griefith/test
1
0
Fork 0
Code Issues Pull Requests Actions 1 Packages Projects Releases Wiki Activity
53,581 Commits 1 Branch 0 Tags
4ed16bc3e69dcdbc9e4e2cd522f380274dd939f7
Commit Graph

2759 Commits

Author SHA1 Message Date
Campbell Barton
3e3efae7e9 Viewport Text Drawing: replace single allocs with a memarena 
- pass label strlen since in many cases its already known.
- use single linked list for cached text drawing.
- add BLI_link_utils.h for single linked list macros.

own tests give approx 22% overall speedup.
 2014-04-17 16:04:28 +10:00
Campbell Barton
1b9db9911d Code cleanup: use bools 
also rename BLI_omp_thread_count -> BLI_system_thread_count_omp
 2014-04-17 16:04:28 +10:00
Campbell Barton
b3972aeea0 Math Lib: optimize axis_dominant_v3_to_m3, approx 6x speedup 
build the matrix directly rather then calculating with axis/angle

also remove unused function calc_poly_plane
 2014-04-16 21:07:28 +10:00
Campbell Barton
233dac1494 Math Lib: increase epsilon for ortho_basis_v3v3_v3 
passing in a unit length vector wouldn't always compute unit length vectors because the epsilon tested was too small.
 2014-04-16 21:07:28 +10:00
Campbell Barton
28a829893c Math Lib: avoid having to pass a pre-calculated normal tot area_poly_v3 
add normal_poly_v3
 2014-04-16 00:29:57 +10:00
Bastien Montagne
d9211b1e7b Fix compilation (size_t not defined) by including stddef.h in BLI_sys_types.h 2014-04-15 09:11:45 +02:00
Campbell Barton
aa10cf7f5c GHash: add BLI_ghashutil_uinthash_v4 for hashing 4 ints at once 2014-04-15 14:40:33 +10:00
Campbell Barton
a7241d09cd GHash: add typed hash functions (were all (void *)) 
- BLI_ghashutil_strhash_n takes string length, to avoid terminating the string before hashing.
- BLI_ghashutil_inthash/uinthash take ints, to avoid casting to (void *)

This also showed up incorrect use of inthash, which was using a pointer.
 2014-04-15 14:22:36 +10:00
Campbell Barton
ad9af56a6e LinkList stack macros to swap and pop into a separate stack. 2014-04-13 20:12:07 +10:00
Campbell Barton
2025e4cbb9 Code cleanup: quiet warnings & style 2014-04-13 11:26:31 +10:00
Campbell Barton
c04e73f386 Math Lib: use less strict epsilon with BLI_ASSERT_UNIT_QUAT 
was causing issues with pointcache
 2014-04-11 18:20:30 +10:00
Campbell Barton
a15b3c4d11 Code cleanup: use bool 2014-04-11 11:33:29 +10:00
Campbell Barton
f700a13eb5 Revert "Mempool: simplify memory chunk list building" 
This reverts commit c82371fc06.

Caused regression in iterator
 2014-04-10 11:35:17 +10:00
Campbell Barton
82628a6b0e Code cleanup: use struct type for mempool & style edits 2014-04-10 06:49:25 +10:00
Campbell Barton
5580afb5df GHash/Edgehash: make simple iterator checking functions inline. 
also remove NULL check, only a few areas made use of this.
 2014-04-08 15:50:38 +10:00
Campbell Barton
c82371fc06 Mempool: simplify memory chunk list building 2014-04-08 13:41:57 +10:00
Campbell Barton
412826a504 Mempool: delay allocating an initial chunk, its not always used 2014-04-08 12:58:56 +10:00
Campbell Barton
5c5d643938 Mempool: use define for used freeword and correct define 2014-04-07 12:52:23 +10:00
Campbell Barton
13d90ab8f7 Mempool: minor optimization to building free pointer list 2014-04-07 12:38:32 +10:00
Campbell Barton
aee82b4b2c Mempool: fix own error in recent commit 
chunks must be added in order for iteration.
 2014-04-07 12:38:32 +10:00
Campbell Barton
2a82b8ade5 Math Lib: add power of 2 min/max for unsigned ints 2014-04-05 21:54:20 +11:00
Campbell Barton
19b7bb5975 Optimize mempool: round chunk size to powers of 2, account for slop-space 
This helps to reduce memory fragmentation.
 2014-04-05 12:14:05 +11:00
Campbell Barton
af59ee340f Mempool: remove BLI_MEMPOOL_SYSMALLOC, MEM_* allocs are more efficient now 2014-04-05 12:14:05 +11:00
Campbell Barton
70353f833a Optimize mempool: replace double linked list with single for memory chunks 2014-04-05 09:49:13 +11:00
Campbell Barton
7cb90a611f Optimization for mempool initial chunk allocation 
Almost all pools allocated 2 chunks on initialization,
every element needed to be added to the free-list which
would never be used for small pools.

