* Add GNUInstallDirs for standard installation directories
Distributions are given standard variables for already existing hooks.
Multiarch libdirs is taken care of automagically.
Raises minimum cmake version by a little.
* Handle CMAKE_INSTALL_xxx being absolute paths for .pc file generation
In some cases the CMAKE_INSTAL_{BIN,MAN,DOC,LIB,INCLUDE}DIR variables
may turn out to be absolute paths in which case prepending ${prefix} in
the pkg-config .pc files will result in incorrect values.
For .pc file generation, figure out if these variables are absolute and
omit the prefix in the configured file when so.
See: ab25e4b7ed
There are a few limitations:
- mixed mode (HT and regular code blocks) is not supported.
- ROI in HT blocks is not supported.
- Placeholder passes are not supported.
- MultiHT sets are not support, only a singleHT set.
- there are known issues with some compliance testing files related to
the parsing of packet header.
Update the bench_dwt utility to have a -decode/-encode switch
Measured performance gains for DWT encoder on a
Intel(R) Core(TM) i7-6700HQ CPU @ 2.60GHz (4 cores, hyper threaded)
Encoding time:
$ ./bin/bench_dwt -encode -num_threads 1
time for dwt_encode: total = 8.348 s, wallclock = 8.352 s
$ ./bin/bench_dwt -encode -num_threads 2
time for dwt_encode: total = 9.776 s, wallclock = 4.904 s
$ ./bin/bench_dwt -encode -num_threads 4
time for dwt_encode: total = 13.188 s, wallclock = 3.310 s
$ ./bin/bench_dwt -encode -num_threads 8
time for dwt_encode: total = 30.024 s, wallclock = 4.064 s
Scaling is probably limited by memory access patterns causing
memory access to be the bottleneck.
The slightly worse results with threads==8 than with thread==4
is due to hyperthreading being not appropriate here.
Instead of being the full tile size.
* Use a sparse array mechanism to store code-blocks and intermediate stages of
IDWT.
* IDWT, DC level shift and MCT stages are done just on that smaller array.
* Improve copy of tile component array to final image, by saving an intermediate
buffer.
* For full-tile decoding at reduced resolution, only allocate the tile buffer to
the reduced size, instead of the full-resolution size.
This saves comparing the current pointer with the end of buffer pointer.
This results at least in tiny speed improvement for raw decoding, and
smaller code size for MQC as well.
This kills the remains of the raw.h/.c files that were only used for
decoding. Encoding using the mqc structure already.
r2077 did expose too much of the detail implementation of OpenJPEG. Rework dependencies in between JPIP file level details and JPIP client/server arch.
Move JPIP file level functionalities back into OpenJP2 as was the case in openjpeg 1.5.x. Use new cio func.
Thomas Bracht Laumann Jespersen
Aous Naman
Anonymous Maarten
Even Rouault
Andrew Murray
Jeroen
Even Rouault
Mathieu Malaterre
Antonin Descampe
Kal Conley
Julien Malik
mayeut
Rex Dieter
Mickael Savinaud