Commit Graph
399 Commits
Author SHA1 Message Date
Nigel Tao b61a827354 Optimize std/adler32 for SSE4.2
On a mid-range x86_64 laptop:

wuffs_adler32_10k/clang9                                  1.35GB/s ± 0%  10.78GB/s ± 1%  +700.65%  (p=0.008 n=5+5)
wuffs_adler32_100k/clang9                                 1.35GB/s ± 0%  12.88GB/s ± 0%  +856.09%  (p=0.008 n=5+5)

wuffs_adler32_10k/gcc10                                   1.85GB/s ± 0%  11.21GB/s ± 0%  +506.34%  (p=0.008 n=5+5)
wuffs_adler32_100k/gcc10                                  1.84GB/s ± 0%  11.43GB/s ± 0%  +521.51%  (p=0.008 n=5+5)

mimic_adler32_10k                                         1.76GB/s ± 0%   1.76GB/s ± 0%      ~     (p=0.841 n=5+5)
mimic_adler32_100k                                        1.72GB/s ± 0%   1.72GB/s ± 0%      ~     (p=0.222 n=5+5)

wuffs_zlib_decode_10k/clang9                               124MB/s ± 0%    139MB/s ± 0%   +12.29%  (p=0.016 n=4+5)
wuffs_zlib_decode_100k/clang9                              148MB/s ± 0%    155MB/s ± 0%    +4.21%  (p=0.008 n=5+5)

wuffs_zlib_decode_10k/gcc10                                133MB/s ± 1%    141MB/s ± 0%    +5.75%  (p=0.008 n=5+5)
wuffs_zlib_decode_100k/gcc10                               159MB/s ± 0%    172MB/s ± 0%    +7.77%  (p=0.016 n=5+4)

mimic_zlib_decode_10k                                      133MB/s ± 0%    133MB/s ± 0%    +0.38%  (p=0.008 n=5+5)
mimic_zlib_decode_100k                                     160MB/s ± 0%    160MB/s ± 0%      ~     (p=0.841 n=5+5)

wuffs_png_decode_image_19k_8bpp/clang9                    90.6MB/s ± 0%   94.7MB/s ± 0%    +4.58%  (p=0.016 n=5+4)
wuffs_png_decode_image_40k_24bpp/clang9                    129MB/s ± 0%    135MB/s ± 0%    +4.49%  (p=0.008 n=5+5)
wuffs_png_decode_image_77k_8bpp/clang9                     335MB/s ± 0%    351MB/s ± 0%    +4.62%  (p=0.008 n=5+5)
wuffs_png_decode_image_552k_32bpp_ignore_checksum/clang9   322MB/s ± 0%    320MB/s ± 1%    -0.71%  (p=0.016 n=4+5)
wuffs_png_decode_image_552k_32bpp_verify_checksum/clang9   256MB/s ± 0%    308MB/s ± 0%   +20.24%  (p=0.008 n=5+5)
wuffs_png_decode_image_4002k_24bpp/clang9                  131MB/s ± 0%    138MB/s ± 0%    +5.24%  (p=0.008 n=5+5)

wuffs_png_decode_image_19k_8bpp/gcc10                     95.1MB/s ± 0%  100.8MB/s ± 1%    +6.08%  (p=0.008 n=5+5)
wuffs_png_decode_image_40k_24bpp/gcc10                     131MB/s ± 0%    136MB/s ± 0%    +3.65%  (p=0.008 n=5+5)
wuffs_png_decode_image_77k_8bpp/gcc10                      350MB/s ± 1%    371MB/s ± 0%    +6.06%  (p=0.008 n=5+5)
wuffs_png_decode_image_552k_32bpp_ignore_checksum/gcc10    327MB/s ± 0%    333MB/s ± 0%    +1.66%  (p=0.008 n=5+5)
wuffs_png_decode_image_552k_32bpp_verify_checksum/gcc10    274MB/s ± 0%    318MB/s ± 0%   +15.93%  (p=0.008 n=5+5)
wuffs_png_decode_image_4002k_24bpp/gcc10                   132MB/s ± 0%    138MB/s ± 0%    +4.73%  (p=0.008 n=5+5)

