Thanks to our macros that abstract SSE use, the functions can use
AVX2 when available (at compile time)
This brings an extra 23% speed improvement on bench_dwt in 64bit builds
with AVX2 compared to SSE2.
* Use single-pass lifting inverse wavelet transform.
* For vertical pass, use SSE2 when available so as to process 8 columns
in parallel. This is the most beneficial improvement, since the
vertical pass involves a lot of cache trashing.
With the bench_dwt utility with default arguments (16383x16383 image),
time goes from 4.064 s to 1.212 s.
Commit 29313eb5 introduced those flags to avoid issues with
-fsanitize=unsigned-integer-overflow
However it is better just to rewrite the loop to avoid such condition
to occur.
There are situations where, given a tile size, at a resolution level,
there are sub-bands with x0==x1 or y0==y1, that consequently don't have any
valid codeblocks, but the other sub-bands may be non-empty.
Given that we recycle the memory from one tile to another one, those
ghost codeblocks might be non-0 and thus candidate for packet inclusion.
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.
Ported from Carl Hetherington work (actually through Matthieu Darbois's port
on top of OpenJPEG 2.1.0)
Can reduce total encoding time by 10-15%
WARNING: VSC mode is not implemented, and so is a temporary regression
that must be fixed.
Decoding some valid .jp2 files like Sentinel2 datasets leads to warnings like:
No incltree created.
tgt_create tree->numnodes == 0, no tree created.
No imsbtree created.
tgt_create tree->numnodes == 0, no tree created.
Besides that, the image is correctly decoded. So there is no reason to emit
those warnings.
Visual Studio 2015 does not pass regression tests with `__restrict` so kept disabled for MSVC.
Need to check proper usage of OPJ_RESTRICT (if correct then there’s
probably a bug in vc14)
Closes#661
The definition of bit-fields with type OPJ_UINT32 caused complilation errors
on IBM iSeries, because OPJ_UINT32 is defined as uint32_t, and
uint32_t is defined as unsigned long in <stdint.h>. The definition of
bit-fields with an integer type of a specific size doesn't make sense
anyway.
By default, only the main thread is used. If opj_codec_set_threads() is not used,
but the OPJ_NUM_THREADS environment variable is set, its value will be
used to initialize the number of threads. The value can be either an integer
number, or "ALL_CPUS". If OPJ_NUM_THREADS is set and this function is called,
this function will override the behaviour of the environment variable.
Addition flag array such that colflags[1+0] is for state of col=0,row=0..3,
colflags[1+1] for col=1, row=0..3, colflags[1+flags_stride] for col=0,row=4..7, ...
This array avoids too much cache trashing when processing by 4 vertical samples
as done in the various decoding steps.
Add a opj_t1_dec_clnpass_step_only_if_flag_not_sig_visit() method that
does the job of opj_t1_dec_clnpass_step_only() assuming the conditions
are met. And use it in opj_t1_dec_clnpass(). The compiler generates
more efficient code.
Fix issue 135
The fix is legal regarding the standard but I did not manage to find out if it covers a bug in opj_t2_read_packet_data or if the file is corrupted
* Fix unsigned int overflow reported by UBSan
Please add -DOPJ_UBSAN_BUILD to CFLAGS when building with
-fsanitize=undefined,unsigned-integer-overflow
It seems clang/gcc do not allow to disable checking for block of code
other than function or file.
dwt_interleave_h.gsr105.jp2 now has the same output as kakadu
issue399 is corrupted. Only the corrupted part changes.
Update known failures for x86 MD5
NR-DEC-kodak_2layers_lrcp.j2c-31-decode-md5
NR-DEC-kodak_2layers_lrcp.j2c-32-decode-md5
NR-DEC-issue135.j2k-68-decode-md5
Follow-up of #757
This shall have no performance impact on 2’s complement machine where
the compiler replaces the multiplication by power of two (constant) by
a left shift.
Verified at least on MacOS Xcode 7.3, same assembly generated after fix.
This shall have no performance impact on 2’s complement machine where
the compiler replaces the multiplication by power of two (constant) by
a left shift.
Verified at least on MacOS Xcode 7.3, same assembly generated after fix.
When trying the GDAL OpenJPEG driver against openjpeg current master HEAD,
I get failures when trying to create .jp2 files. The driver uses
opj_write_tile() and in some tests numresolutions = 1.
In openjp2/dwt.c:410, l_data_size = opj_dwt_max_resolution( tilec->resolutions,tilec->numresolutions) * (OPJ_UINT32)sizeof(OPJ_INT32);
is called and returns l_data_size = 0. Now in git opj_malloc() has a special case
for 0 to return a NULL pointer whereas previously it relied on system malloc(),
which in my case didn't return NULL.
So only test the pointer value if l_data_size != 0. This makes the GDAL
autotest suite to pass again.
By default, OpenJPEG uses the function memalign to allocate aligned
memory on Linux systems. That function needs malloc.h which was
missing. This results in a compiler warning:
openjpeg/src/lib/openjp2/opj_malloc.c:63:3: warning:
implicit declaration of function ‘memalign’
[-Wimplicit-function-declaration]
On hosts where sizeof(int) < sizeof(void *) the return value of memalign
will be truncated which results in an invalid pointer.
That caused "make test" to produce lots of segmentation faults when
running on a 64 bit Linux host.
Signed-off-by: Stefan Weil <sw@weilnetz.de>
posix_memalign is only declared conditionally in stdlib.h,
so add one of the possible definitions to get the declaration.
Signed-off-by: Stefan Weil <sw@weilnetz.de>
HP compiler warns:
cc: "dwt.c", line 798: warning 562: Redeclaration of "opj_v4dwt_decode"
with a different storage class specifier: "opj_v4dwt_decode" will have
internal linkage.
cc: "t2.c", line 1341: warning 562: Redeclaration of "opj_t2_init_seg"
with a different storage class specifier: "opj_t2_init_seg" will have
internal linkage.
Even Rouault
Jeroen Ooms
Mathieu Malaterre
Antonin Descampe
Matthieu Darbois
Stephan Mühlstrasser
trylab
Stephan Mühlstrasser
Kal Conley
Stefan Weil
Julien Malik
julienmalik
Aaron Boxer