| 00:08.49 | ``Erik | NARF! |
| 00:13.00 | tegtmeye | eh! which one of our machines has the most memory? |
| 00:13.21 | tegtmeye | rack has 8 and I jsut killed it... |
| 00:18.38 | ``Erik | um, the altix has like 24g |
| 00:19.20 | tegtmeye | I don't think I have an account on wopr |
| 00:19.31 | tegtmeye | can you give me one from home? |
| 00:19.35 | ``Erik | no |
| 00:19.37 | brlcad | not really public conversation material |
| 00:19.39 | ``Erik | it's not mine |
| 00:19.46 | ``Erik | and we dont' say actual machine names in public |
| 00:19.53 | tegtmeye | gotcha |
| 00:20.35 | ``Erik | the second most I can think of off the top of my head is the amd64 minicluster with 8g/node |
| 00:21.08 | tegtmeye | yea, jsut killed one with 8 |
| 00:21.12 | tegtmeye | stupid perl |
| 00:21.17 | brlcad | heh |
| 00:21.18 | ``Erik | didja cook up much swap on it? |
| 00:21.19 | brlcad | bad regex |
| 00:21.21 | ``Erik | heh, perl sucks :D |
| 00:21.56 | brlcad | if it wasn't a regex, I don't want to know how you ran through that much memory |
| 00:22.12 | tegtmeye | graph stuff |
| 00:22.17 | ``Erik | perl is as perl does? :) |
| 00:22.26 | brlcad | if it was, you should be able to cut that down with different expressions |
| 00:22.46 | brlcad | ahm.. storing sparse matrices in hash tables? :) |
| 00:22.48 | tegtmeye | no regex at all |
| 00:23.08 | tegtmeye | no |
| 00:23.14 | tegtmeye | adjacency lists |
| 00:23.17 | tegtmeye | sp |
| 00:24.16 | brlcad | perl isn't "horrible" on memory usage generally speaking.. not as bad as some other languages at least; sure you're not doing something that's causing some explosive allocation? |
| 00:24.18 | tegtmeye | Mar 27 19:06:10 ?? kernel: Node 0 Normal: 0*4kB 0*8kB 1*16kB 0*32kB 1*64kB 1*128kB 1*256kB 1*51 |
| 00:24.21 | tegtmeye | 2kB 1*1024kB 0*2048kB 0*4096kB = 2000kB |
| 00:24.24 | tegtmeye | Mar 27 19:06:10 ?? kernel: Node 0 HighMem: empty |
| 00:24.26 | tegtmeye | Mar 27 19:06:10 ?? kernel: Swap cache: add 516944, delete 516944, find 255/1278, race 0+0 |
| 00:24.30 | tegtmeye | Mar 27 19:06:10 ?? kernel: Free swap: 0kB |
| 00:24.32 | tegtmeye | Mar 27 19:06:10 ?? kernel: 2097150 pages of RAM |
| 00:24.35 | tegtmeye | Mar 27 19:06:10 ?? kernel: 249625 reserved pages |
| 00:24.37 | tegtmeye | Mar 27 19:06:10 ?? kernel: 6527 pages shared |
| 00:24.40 | tegtmeye | Mar 27 19:06:10 ?? kernel: 0 pages swap cached |
| 00:24.42 | tegtmeye | Mar 27 19:06:10 ?? kernel: Out of Memory: Killed process 31308 (bash). |
| 00:24.55 | tegtmeye | yea, doing static analysis on doxygen... |
| 00:25.00 | tegtmeye | it has like 20k functions |
| 00:25.21 | ``Erik | erm, y'know, we're only allowed to use those computers for official work related business o.O |
| 00:25.26 | ``Erik | :) |
| 00:25.50 | brlcad | on doxygen?? |
| 00:26.21 | brlcad | it's a document tool, i could care less how poorly written it is :) |
| 00:26.28 | brlcad | but then that's just me |
| 00:26.47 | ``Erik | doesn't it use crap from like kde or qt? |
| 00:27.00 | brlcad | not last I payed attention |
| 00:27.08 | brlcad | at least not required |
| 00:48.52 | ``Erik | ahhh O.o |