mimic_png_decode_image_19k_8bpp                           58.1MB/s ± 0%   58.2MB/s ± 0%    +0.27%  (p=0.008 n=5+5)
mimic_png_decode_image_40k_24bpp                          73.0MB/s ± 1%   73.3MB/s ± 0%    +0.45%  (p=0.008 n=5+5)
mimic_png_decode_image_77k_8bpp                            177MB/s ± 0%    177MB/s ± 0%      ~     (p=0.421 n=5+5)
mimic_png_decode_image_552k_32bpp_ignore_checksum               skipped
mimic_png_decode_image_552k_32bpp_verify_checksum          146MB/s ± 0%    146MB/s ± 0%      ~     (p=0.310 n=5+5)
mimic_png_decode_image_4002k_24bpp                         104MB/s ± 0%    104MB/s ± 0%      ~     (p=0.056 n=5+5)
2021-01-26 22:31:54 +11:00
Nigel Tao cee57bc4ae Add some more base.x86_m128i methods 2021-01-26 10:46:36 +11:00
Nigel Tao 2f2bbc6ab0 Rename default github.com branch s/master/main/ 2021-01-25 11:30:15 +11:00
Nigel Tao 14f194cdfc Remove built-in range reset methods 2021-01-24 22:16:57 +11:00
Nigel Tao 77b9316aad Have "SSE42 capable" also include PCLMUL 2021-01-24 22:11:17 +11:00
Nigel Tao 73d95387f5 Change some more s/sse128/sse42/ 2021-01-24 22:04:53 +11:00
Nigel Tao 5941e7e5cc Rename sse128/sse128i to x86_sse42/x86_m128i 2021-01-24 21:58:00 +11:00
Nigel Tao 0c1f083522 Change _mm_loadu_si128 to _mm_lddqu_si128 2021-01-23 23:01:18 +11:00
Nigel Tao de9affaed8 Have _mm_cvtsi32_si128 cast its arg to int32_t 2021-01-23 22:59:52 +11:00
Nigel Tao d1b741338b Add sse128_i.load_slice128 2021-01-23 22:58:21 +11:00
Nigel Tao 2f78804e62 Use more inclusive language 2021-01-23 19:50:47 +11:00
Nigel Tao 5d88f3cb37 Remove some wuffs_base__make_empty_struct calls 2021-01-22 23:01:27 +11:00
Nigel Tao b2fe296a43 Add WUFFS_BASE__QUIRK_IGNORE_CHECKSUM 2021-01-21 22:43:24 +11:00
Nigel Tao 04ab103809 Have std/png check CRC-32 checksums
wuffs_png_decode_19k_8bpp/clang9       90.3MB/s ± 0%  89.0MB/s ± 0%  -1.49%  (p=0.008 n=5+5)
wuffs_png_decode_40k_24bpp/clang9       128MB/s ± 0%   126MB/s ± 0%  -1.42%  (p=0.008 n=5+5)
wuffs_png_decode_77k_8bpp/clang9        340MB/s ± 0%   333MB/s ± 0%  -1.81%  (p=0.008 n=5+5)
wuffs_png_decode_552k_32bpp/clang9      256MB/s ± 0%   253MB/s ± 0%  -1.06%  (p=0.008 n=5+5)
wuffs_png_decode_4002k_24bpp/clang9     128MB/s ± 0%   126MB/s ± 0%  -1.77%  (p=0.008 n=5+5)

