| 00:19.48 | *** join/#brlcad infobot (ibot@rikers.org) | |
| 00:19.48 | *** topic/#brlcad is BRL-CAD and open source CAx discussion ! Also @ http://brlcad.zulipchat.com ! Logs @ http://infobot.rikers.org/%23brlcad/ | |
| 05:43.22 | *** join/#brlcad infobot (ibot@rikers.org) | |
| 05:43.22 | *** topic/#brlcad is BRL-CAD and open source CAx discussion ! Also @ http://brlcad.zulipchat.com ! Logs @ http://infobot.rikers.org/%23brlcad/ | |
| 05:47.26 | *** join/#brlcad zwliew (uid161395@gateway/web/irccloud.com/x-ggdhjfcwwaqbenja) | |
| 07:11.16 | *** join/#brlcad raingloo1 (~raingloom@C2267141.catv.pool.telekom.hu) | |
| 07:58.12 | *** join/#brlcad merzo (~merzo@123-57-133-95.pool.ukrtel.net) | |
| 09:46.55 | *** join/#brlcad teepee (~teepee@unaffiliated/teepee) | |
| 10:08.39 | *** join/#brlcad teepee- (bc5c2134@gateway/web/freenode/ip.188.92.33.52) | |
| 10:20.55 | *** join/#brlcad teepee (~teepee@unaffiliated/teepee) | |
| 11:06.30 | *** join/#brlcad teepee (~teepee@unaffiliated/teepee) | |
| 11:58.57 | *** join/#brlcad merzo (~merzo@185.39.197.205) | |
| 12:07.41 | *** join/#brlcad teepee (~teepee@unaffiliated/teepee) | |
| 12:34.01 | *** join/#brlcad yorik (~yorik@2804:431:f721:327d:290:f5ff:fedc:3bb2) | |
| 13:45.50 | *** join/#brlcad teepee (~teepee@unaffiliated/teepee) | |
| 14:30.26 | brlcad | dang it.. looks like remrt/rtsrv isn't quite cooperating correctly |
| 15:01.56 | brlcad | aaand, I'm the culprit |
| 15:22.57 | brlcad | looks like there's at least a workaround, but fixing it now |
| 15:56.55 | *** join/#brlcad raingloom (~raingloom@C2267141.catv.pool.telekom.hu) | |
| 19:15.51 | *** join/#brlcad raingloom (~raingloom@C2267141.catv.pool.telekom.hu) | |
| 19:38.18 | *** join/#brlcad infobot (ibot@rikers.org) | |
| 19:38.18 | *** topic/#brlcad is BRL-CAD and open source CAx discussion ! Also @ http://brlcad.zulipchat.com ! Logs @ http://infobot.rikers.org/%23brlcad/ | |
| 19:49.33 | starseeker | brlcad: what's the error on non x86? |
| 19:49.48 | starseeker | we're trying to call the wrong logic? |
| 20:13.56 | brlcad | it was a few issues, mismatch between the decl and the impl |
| 20:14.41 | brlcad | impl limited to x86 in mm code, limited to windows in tc code |
| 20:18.27 | starseeker | ah. |
| 20:18.50 | starseeker | brlcad: do you want me to toggle the isst code and the gct code with one flag? |
| 20:18.51 | brlcad | patched up my specific error but main issue remains that mm isn't very portabile, we should have some way to toggle off gct |
| 20:19.10 | brlcad | isst's using something? haven't hit an error there lately |
| 20:19.15 | starseeker | (isst is using tc as well) |
| 20:19.23 | starseeker | or adrt is |
| 20:19.28 | starseeker | something in there... one sec |
| 20:19.46 | starseeker | tienet.h |
| 20:20.32 | starseeker | tinet.c and master/tienet_master.c are the .c users |
| 20:21.16 | brlcad | so far tc hasn't really bitten us since usage is pretty limited -- the issue was limited to missing implementation of trylock |
| 20:21.20 | brlcad | for pthreading |
| 20:21.31 | starseeker | OK, so GCT only for the moment |
| 20:21.47 | brlcad | that's there now, so tc's probably fine for now.. at least until new threads |
| 20:22.04 | starseeker | hopefully TBB will make the whole issue moot |
| 20:22.29 | brlcad | gct just doubles up with hardware/assembly or .. tc threading |
| 20:22.30 | starseeker | is Stragus's more modern mmthread.h more portable? I hadn't dug that deeply on it yet |
| 20:23.35 | brlcad | I haven't looked into it that closely, but a good test would be a non-desktop system |
| 20:24.12 | brlcad | like a Power system or AIX or just about anything HPC |
| 20:25.11 | Stragus | It's gcc/clang/MSVC (including very old MSVC, like 2003 or whatever), pthreads and winthreads. The only broken part that I'm aware of is providing a stack for Winthreads (the API doesn't support it) |
| 20:25.18 | brlcad | this is admittedly a bit frustrating ... I needed to run remrt real quick on some bigger iron and none of them are working for various reasons |
| 20:25.38 | brlcad | systems that have worked, but not tested frequently |
| 20:26.05 | Stragus | It should work with other compilers on POSIX systems, but that's untested |
| 20:26.20 | brlcad | Stragus: my woes are not entirely this issue, just fyi |
| 20:27.39 | brlcad | it's more that we need a CI telling us when certain systems stop working, not finding out the day I need them to be working :) |
| 20:37.09 | Stragus | Okay :) |
| 20:37.20 | starseeker | brlcad: for AIX that means the GCC farm I guess? |
| 20:37.44 | starseeker | wonders how POSIX-ish AIX is... |
| 20:38.12 | brlcad | farm or hpc |
| 20:38.28 | brlcad | I believe there's an aix in there |
| 20:38.51 | brlcad | but definitely a whole fleet of them in the farm, all the biggest ones with 64-160 cores |
| 20:42.02 | brlcad | I'm going to be doing a production render this week if you have some cores to spare and an rtsrv compiled up |
| 20:52.38 | starseeker | brlcad: OK, cmake -DBRLCAD_ENABLE_GCT=OFF should do the trick |
| 20:52.50 | starseeker | r70913 |
| 22:07.11 | brlcad | awesome! |
| 22:16.39 | LoH | starseeker: ick, aix |
| 22:34.32 | *** join/#brlcad merzo (~merzo@36-53-132-95.pool.ukrtel.net) | |
| 22:37.28 | *** join/#brlcad raingloom (~raingloom@C2267141.catv.pool.telekom.hu) | |
| 23:33.45 | *** join/#brlcad yorik (~yorik@2804:431:f721:347:290:f5ff:fedc:3bb2) | |