| 00:50.00 | pra5ad | sonofa.. oblivion is xp only too |
| 00:50.04 | pra5ad | this is ridiculous |
| 03:13.41 | *** join/#brlcad pra5ad (n=pra5ad@pool-141-157-106-216.balt.east.verizon.net) | |
| 03:19.28 | *** join/#brlcad pra5ad (n=pra5ad@pool-141-157-106-216.balt.east.verizon.net) | |
| 03:20.29 | *** join/#brlcad pra5ad (n=pra5ad@pool-141-157-106-216.balt.east.verizon.net) | |
| 03:21.05 | *** join/#brlcad pra5ad (n=pra5ad@pool-141-157-106-216.balt.east.verizon.net) | |
| 03:21.28 | *** join/#brlcad pra5ad (n=pra5ad@pool-141-157-106-216.balt.east.verizon.net) | |
| 03:21.52 | *** join/#brlcad pra5ad (n=pra5ad@pool-141-157-106-216.balt.east.verizon.net) | |
| 03:22.17 | *** join/#brlcad pra5ad (n=pra5ad@pool-141-157-106-216.balt.east.verizon.net) | |
| 03:22.41 | *** join/#brlcad pra5ad (n=pra5ad@pool-141-157-106-216.balt.east.verizon.net) | |
| 03:23.16 | *** join/#brlcad pra5ad (n=pra5ad@pool-141-157-106-216.balt.east.verizon.net) | |
| 03:23.41 | *** join/#brlcad pra5ad (n=pra5ad@pool-141-157-106-216.balt.east.verizon.net) | |
| 03:24.04 | *** join/#brlcad pra5ad (n=pra5ad@pool-141-157-106-216.balt.east.verizon.net) | |
| 03:24.35 | brlcad | grr |
| 03:25.05 | *** join/#brlcad pra5ad (n=pra5ad@pool-141-157-106-216.balt.east.verizon.net) | |
| 03:25.22 | *** mode/#brlcad [+b pra5ad!pra5ad@pool-141-157-106-216.balt.east.verizon.net] by brlcad | |
| 03:25.28 | *** join/#brlcad pra5ad (n=pra5ad@pool-141-157-106-216.balt.east.verizon.net) | |
| 03:25.34 | brlcad | hrmph |
| 03:25.40 | *** mode/#brlcad [-b pra5ad!pra5ad@pool-141-157-106-216.balt.east.verizon.net] by brlcad | |
| 03:25.53 | *** join/#brlcad pra5ad (n=pra5ad@pool-141-157-106-216.balt.east.verizon.net) | |
| 03:26.22 | *** mode/#brlcad [+b pra5ad*!*@*] by brlcad | |
| 03:26.55 | brlcad | sheesh |
| 03:29.49 | *** mode/#brlcad [-b pra5ad*!*@*] by brlcad | |
| 03:30.08 | *** join/#brlcad pra5ad (n=pra5ad@pool-141-157-106-216.balt.east.verizon.net) | |
| 03:33.36 | pra5ad | ok so im trying to use tinyxml for the c++ gl test |
| 03:33.49 | pra5ad | its makefile only produces obj files |
| 03:34.00 | pra5ad | how do i incorporate that into my gnu build system? |
| 03:34.05 | pra5ad | or whats the best way |
| 03:36.00 | brlcad | who would only make .o files |
| 03:36.36 | brlcad | you can half-ass it and just add them to the LIBS or LDFLAGS |
| 03:36.45 | pra5ad | *** i have a 15% gdc discount at www.akpeters.com (till 4/30/06) ... let me know if u are planning to get some books from that site |
| 03:36.47 | brlcad | or convert the build to a gbs one |
| 03:37.33 | pra5ad | hm |
| 03:38.59 | brlcad | actually yeah, http://www.akpeters.com/product.asp?ProdCode=2175 |
| 03:43.01 | pra5ad | since i'm not selling this gl test prog, i can technically rip out the src (3 cpp files) and put em under my build, right? |
| 03:43.10 | brlcad | mm.. http://www.akpeters.com/product.asp?ProdCode=1772 |
| 03:43.13 | pra5ad | iirc its under gpl |
| 03:43.56 | pra5ad | dude, email me the list of books |
| 03:44.11 | pra5ad | im sorta curious about em too |
| 03:49.23 | brlcad | only other one maybe, http://www.akpeters.com/product.asp?ProdCode=0903 |