Now allocate only one, gives minor speedup for some bmesh operations.
 2014-04-04 21:25:08 +11:00
Campbell Barton
5770d691bb Optimize BLI_convexhull_aabb_fit_hull_2d, avoid atan2, sin, cos 
add utility functions for using a 2d unit vector as a rotation matrix
mul_v2_v2_cw & mul_v2_v2_ccw
 2014-04-03 21:47:03 +11:00
Bastien Montagne
35747ee3d5 Rename "BLI_cpu.h" to "BLI_system.h" and add to it BLI_getpid() helper. 
There is not much sense to have a whole BLI file just to check SSE2 on CPUs...
So idea is to rename it to more generic "BLI_system", and add to it more system-related
utils, like e.g. an include helper for getpid(), which allows to hide unix/windows
internals from rest of the code...

Reviewers: campbellbarton

Reviewed By: campbellbarton

Differential Revision: https://developer.blender.org/D439
 2014-04-02 11:59:43 +02:00
Campbell Barton
1d9e0c373d Fix for BLI_rng_shuffle_array noted by mont29 2014-04-02 10:09:48 +11:00
Campbell Barton
da4b90a331 Code cleanup: use uint64_t for BLI_rand 2014-04-02 09:58:27 +11:00
Jens Verwiebe
62dc18c717 Fix an unused function warning without openmp present, some typos 2014-04-01 11:23:28 +02:00
Campbell Barton
e95fd79258 Correction for error in own recent commit (makesrna c++ api, un-init var) 2014-04-01 16:53:15 +11:00
Campbell Barton
617557b08e Code cleanup: remove TRUE/FALSE & WITH_BOOL_COMPAT define 2014-04-01 15:22:28 +11:00
Campbell Barton
097a3756c0 Code cleanup: use bool 2014-03-31 23:39:08 +11:00
Jens Verwiebe
277fb1a31f Sculpt/dyntopo: Make the omp threads configurable to overcome performance issues 
- autodetect optimal default, which typically avoids HT threads
- can store setting in .blend per scene
- this does not touch general omp max threads, due i found other areas where the calculations are fitting for huge corecount
- Intel notes, some of the older generation processors with HyperThreading would not provide significant performance boost for FPU intensive applications. On those systems you might want to set OMP_NUM_THREADS = total number of cores (not total number of hardware theads).
 2014-03-31 13:51:49 +02:00
Campbell Barton
0055162b75 Tweak slerp assert to be less picky 2014-03-31 14:55:12 +11:00
Campbell Barton
55f83e36cc Py API: Vector.slerp(). also added interp_v3_v3v3_slerp(_safe) functions 2014-03-31 13:28:37 +11:00
Campbell Barton
6aa75d3b2c Fix for error in normalize_vn_vn(), add len_squared_vn 2014-03-31 11:19:32 +11:00
Thomas Dinges
b99977ac60 And another compile fix for M_PI and msvc2008... 2014-03-30 15:36:09 +02:00
Campbell Barton
0319db1063 Correct own recent changes broke release builds 2014-03-30 20:35:59 +11:00
Campbell Barton
8d1b289b78 Code cleanup: warnings (clang) 2014-03-30 19:37:43 +11:00
Campbell Barton
c16bd951cd Enable GCC pedantic warnings with strict flags, 
also modify MIN/MAX macros to prevent shadowing.
 2014-03-30 15:04:20 +11:00
Campbell Barton
5819421e1b correct recent commit 2014-03-30 15:04:20 +11:00
Campbell Barton
bbfeb120fc Code cleanup: use strict flags for BLI_rand 2014-03-30 15:04:20 +11:00
Campbell Barton
faf529d036 BLI_rand: Add BLI_rng_get_float_unit_v3, was static rayshade func 2014-03-30 15:04:20 +11:00
Campbell Barton
a6e8137983 Convenience macro for print_ funcs, saves passing id each time 2014-03-30 15:04:20 +11:00
Campbell Barton
480c5019bb Code cleanup: reflect_v3_v3v3 made redundant copies 2014-03-30 12:28:09 +11:00
Campbell Barton
3bd15fcf15 Correction for poly_to_tri_count (used over-complicated method) 2014-03-30 12:05:42 +11:00
Campbell Barton
1f58bfb8be Code cleanup: de-duplicate cotangent weight function & add arg sizes 2014-03-30 11:08:33 +11:00
Campbell Barton
23ef10c705 Code cleanup: kdopbvh, move minmax init into function 2014-03-30 10:33:01 +11:00
Campbell Barton
2100fb4094 Code cleanup: avoid setting du/dv twice with ray intersection 2014-03-29 22:38:01 +11:00