.57 | andrei_ | brlcad, what should I do . 1 - 2048 simply takes too long. |
| 20:59.38 | andrei_ | there are a couple of issues, will fix them in a few hours. I need to rethink the whole script. |
| 21:01.29 | andrei_ | this is the current status : http://slexy.org/view/s2IbZHz6oY . Sometimes it does not connect properly and it takes way too long to process the parameters you suggested. |