| 03:52.29 | brlcad | sent |
| 04:04.48 | Twingy | unf |
| 04:04.54 | Twingy | soooo many boards |
| 04:21.45 | Maloeran | Eh, coerced ``Erik into giving a hand yet? |
| 04:28.36 | Twingy | he's been sick for like 10 days? |
| 04:28.44 | Twingy | and refuses to get antibiotics |
| 04:32.36 | Maloeran | Ouch, so he's still sick... |
| 04:33.01 | Twingy | he should be using sick days :) |
| 04:33.14 | Twingy | with all the sick people I've been around at work |
| 04:33.20 | Twingy | you'd think I'd be on my death bed |
| 04:33.53 | pra5ad | gym is my antibiotic |
| 04:34.21 | Maloeran | Sure, regular exercise and a good cardiovascular system greatly reinforces the immunity system |
| 04:34.48 | Maloeran | I haven't been sick for... many years either |
| 04:35.59 | pra5ad | so when do i get to see ur raytracer? |
| 04:36.23 | Maloeran | Hrm, what about at Siggraph? :) |
| 04:36.35 | Maloeran | I should have dynamic geometry support nicely completed by then |
| 04:36.42 | pra5ad | oh yea? nice |
| 04:37.04 | pra5ad | what are u using as the spatial structure? kinetic octree? |
| 04:37.13 | Maloeran | Not a known technique |
| 04:37.18 | pra5ad | ohh |
| 04:37.20 | pra5ad | heh =) |
| 04:37.21 | Maloeran | The ray-triangle intersection test is a new technique too |
| 04:38.38 | Maloeran | pra5ad, I vaguely understood you were interested by low-level hardware, have you played with FPGAs? |
| 04:39.10 | pra5ad | sure have |
| 04:39.14 | brlcad | you're presenting a paper at siggraph? |
| 04:39.16 | pra5ad | xilinx mostly |
| 04:39.33 | pra5ad | done a little abel and verliog |
| 04:40.31 | brlcad | or going to demo it at the brl-cad bof? :) |
| 04:40.46 | Twingy | no no |
| 04:40.52 | Twingy | the Justin is your God bof |
| 04:40.58 | Twingy | :} |
| 04:41.03 | brlcad | we need better compute power this year |
| 04:41.10 | Twingy | I got that covered |
| 04:41.17 | brlcad | heh |
| 04:41.28 | Twingy | mal, I'll have a dual cpu dual core 2ghz opteron there for yah |
| 04:41.30 | brlcad | should still.. press for new equipment by then |
| 04:42.05 | brlcad | hopefully it'll arive in time |
| 04:42.13 | Twingy | the craptops? |
| 04:42.14 | brlcad | i needs it, i vants it |
| 04:42.35 | pra5ad | help me push for a dual cell blade |
| 04:42.38 | pra5ad | only $50k |
| 04:42.39 | pra5ad | =) |
| 04:43.32 | Twingy | heh, in that case let me just cash out my trust fund |
| 04:58.29 | pra5ad | http://www.flickr.com/photos/ps3/117907118/ |
| 04:58.46 | pra5ad | warhawk on ps3; clouds rendered using raycasting |
| 04:58.50 | pra5ad | on a single spu |
| 05:07.19 | Maloeran | Sorry about that, phone call from friend lawyer. So pra5ad, ever pondered about putting together ray-tracing hardware? :} |
| 05:08.38 | pra5ad | my interest has peaked recently |
| 05:09.00 | pra5ad | would be a fun project :) |
| 05:09.26 | Maloeran | For my part, I am really interested. I'm the kind of guy who write assembly for any common archs and stomp compilers, but I never went below that |
| 05:10.09 | Maloeran | As it was for rasterization many years ago, specialized hardware will be the key for ray-tracing |
