From schnetter at cct.lsu.edu Sat Mar 6 14:10:56 2010 From: schnetter at cct.lsu.edu (Erik Schnetter) Date: Sat, 6 Mar 2010 14:10:56 -0600 Subject: [Developers] Scheduling HydroBase's ExcisionMaskSetup Message-ID: <74267A63-5193-4805-8653-DC8DACA0FE95@cct.lsu.edu> I think HydroBase's ExcisionMaskSetup group could be scheduled in MoL_PseudoEvolution. This would simplify the schedule. -erik -- Erik Schnetter http://www.cct.lsu.edu/~eschnett/ -------------- next part -------------- A non-text attachment was scrubbed... Name: smime.p7s Type: application/pkcs7-signature Size: 2712 bytes Desc: not available Url : http://www.cactuscode.org/pipermail/developers/attachments/20100306/475ff89c/attachment.bin From schnetter at cct.lsu.edu Sat Mar 27 08:05:01 2010 From: schnetter at cct.lsu.edu (Erik Schnetter) Date: Sat, 27 Mar 2010 09:05:01 -0500 Subject: [Developers] [Carpet] Carpet 2D HDF5 output overwritten after recovery? In-Reply-To: References:

Message-ID: <614CA2F8-1EBE-4B28-B98A-C62C98674E48@cct.lsu.edu> Bruno I committed the change. -erik On Mar 24, 2010, at 14:19 , Bruno Giacomazzo wrote: > Hi, > was the attached patch ever committed to Carpet? It seems not since after downloading Carpet from scratch I had again the same problems with the 2D HDF5 files being overwritten when restarting from checkpoint (i.e. all the 2D data produced form the previous run were deleted). > > Is there a reason for which it was not applied? In my case it worked. > > Thank you, > Bruno > > On Aug 5, 2009, at 10:54 AM, Erik Schnetter wrote: > >> Bruno >> >> It seems that Carpet's logic to detect whether there is an existing valid HDF5 file is flawed. Can you try the attached patch? >> >> -erik >> >> -- >> Erik Schnetter http://www.cct.lsu.edu/~eschnett/ >> >> My email is as private as my paper mail. I therefore support encrypting >> and signing email messages. Get my PGP key from . >> >> >> _______________________________________________ >> developers mailing list >> developers at lists.carpetcode.org >> http://lists.carpetcode.org/listinfo/developers > > Dr. Bruno Giacomazzo > Department of Astronomy > University of Maryland > College Park > MD 20742-2421 > USA > > Tel. : +1 301 405 0676 > Fax : +1 301 314 9067 > Tel. : +1 301 286 3809 (NASA/GSFC) > Fax : +1 301 286 7230 (NASA/GSFC) > Cell. : +1 443 535 5846 > email : bgiacoma at astro.umd.edu > > ---------------------------------------------------------------------- > There are only 10 types of people in the world: > Those who understand binary, and those who don't > ---------------------------------------------------------------------- > > -- Erik Schnetter http://www.cct.lsu.edu/~eschnett/ -------------- next part -------------- A non-text attachment was scrubbed... Name: smime.p7s Type: application/pkcs7-signature Size: 2712 bytes Desc: not available Url : http://www.cactuscode.org/pipermail/developers/attachments/20100327/13b4deb0/attachment.bin From roland.haas at physics.gatech.edu Mon Mar 29 08:43:17 2010 From: roland.haas at physics.gatech.edu (Roland Haas) Date: Mon, 29 Mar 2010 10:43:17 -0400 Subject: [Developers] global-early during ANALYSIS using Carpet Message-ID: <4BB0BC85.9020209@mail.gatech.edu> Hello all, it seems to me as if the implementation of global-early during ANALYSIS in the commit e91c85225521026738519d1d81455b22bbdcb2ee might not be correct; a routine that I scheduled global-early was called on every refinement level instead of only one. It seems as if the condition in line 528 have_done_global_mode |= do_global_mode should use do_early_global_mode instead. Attached please find a patch which changes this (and renames have_done_global_mode to have_done_early_global_mode since the global mode is late in ANALYSIS). Yours, Roland -- My email is as private as my paper mail. I therefore support encrypting and signing email messages. Get my PGP key from http://keys.gnupg.net. -------------- next part -------------- A non-text attachment was scrubbed... Name: 0001-fix-global_early-in-ANALYSIS.patch Type: text/x-patch Size: 2052 bytes Desc: not available Url : http://www.cactuscode.org/pipermail/developers/attachments/20100329/494408d2/attachment.bin -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 189 bytes Desc: OpenPGP digital signature Url : http://www.cactuscode.org/pipermail/developers/attachments/20100329/494408d2/attachment-0001.bin From schnetter at cct.lsu.edu Mon Mar 29 12:28:02 2010 From: schnetter at cct.lsu.edu (Erik Schnetter) Date: Mon, 29 Mar 2010 13:28:02 -0500 Subject: [Developers] global-early during ANALYSIS using Carpet In-Reply-To: <4BB0BC85.9020209@mail.gatech.edu> References: <4BB0BC85.9020209@mail.gatech.edu> Message-ID: On Mar 29, 2010, at 9:43 , Roland Haas wrote: > Hello all, > > it seems to me as if the implementation of global-early during > ANALYSIS in the commit e91c85225521026738519d1d81455b22bbdcb2ee > might not be correct; a routine that I scheduled global-early was > called on every refinement level instead of only one. It seems as if > the condition in line 528 > > have_done_global_mode |= do_global_mode > > should use do_early_global_mode instead. > > Attached