summaryrefslogtreecommitdiff
AgeCommit message (Collapse)AuthorFilesLines
2015-05-02Use rotation matrixLars-Dominik Braun4-26/+60
Reuse affine transformation code from variations.
2015-05-02Factor out genome printfLars-Dominik Braun1-10/+9
Fixes -random’s output.
2015-05-02Move width/height options from render to randomLars-Dominik Braun1-29/+63
flam3’s design is a little wonky here. An explicit bounding box in the flame’s coordinate system would help.
2015-05-02random: Adjust bounding boxLars-Dominik Braun1-0/+34
2015-05-02render: Add width/height command line optionsLars-Dominik Braun1-3/+33
2015-05-02Factor out palette interpolationLars-Dominik Braun5-41/+47
2015-05-02Vectorize color clippingLars-Dominik Braun11-166/+143
Replaces redundant code with one function. Oddly this fixes rendering with earlyclip – not sure why. Drop transparency and channel settings (always transparent, always four channels).
2015-05-02render: Add oversample argumentLars-Dominik Braun1-1/+13
2015-05-02Use floating point color values in [0,1]Lars-Dominik Braun4-40/+37
Internal color representation is [0,1] now for every channel.
2015-05-02Remove all uses of getenvLars-Dominik Braun6-213/+14
Not much functionality lost here. Move on.
2015-05-02Drop unused png read functionLars-Dominik Braun1-108/+0
2015-05-02Improve command line interfaceLars-Dominik Braun22-3289/+443
Initial arguments-based cli, replacing the three tools -animate, -genome and -render with just one. Still need to get rid of most calls to getenv.
2015-05-02Fix memory leaksLars-Dominik Braun1-2/+3
Introduced by 15f89a8390af9338227c2ee3f603b32a1a206a8f
2015-05-02Correctly align malloc’d vector arraysLars-Dominik Braun2-4/+17
Fixes segfault with -march=native
2015-05-02Support clang vector attributeLars-Dominik Braun1-0/+9
2015-05-02Vectorize accumulationLars-Dominik Braun8-178/+97
Also vectorizes some color functions and switches to double accumulation buffer. Does not seem to be slower.
2015-05-02Drop density estimatorLars-Dominik Braun9-545/+22
First of all, it does not look too pretty. But more importantly now we can reduce the bucket size to four (instead of five) and vectorize that stuff.
2015-05-02Remove unused variables, fix compiler warningsLars-Dominik Braun7-20/+11
2015-05-02Switch rng seed to /dev/urandomLars-Dominik Braun2-13/+12
2015-05-02Unify prng useLars-Dominik Braun19-433/+373
Finally drop all isaac references, do not use system rng any more. Drop rng state from flam3_frame – it was not used anyway.
2015-05-02Fix fuse iterationsLars-Dominik Braun1-1/+1
Bug introduced by commit 37b8237136afee2d47d3899b87a3bafdc40acce3
2015-05-02Drop striped renderingLars-Dominik Braun4-174/+35
The amount of memory available is not a problem anymore.
2015-05-02Drop support for jpeg/netpbm outputLars-Dominik Braun6-206/+6
2015-05-02Drop unused, exported functionsLars-Dominik Braun6-709/+19
2015-05-02Drop Windows/Mac support, require pthreadsLars-Dominik Braun12-279/+6
2015-05-02Drop support for atomic ops and bucket formatsLars-Dominik Braun3-257/+2
Atomic ops are slower than locks, probably going to switch buckets to uint64_t later.
2015-05-02Fix curve variationLars-Dominik Braun1-2/+2
2015-05-02Vectorize flam3_iterateLars-Dominik Braun4-52/+51
2015-05-02Vectorize c and post matrices as well apply_xformLars-Dominik Braun8-74/+66
2015-05-02Bugfixes for variations vectorizationLars-Dominik Braun1-5/+5
2015-05-02Explicit input coordinatesLars-Dominik Braun2-392/+389
Slower, obviously.
