Age | Commit message (Collapse) | Author | Files | Lines | |
---|---|---|---|---|---|
2015-05-02 | Move iter_storage allocation to thread function | Lars-Dominik Braun | 1 | -26/+24 | |
2015-05-02 | Prefetch bucket accumulator | Lars-Dominik Braun | 1 | -11/+13 | |
The bucket add is a hotspot. | |||||
2015-05-02 | Use rotation matrix | Lars-Dominik Braun | 1 | -15/+10 | |
Reuse affine transformation code from variations. | |||||
2015-05-02 | Factor out palette interpolation | Lars-Dominik Braun | 1 | -32/+38 | |
2015-05-02 | Vectorize color clipping | Lars-Dominik Braun | 1 | -146/+61 | |
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). | |||||
2015-05-02 | Use floating point color values in [0,1] | Lars-Dominik Braun | 1 | -23/+21 | |
Internal color representation is [0,1] now for every channel. | |||||
2015-05-02 | Remove all uses of getenv | Lars-Dominik Braun | 1 | -16/+1 | |
Not much functionality lost here. Move on. | |||||
2015-05-02 | Fix memory leaks | Lars-Dominik Braun | 1 | -2/+3 | |
Introduced by 15f89a8390af9338227c2ee3f603b32a1a206a8f | |||||
2015-05-02 | Correctly align malloc’d vector arrays | Lars-Dominik Braun | 1 | -3/+14 | |
Fixes segfault with -march=native | |||||
2015-05-02 | Vectorize accumulation | Lars-Dominik Braun | 1 | -84/+47 | |
Also vectorizes some color functions and switches to double accumulation buffer. Does not seem to be slower. | |||||
2015-05-02 | Drop density estimator | Lars-Dominik Braun | 1 | -358/+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. | |||||
2015-05-02 | Remove unused variables, fix compiler warnings | Lars-Dominik Braun | 1 | -2/+0 | |
2015-05-02 | Unify prng use | Lars-Dominik Braun | 1 | -6/+6 | |
Finally drop all isaac references, do not use system rng any more. Drop rng state from flam3_frame – it was not used anyway. | |||||
2015-05-02 | Drop Windows/Mac support, require pthreads | Lars-Dominik Braun | 1 | -60/+4 | |
2015-05-02 | Drop support for atomic ops and bucket formats | Lars-Dominik Braun | 1 | -26/+0 | |
Atomic ops are slower than locks, probably going to switch buckets to uint64_t later. | |||||
2015-05-02 | Vectorize flam3_iterate | Lars-Dominik Braun | 1 | -10/+12 | |
2015-05-02 | Replace isaac with xorshift | Lars-Dominik Braun | 1 | -4/+1 | |
No difference in performance. rdrand instruction alone is too slow. | |||||
2015-05-02 | Fix crash | Lars-Dominik Braun | 1 | -3/+2 | |
For some reason filter is overwritten by pthread_join, removing its second argument fixes the problem. It is unused anyway. | |||||
2015-05-02 | Move out of subdir | Lars-Dominik Braun | 1 | -0/+1287 | |