Commit Graph

46 Commits

Author SHA1 Message Date
DRC
1db1ce45da TJBench: Improve consistency of results
Given that libjpeg-turbo can often process hundreds of megapixels/second
on modern hardware, the default of one warmup iteration was essentially
meaningless.  Furthermore, the -warmup option was a bit clunky, since
it required some foreknowledge of how fast the benchmarks were going to
execute.

This commit introduces a 1-second warmup interval for each benchmark by
default, and the -warmup option has been retasked to control the length
of that interval.
2017-06-27 14:28:43 -05:00
DRC
6a2b067428 TJBench: Code formatting tweaks
Spaces-->tab + remove stray control character that was introduced in
95e4cb2060
2017-06-26 21:15:15 -05:00
DRC
11eec4a398 TJBench: Fix errors when decomp. files w/ ICC data
Embedded ICC profiles can cause the size of a JPEG file to exceed the
size returned by tjBufSize() (which is really meant to be used for
compression anyhow, not for decompression), and this was causing a
segfault (C) or an ArrayIndexOutOfBoundsException (Java) when
decompressing such files with TJBench.  This commit modifies the
benchmark such that, when tiled decompression is disabled, it re-uses
the source buffer as the primary JPEG buffer.
2017-06-26 21:12:40 -05:00
DRC
db04435165 Silence pedantic GCC6 code formatting warnings
Apparently it's "misleading" to put two self-contained if statements
on a single line.  Who knew?
2016-07-14 13:36:47 -05:00
DRC
3128061382 TJBench: Fix segfault on Android
For whatever reason, the "write" global variable in tjbench.c was
overriding the linkage with the write() system function.  This may have
affected other platforms as well but was not known to.
2016-02-01 11:03:39 -06:00
DRC
9c8cee8c29 tjbench: Further tweaks to -nowrite feature
* Do not compute compression error if -nowrite is specified
* Adjust formatting of -nowrite usage description
2016-01-13 22:32:19 -06:00
DRC
eb59b6e72d Add -nowrite arg to TJBench to improve consistency
Prevents any images from being written to disk, thus making the
performance of the benchmark as CPU-bound as possible.
2016-01-12 00:36:32 -06:00
DRC
2ab5ea6c48 Add the ability to benchmark YCCK JPEG compression/decompression. This is particularly useful since that is the only way to test the performance of the "plain" upsampling routines, which are accelerated on some platforms.
git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/trunk@1505 632fc199-4ca6-4c93-a231-07263d6284db
2015-01-14 19:39:01 +00:00
DRC
f5644c3498 Ensure that tjFree() is used for any JPEG buffers that might have been dynamically allocated by the compress/transform functions. To keep things simple, we use tjAlloc() for the statically-allocated buffer as well, so that tjFree() can always be used to free the buffer, regardless of whether it was allocated by tjbench or by the TurboJPEG library. This fixes crashes that occurred on Windows when running tjunittest or tjbench with the -alloc flag.
git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/trunk@1383 632fc199-4ca6-4c93-a231-07263d6284db
2014-08-22 19:27:28 +00:00
DRC
eb66974f3a Ensure that tjFree() is used for any JPEG buffers that might have been dynamically allocated by the compress/transform functions. To keep things simple, we use tjAlloc() for the statically-allocated buffer as well, so that tjFree() can always be used to free the buffer, regardless of whether it was allocated by tjbench or by the TurboJPEG library. This fixes crashes that occurred on Windows when running tjunittest or tjbench with the -alloc flag.
git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/branches/1.3.x@1384 632fc199-4ca6-4c93-a231-07263d6284db
2014-08-22 19:59:51 +00:00
DRC
684ace1924 Extend tjbenchtest so that it tests the dynamic JPEG buffer allocation feature in TurboJPEG. Disable the tiling feature in TJBench whenever dynamic buffer allocation is enabled (because the tiling feature requires a separate buffer for each tile, using it successfully with dynamic buffer allocation would require a separate TurboJPEG compressor instance for each tile, and it's not worth going to that trouble right now.)
git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/trunk@1374 632fc199-4ca6-4c93-a231-07263d6284db
2014-08-22 03:04:06 +00:00
DRC
43513893f0 Extend tjbenchtest so that it tests the dynamic JPEG buffer allocation feature in TurboJPEG. Disable the tiling feature in TJBench whenever dynamic buffer allocation is enabled (because the tiling feature requires a separate buffer for each tile, using it successfully with dynamic buffer allocation would require a separate TurboJPEG compressor instance for each tile, and it's not worth going to that trouble right now.)