| 05:10.21 | pra5ad | i agree |
| 05:11.09 | pra5ad | wonder if xilinx is selling a pci-x fpga prototyping board |
| 05:12.19 | Maloeran | This sure would be appropriate, such hardware would require pci-x's bandwidth |
| 05:13.09 | Maloeran | What emulation software do you use or would recommend? |
| 05:13.10 | pra5ad | personally id like to try out two things: the pci-x path or build it against cell |
| 05:13.35 | pra5ad | heh i cheated |
| 05:13.40 | pra5ad | xilinx's own tools |
| 05:13.55 | pra5ad | i dont believe they have any for nix/osx |
| 05:14.10 | pra5ad | then again this was 3 yrs ago |
| 05:14.21 | Maloeran | Mmhm, I see |
| 05:14.46 | pra5ad | the one we used in 2k4 was pretty powerful imo |
| 05:14.54 | pra5ad | full system simulation |
| 05:15.04 | pra5ad | and our schematic was pretty complex |
| 05:15.15 | pra5ad | twas a DLA |
| 05:15.39 | pra5ad | iirc it allowed us to create components and tie em together; building blocks |
| 05:15.45 | pra5ad | and simulate it that way |
| 05:16.12 | pra5ad | and i do believe it cost around the order of $200 |
| 05:16.35 | pra5ad | (project budget was somewhere around that) |
| 05:16.39 | Maloeran | Interesting. What software would you recommend for an initiation into the path that would lead to efficient ray-tracing hardware many years from now? |
| 05:17.54 | Maloeran | Most preferabily nix software... :) |
| 05:17.59 | Maloeran | preferably, even |
| 05:18.07 | pra5ad | i really shouldnt have any say in this, but id stick with the hardware vendor tools |
| 05:18.12 | pra5ad | ah.. that i dont know |
| 05:18.23 | pra5ad | never really did look into nix stuff |
| 05:18.52 | pra5ad | i believe altera's basic dev kits run about $150-200 |
| 05:19.03 | pra5ad | includes vhdl compilers |
| 05:19.05 | pra5ad | iirc |
| 05:20.16 | Maloeran | I see, thanks for the tip |
| 05:22.29 | pra5ad | ah yea, we used the spartan 3E's |
| 05:24.14 | pra5ad | ohh interesting |
| 05:24.22 | pra5ad | logic building tools |
| 05:24.23 | pra5ad | System Requirements |
| 05:24.23 | pra5ad | <PROTECTED> |
| 05:24.23 | pra5ad | <PROTECTED> |
| 05:24.23 | pra5ad | <PROTECTED> |
| 05:24.29 | pra5ad | might wanna try that |
| 05:25.05 | Maloeran | Nice, that seems more convenient |
| 05:25.13 | pra5ad | http://www.xilinx.com/ise/logic_design_prod/foundation.htm |
| 05:25.20 | pra5ad | evaluation cd |
| 05:25.24 | pra5ad | heh full version is 2.5k |
| 05:26.41 | Maloeran | Ouch. I really want to explore this avenue soon, after I'm done with the software side |
| 05:28.27 | pra5ad | http://www.xilinx.com/xlnx/xebiz/designResources/ip_product_details.jsp?key=HW-V4-ML455&sGlobalNavPick=PRODUCTS&sSecondaryNavPick=BOARDS |
| 05:28.32 | pra5ad | pci-x dev kit |
| 05:28.38 | pra5ad | for virtex-4s |
| 05:28.46 | pra5ad | 1.2k |
| 05:30.55 | pra5ad | heh now i want to get a dev kit |
| 05:30.58 | Maloeran | Eh, I think I can be satisfied with software emulation for the moment, to gain some experience with this strange new world |
| 05:31.44 | pra5ad | it's really not that hard |
| 05:31.52 | pra5ad | id say it's easier than asm |