2015-05-02Vectorize variantsLars-Dominik Braun2-657/+562
Slightly worse with amdlibm, slightly better with gcc’s builtins (no optimization) and worse with optimization
2015-05-02Move to wafLars-Dominik Braun14-3001/+43
2015-05-02Declare variation functions staticLars-Dominik Braun2-220/+107
2015-05-02Fix compiler warningLars-Dominik Braun1-1/+1
2015-05-02Remove autoconf stuff from gitLars-Dominik Braun21-50796/+13
2015-05-02Replace isaac with xorshiftLars-Dominik Braun14-271/+79
No difference in performance. rdrand instruction alone is too slow.
2015-05-02Fix crashLars-Dominik Braun1-3/+2
For some reason filter is overwritten by pthread_join, removing its second argument fixes the problem. It is unused anyway.
2015-05-02Move out of subdirLars-Dominik Braun65-0/+0
2015-02-15Modified for libpng 1.5 API support.Erik Reckase1-14/+14
git-svn-id: https://flam3.googlecode.com/svn/trunk@164 77852712-ef1d-11de-8684-7d64432d61a3
2015-02-15s array in interpolate_n was one element too smallErik Reckase1-1/+1
git-svn-id: https://flam3.googlecode.com/svn/trunk@163 77852712-ef1d-11de-8684-7d64432d61a3
2015-02-15moved #defines from filters.h to flam3.h so that they can be used by ↵Erik Reckase2-21/+23
external programs git-svn-id: https://flam3.googlecode.com/svn/trunk@162 77852712-ef1d-11de-8684-7d64432d61a3
2015-02-15was incorrectly adding symmetry xforms when a final xform was present, as ↵Erik Reckase1-1/+12
well as a bug in handling final xforms in genetic crosses. thanks david bitseff git-svn-id: https://flam3.googlecode.com/svn/trunk@161 77852712-ef1d-11de-8684-7d64432d61a3
2015-02-15When using motion elements with the final xform, the closing tag on the ↵Erik Reckase1-1/+5
final xform was incorrectly written as </xform>. git-svn-id: https://flam3.googlecode.com/svn/trunk@160 77852712-ef1d-11de-8684-7d64432d61a3
2015-02-15fixed flam3_version for 'exported' SVN_REVErik Reckase1-9/+37
git-svn-id: https://flam3.googlecode.com/svn/trunk@159 77852712-ef1d-11de-8684-7d64432d61a3
2015-02-15roundoff error was introducing some strangeness in smoothly interpolated ↵Erik Reckase1-0/+7
genome palettes - now ensure that all 1.0 alpha channels -> 1.0 alpha in output git-svn-id: https://flam3.googlecode.com/svn/trunk@156 77852712-ef1d-11de-8684-7d64432d61a3
2015-02-15handled missing unistd.h on windowsErik Reckase1-1/+3
git-svn-id: https://flam3.googlecode.com/svn/trunk@155 77852712-ef1d-11de-8684-7d64432d61a3
2015-02-15in some circumstances, temporary files are not created as they are supposed ↵Erik Reckase2-1/+29
to (windows tries to create it in a place protected by Admin.) We need to be able to try to create them in the location pointed to by the TEMP env var. Note that this is not threadsafe, but it's unlikely to cause a problem at the moment. git-svn-id: https://flam3.googlecode.com/svn/trunk@154 77852712-ef1d-11de-8684-7d64432d61a3
2015-02-15CMAP_SIZE can be used when creating the dmap - I had erroneously thought ↵Erik Reckase1-3/+3
that cmap_size was necessary (the apo-corrected number of elements in the palette) git-svn-id: https://flam3.googlecode.com/svn/trunk@153 77852712-ef1d-11de-8684-7d64432d61a3
2015-02-15apophysis only uses the bottom 255 elements of the palette (instead of all ↵Erik Reckase4-13/+20
256). In many cases this does not have a noticable effect on images, but in others, the color difference is very striking. the env var 'apo_palette' may be set to 1 in order to render the image in flam3 as it would appear in apophysis. git-svn-id: https://flam3.googlecode.com/svn/trunk@152 77852712-ef1d-11de-8684-7d64432d61a3