git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/branches/1.3.x@1372 632fc199-4ca6-4c93-a231-07263d6284db
2014-08-22 02:57:34 +00:00
DRC
1a45b81fa2 Remove trailing spaces (+ one additional tab in TJUnitTest.java that was missed in the previous commit)
git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/trunk@1279 632fc199-4ca6-4c93-a231-07263d6284db
2014-05-09 18:06:58 +00:00
DRC
0c989d9e90 Fix warnings about unused variables when building with GCC 4.8.x
git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/branches/1.3.x@1242 632fc199-4ca6-4c93-a231-07263d6284db
2014-04-17 00:47:18 +00:00
DRC
4186162665 Fix warnings about unused variables when building with GCC 4.8.x
git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/trunk@1241 632fc199-4ca6-4c93-a231-07263d6284db
2014-04-16 23:38:37 +00:00
DRC
dd5fcdd007 formatting tweaks
git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/trunk@1183 632fc199-4ca6-4c93-a231-07263d6284db
2014-03-17 10:40:10 +00:00
DRC
78b227e251 Oops. This code should not have been removed. It is necessary for generating the error images.
git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/trunk@1165 632fc199-4ca6-4c93-a231-07263d6284db
2014-03-14 04:15:43 +00:00
DRC
ce6891f048 If we have transformed a 4:2:2 or 4:4:0 image in any way that transposes the horizontal and vertical dimensions, then we need to inform the decompression function that the subsampling type has changed. Otherwise, tjDecodeYUV() will not produce correct results.
git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/trunk@1162 632fc199-4ca6-4c93-a231-07263d6284db
2014-03-13 20:33:43 +00:00
DRC
adf5551ba9 Execute warmup runs with tjTransform() as well
git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/trunk@1159 632fc199-4ca6-4c93-a231-07263d6284db
2014-03-12 07:17:23 +00:00
DRC
e1efc8a0ef Fix a couple of output formatting issues
git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/trunk@1158 632fc199-4ca6-4c93-a231-07263d6284db
2014-03-12 07:16:17 +00:00
DRC
95e4cb2060 Since we now have a complete set of YUV functions, TJBench can be greatly simplified. It now tests YUV encoding/decoding as an intermediate step of JPEG compression/decompression, which eliminates the need for a separate YUV mode. Several other things have been streamlined in the process.
git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/trunk@1157 632fc199-4ca6-4c93-a231-07263d6284db
2014-03-12 06:51:39 +00:00
DRC
0e9bf6442b Fix a segfault that would occur in decompress-only mode if -alloc was specified without -tile.
git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/trunk@1156 632fc199-4ca6-4c93-a231-07263d6284db
2014-03-12 06:17:12 +00:00
DRC
dd59233dd4 Fix a segfault that would occur in decompress-only mode if -alloc was specified without -tile.
git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/branches/1.3.x@1155 632fc199-4ca6-4c93-a231-07263d6284db
2014-03-12 06:15:51 +00:00
DRC
a15f19f2d2 Wordsmithing & formatting tweaks
git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/branches/1.3.x@1154 632fc199-4ca6-4c93-a231-07263d6284db
2014-03-11 09:46:50 +00:00
DRC
d4b453c26e Back-port the -subsamp option from 1.4 rather than use the hackish approach of replacing 4:2:2 with 4:4:0. This has the added advantage of allowing the user to test only a specific level of subsampling.
git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/branches/1.3.x@1153 632fc199-4ca6-4c93-a231-07263d6284db
2014-03-11 08:28:47 +00:00
DRC
c6501f7f5f Add the ability to test scaling when decompressing to YUV; compression from YUV to JPEG; and YUV padding. Replace clunky -411 and -440 parameters with a -subsamp parameter that allows any of the subsampling options to be tested in isolation.
git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/trunk@1074 632fc199-4ca6-4c93-a231-07263d6284db
2013-11-04 23:07:54 +00:00
DRC
07e982d9b4 Deprecate and undocument the FORCE{MMX|SSE|SSE2|SSE3} flags. These were originally introduced in TurboJPEG/IPP as a way to override the automatic CPU selection in the underlying IPP codec, which was closed source. They are not meaningful anymore, since libjpeg-turbo provides environment variables to accomplish the same thing and since it no longer necessarily uses x86 SIMD code behind the scenes.
git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/trunk@1072 632fc199-4ca6-4c93-a231-07263d6284db
2013-10-31 07:11:39 +00:00
DRC
c45653e471 Fix a couple of issues with return value checking. JFREAD(), which wraps fread(), will never return -1. fread() will instead return 0 or a short object count if an error occurs, and ferror() will return 1 in that case. The second issue was that we were assigning the return value of ftell() to an unsigned long prior to checking the value, so the value would never be < 0 if an error occurred. It would instead be (unsigned long)-1.