| 05:31.54 | pra5ad | much easier |
| 05:33.29 | pra5ad | so i get to learn about this new raytracing algorithm in august eh? |
| 05:34.17 | Maloeran | I'm afraid it's unlikely that you will learn more than what level of performance it reaches... :) |
| 05:36.55 | pra5ad | collaboration? =) |
| 05:38.52 | Maloeran | That might be an option, I'm really unsure of how things will unfold at the moment |
| 05:39.15 | pra5ad | ah, ok |
| 06:01.33 | Twingy | damn I'm getting close to producing working g-code |
| 06:16.53 | Twingy | bed time |
| 06:46.01 | *** join/#brlcad Twingy (n=justin@c-68-33-163-43.hsd1.md.comcast.net) | |
| 15:41.12 | CIA-5 | BRL-CAD: 03lbutler * 10brlcad/regress/master_prep.sh: |
| 15:41.12 | CIA-5 | BRL-CAD: Added --enable-everything to the configure line |
| 15:41.12 | CIA-5 | BRL-CAD: Printing old and new version numbers for confirmation |
| 15:46.56 | docelic | hi folks.. any of sourceforge project owners here ? |
| 15:47.20 | docelic | s/owners/brlcad admins/ |
| 15:47.39 | docelic | haha good :) |
| 15:48.07 | ``Erik | o.O |
| 15:49.42 | docelic | I am working on a large project too, and we are thinking of opening a sourceforge project.. but if I remember, brlcad was specially announced on sf.net frontpage, so maybe such a deal could be made with us as well |
| 15:50.04 | docelic | I'd like to find out the people I could talk to about such possibility |
| 15:53.43 | CIA-5 | BRL-CAD: 03lbutler * 10brlcad/regress/Makefile.am: Now does all the regression tests even if one fails |
| 15:56.17 | ``Erik | I don't think anything special was done for that... we had a cvs repo that we wanted imported |
| 15:57.14 | ``Erik | so brlcad over there *point* talked to them and gave them a tarball of goop, as well as sending a blurb to slashdot... I d'no if he did anything special for sf news, or if they just did it themselves |
| 19:01.51 | brlcad | docelic: yes |
| 19:02.25 | docelic | hi brlcad |
| 19:02.55 | brlcad | sf.net picks up news for their main page at their discretion based on when you post news to your project :) |
| 19:03.55 | docelic | so you did not make any special arrangements to have it appear there? I might be wrong, but I remember them having the 1:1 picture of your logo on the main site for a week |
| 19:04.17 | brlcad | really? |
| 19:04.22 | brlcad | i never knew that |
| 19:05.10 | brlcad | so yeah.. no special arrangements other than being a high-profile popular newsworthy project ;) |
| 19:06.17 | docelic | interesting.. |
| 19:08.35 | brlcad | likewise |
| 19:08.54 | brlcad | i could just be forgetting, but I didn't know about the sf.net news |
| 19:27.20 | brlcad | time to get out the hose |
| 20:02.48 | CIA-5 | BRL-CAD: 03bob1961 * 10brlcad/src/tclscripts/archer/Archer.tcl: Set the default button for the preferences dialog. |
| 20:12.21 | CIA-5 | BRL-CAD: 03lbutler * 10brlcad/regress/master_prep.sh: |
| 20:12.21 | CIA-5 | BRL-CAD: Fixed bug where master_prep.sh was being run multiple times and |
| 20:12.21 | CIA-5 | BRL-CAD: alternating between release numbers. Now gets consistent release |
| 20:12.22 | CIA-5 | BRL-CAD: number from README file |