please find a patch which changes this (and renames > have_done_global_mode to have_done_early_global_mode since the > global mode is late in ANALYSIS). Roland Thanks for the patch. I have applied a similar patch with a bit more error checking. -erik -- Erik Schnetter http://www.cct.lsu.edu/~eschnett/ From schnetter at cct.lsu.edu Mon Mar 29 15:12:58 2010 From: schnetter at cct.lsu.edu (Erik Schnetter) Date: Mon, 29 Mar 2010 16:12:58 -0500 Subject: [Developers] Fwd: Speed up USES SOURCE mechanism References: <200401111429.20841.schnetter@uni-tuebingen.de> Message-ID: The issue mentioned below has come up again. Should I apply the suggested patch, or a modernised version of it? -erik Begin forwarded message: > From: Erik Schnetter > Date: January 11, 2004 7:29:20 CST > To: Cactus Developers > Subject: [Developers] Speed up USES SOURCE mechanism > > -----BEGIN PGP SIGNED MESSAGE----- > Hash: SHA1 > > The USES SOURCE mechanism currently calls CCTK_IsThornActive at the > beginning of every included section. This is a very expensive routine > that traverses the tree of active thorns and performs many string > comparisons. This is e.g. the case for BSSN_MoL with Whisky. More > than 1/4 of the total run time is spent comparing strings. > > The patch below introduces "thorn handles" (which are pointers to the > thorn structure). It also creates a wrapper function for each thorn > that caches this handle. Okay to apply? > > - -erik > > - -- > Erik Schnetter http://www.aei.mpg.de/~eschnett/ > > My email is as private as my paper mail. I therefore support > encrypting > and signing email messages. Get my PGP key from www.keyserver.net. > -----BEGIN PGP SIGNATURE----- > Version: GnuPG v1.2.4 (GNU/Linux) > > iD4DBQFAAU+wm3uiSwno3f0RAiNOAJ9a402/StgiyNqx5r40HjHsgd3i/gCXdYuk > Mxtk3uapQ6tiCYljdX4qvg== > =76ce > -----END PGP SIGNATURE----- -------------- next part -------------- A non-text attachment was scrubbed... Name: usessource-lib.diff Type: text/x-diff Size: 13685 bytes Desc: not available Url : http://www.cactuscode.org/pipermail/developers/attachments/20100329/a6bc2218/attachment-0002.bin -------------- next part -------------- A non-text attachment was scrubbed... Name: usessource-src.diff Type: text/x-diff Size: 14256 bytes Desc: not available Url : http://www.cactuscode.org/pipermail/developers/attachments/20100329/a6bc2218/attachment-0003.bin -------------- next part -------------- > _______________________________________________ > Developers mailing list > Developers at cactuscode.org > http://www.cactuscode.org/mailman/listinfo/developers -- Erik Schnetter http://www.cct.lsu.edu/~eschnett/ From schnetter at cct.lsu.edu Mon Mar 29 15:18:39 2010 From: schnetter at cct.lsu.edu (Erik Schnetter) Date: Mon, 29 Mar 2010 16:18:39 -0500 Subject: [Developers] [Patches] TimerReport: Output more than one clock In-Reply-To: <2B18B574-49B6-4A69-AF51-36834D662089@cct.lsu.edu> References: <2B18B574-49B6-4A69-AF51-36834D662089@cct.lsu.edu> Message-ID: On Jan 18, 2010, at 14:18 , Erik Schnetter wrote: > Currently, thorn CactusUtils/TimerReport outputs one clock only for > each timer. This clock can be selected with the parameter > all_timers_clock. This makes the output much more readable, since > there may be many clocks. > > The enclosed patch adds a new array parameter all_timers_clocks[] > that allows multiple clocks to be specified. They are output into a > different file each, with the file name containing the clock name. > This affects the output_all_timers, output_all_timers_together, and > output_all_timers_readable functionality. > > Implementation-wise, there is now a for loop surrounding the output > code that traverses all clocks. > > This makes is possible to output e.g. the wall time as well as > several PAPI clocks. I solicitate input and comments regarding this patch. This patch makes it much easier to output several clocks at once, which is important e.g. for PAPI profiling. Ian Hinder, do you want to comment since you have experience with this thorn? -erik -- Erik Schnetter http://www.cct.lsu.edu/~eschnett/ From knarf at cct.lsu.edu Mon Mar 29 16:01:46 2010 From: knarf at cct.lsu.edu (Frank Loeffler) Date: Mon, 29 Mar 2010 17:01:46 -0500 Subject: [Developers] Fwd: Speed up USES SOURCE mechanism In-Reply-To: References: <200401111429.20841.schnetter@uni-tuebingen.de> Message-ID: <20100329220146.GJ7849@numrel07.cct.lsu.edu> On Mon, Mar 29, 2010 at 04:12:58PM -0500, Erik Schnetter wrote: > The issue mentioned below has come up again. Should I apply the > suggested patch, or a modernised version of it? Applying a patch like this would certainly be a good idea. In the end it might not matter for future production runs because they don't use the old CalcTmunu interface (probably the most prominent user of the include mechanism), but there are other uses of that mechanism which are unlikely to go away any time soon and which might be or become performance critical. Frank -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 197 bytes Desc: not available Url : http://www.cactuscode.org/pipermail/developers/attachments/20100329/50b2c853/attachment.bin