git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/branches/1.3.x@1061 632fc199-4ca6-4c93-a231-07263d6284db
2013-10-12 21:52:48 +00:00
DRC
490c7f955a Fix a couple of issues with return value checking. JFREAD(), which wraps fread(), will never return -1. fread() will instead return 0 or a short object count if an error occurs, and ferror() will return 1 in that case. The second issue was that we were assigning the return value of ftell() to an unsigned long prior to checking the value, so the value would never be < 0 if an error occurred. It would instead be (unsigned long)-1.
git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/trunk@1060 632fc199-4ca6-4c93-a231-07263d6284db
2013-10-12 21:51:16 +00:00
DRC
38cb1ec2a7 Add CMYK support to the TurboJPEG Java API & clean up a few things in the C API
git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/trunk@1020 632fc199-4ca6-4c93-a231-07263d6284db
2013-08-23 04:45:43 +00:00
DRC
cd7c3e6672 Add CMYK support to the TurboJPEG C API
git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/trunk@1019 632fc199-4ca6-4c93-a231-07263d6284db
2013-08-23 02:49:25 +00:00
DRC
1f3635c496 Add 4:1:1 subsampling support in the TurboJPEG C API
git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/trunk@1014 632fc199-4ca6-4c93-a231-07263d6284db
2013-08-18 10:19:00 +00:00
DRC
d081fd728c Add note regarding the fact that 4:4:0 lacks full SIMD support; Add an option for benchmarking 4:4:0 subsampling in tjbench.
git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/trunk@1009 632fc199-4ca6-4c93-a231-07263d6284db
2013-08-18 09:15:37 +00:00
DRC
07796da4e5 Add note regarding the fact that 4:4:0 lacks full SIMD support; Add an option for benchmarking 4:4:0 subsampling in tjbench.
git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/branches/1.3.x@1008 632fc199-4ca6-4c93-a231-07263d6284db
2013-08-18 09:13:00 +00:00
DRC
73d74c132b Add flags to the TurboJPEG API that allow the caller to force the use of either the fast or the accurate DCT/IDCT algorithms in the underlying codec.
git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/trunk@851 632fc199-4ca6-4c93-a231-07263d6284db
2012-06-29 23:46:38 +00:00
DRC
ffe4fc182f Fix the behavior of the alpha-enabled colorspace constants whenever libjpeg-turbo is built without SIMD support and merged upsampling is used.
git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/trunk@810 632fc199-4ca6-4c93-a231-07263d6284db
2012-03-16 14:30:46 +00:00
DRC
cac105133e Fix the behavior of the alpha-enabled colorspace constants whenever libjpeg-turbo is built without SIMD support and merged upsampling is used.
git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/branches/1.2.x@811 632fc199-4ca6-4c93-a231-07263d6284db
2012-03-16 14:37:36 +00:00
DRC
ab2df6ea1b Expose new scaling factors in TurboJPEG API
git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/trunk@757 632fc199-4ca6-4c93-a231-07263d6284db
2012-01-28 06:49:56 +00:00
DRC
e006f5c427 Oops. Height of output buffer should equal scaled height, not JPEG height.
git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/branches/1.2.x@755 632fc199-4ca6-4c93-a231-07263d6284db
2012-01-28 06:31:44 +00:00
DRC
20c7dbc64f Fix compiler warning
git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/trunk@713 632fc199-4ca6-4c93-a231-07263d6284db
2011-10-18 22:06:22 +00:00
DRC
f546711076 Implement custom filter callback in Java
git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/trunk@707 632fc199-4ca6-4c93-a231-07263d6284db
2011-09-20 05:02:19 +00:00
DRC
57a3736243 git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/trunk@704 632fc199-4ca6-4c93-a231-07263d6284db 2011-09-17 00:41:14 +00:00
DRC
7bf04d399e Implement a custom DCT filter callback for lossless transforms
git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/trunk@703 632fc199-4ca6-4c93-a231-07263d6284db
2011-09-17 00:18:31 +00:00
DRC
9b49f0e4c7 Re-work TJBUFSIZE() to take into account the level of chrominance subsampling
git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/trunk@668 632fc199-4ca6-4c93-a231-07263d6284db
2011-07-12 03:17:23 +00:00
DRC
94f0e03f90 Oops. Add the already-advertised -bmp option
git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/trunk@647 632fc199-4ca6-4c93-a231-07263d6284db
2011-05-25 04:35:53 +00:00
DRC
b8b359ae97 "jpegut"="tjunittest" & "jpgtest"="tjbench"
git-svn-id: svn+ssh://svn.code.sf.net/p/libjpeg-turbo/code/trunk@644 632fc199-4ca6-4c93-a231-07263d6284db
2011-05-25 03:54:56 +00:00