wuffs_png_decode_19k_8bpp/gcc10        97.7MB/s ± 0%  95.2MB/s ± 0%  -2.52%  (p=0.008 n=5+5)
wuffs_png_decode_40k_24bpp/gcc10        136MB/s ± 3%   129MB/s ± 0%  -5.13%  (p=0.008 n=5+5)
wuffs_png_decode_77k_8bpp/gcc10         362MB/s ± 1%   353MB/s ± 1%  -2.52%  (p=0.008 n=5+5)
wuffs_png_decode_552k_32bpp/gcc10       273MB/s ± 0%   271MB/s ± 0%  -0.54%  (p=0.008 n=5+5)
wuffs_png_decode_4002k_24bpp/gcc10      136MB/s ± 0%   130MB/s ± 0%  -4.45%  (p=0.008 n=5+5)
2021-01-21 13:28:10 +11:00
Nigel Tao 909b933695 Fix iterate end-pointer calculation 2021-01-21 09:29:32 +11:00
Nigel Tao c9221e7c27 Clear iterate slice-variables after iteration 2021-01-19 21:59:29 +11:00
Nigel Tao a1788a0808 Add png.decoder.filter_3_distance_4_sse128
wuffs_png_decode_filt_3_dist_3/clang9   905MB/s ± 0%   913MB/s ± 0%    +0.81%  (p=0.008 n=5+5)
wuffs_png_decode_filt_3_dist_4/clang9   944MB/s ± 1%  1780MB/s ± 0%   +88.51%  (p=0.008 n=5+5)
wuffs_png_decode_filt_3_dist_3/gcc10   1.07GB/s ± 0%  1.08GB/s ± 0%    +0.42%  (p=0.008 n=5+5)
wuffs_png_decode_filt_3_dist_4/gcc10   1.01GB/s ± 1%  2.11GB/s ± 0%  +109.79%  (p=0.008 n=5+5)
2021-01-18 23:44:57 +11:00
Nigel Tao 4024f5f051 Tweak builtin.SliceU8Funcs peek method names 2021-01-18 16:01:44 +11:00
Nigel Tao 26a7280900 Add slice base.u8 peek/poke methods
wuffs_png_decode_filter_1_sub/clang9  4.59GB/s ± 0%   4.59GB/s ± 0%      ~     (p=0.151 n=5+5)
wuffs_png_decode_filter_1_sub/gcc10   1.85GB/s ± 0%   4.11GB/s ± 0%  +122.24%  (p=0.016 n=5+4)
2021-01-17 23:47:37 +11:00
Nigel Tao a1c22caa15 Rename load and store to peek and poke 2021-01-17 22:44:39 +11:00
Nigel Tao 0b6865e92c Optimize png.decoder.filter_1_distance_4_sse128
wuffs_png_decode_filter_1_sub/clang9  1.26GB/s ± 0%  4.59GB/s ± 0%  +263.24%  (p=0.008 n=5+5)
wuffs_png_decode_filter_1_sub/gcc10   1.85GB/s ± 0%  1.85GB/s ± 0%      ~     (p=0.333 n=5+5)
2021-01-15 23:28:58 +11:00
Nigel Tao afc6bb9ec3 Add base.sse128_i 2021-01-15 15:47:56 +11:00
Nigel Tao 020b8e6903 Add cpu_arch 2021-01-15 10:26:55 +11:00
Nigel Tao 6124fda77f Fix some SIMD compiler warnings 2021-01-13 23:57:58 +11:00
Nigel Tao a882558900 Spell nonpremul consistently 2021-01-13 21:37:55 +11:00
Nigel Tao dd75b6e545 SIMD-optimize some pixel swizzler loops
wuffs_pixel_swizzler_bgra_non_premul_rgba_nonpremul_src/clang9   1.75GB/s ± 0%  13.10GB/s ± 0%  +650.08%  (p=0.008 n=5+5)
wuffs_pixel_swizzler_bgra_premul_rgb_src/clang9                  1.85GB/s ± 0%  12.24GB/s ± 0%  +563.16%  (p=0.008 n=5+5)

wuffs_pixel_swizzler_bgra_non_premul_rgba_nonpremul_src/gcc10    7.41GB/s ± 0%  11.90GB/s ± 0%   +60.73%  (p=0.008 n=5+5)
wuffs_pixel_swizzler_bgra_premul_rgb_src/gcc10                   1.85GB/s ± 0%  10.66GB/s ± 0%  +474.51%  (p=0.016 n=5+4)
2021-01-13 21:36:14 +11:00
Nigel Tao 79b39e14bf Add SIMD
wuffs_pixel_swizzler_bgra_premul_y_src/clang9  1.94GB/s ± 1%  12.08GB/s ± 0%  +521.63%  (p=0.008 n=5+5)
wuffs_pixel_swizzler_bgra_premul_y_src/gcc10   2.92GB/s ± 0%  16.00GB/s ± 0%  +448.86%  (p=0.008 n=5+5)
2021-01-13 15:22:15 +11:00
Nigel Tao a1a70da539 Add iterate advance parameter 2021-01-11 00:03:44 +11:00
Nigel Tao 2b17cf72ca Fix uint64_t vs size_t when compiling with -m32 2021-01-08 23:07:07 +11:00
Nigel Tao dcdd7928c1 Let iterate have multiple assigns
wuffs_png_decode_19k_8bpp/clang9          92.2MB/s ± 0%  93.4MB/s ± 0%   +1.33%  (p=0.008 n=5+5)
wuffs_png_decode_40k_24bpp/clang9          116MB/s ± 0%   115MB/s ± 0%     ~     (p=1.000 n=5+5)
wuffs_png_decode_77k_8bpp/clang9           347MB/s ± 0%   351MB/s ± 0%   +1.19%  (p=0.016 n=5+4)
wuffs_png_decode_552k_32bpp/clang9         150MB/s ± 0%   156MB/s ± 0%   +3.57%  (p=0.008 n=5+5)
wuffs_png_decode_4002k_24bpp/clang9        117MB/s ± 1%   118MB/s ± 0%   +0.62%  (p=0.008 n=5+5)

wuffs_png_decode_filter_1_sub/clang9      1.34GB/s ± 0%  1.24GB/s ± 0%   -7.30%  (p=0.008 n=5+5)
wuffs_png_decode_filter_2_up/clang9       13.3GB/s ± 0%  13.4GB/s ± 0%   +0.96%  (p=0.008 n=5+5)
wuffs_png_decode_filter_3_average/clang9   754MB/s ± 0%   948MB/s ± 0%  +25.62%  (p=0.008 n=5+5)
wuffs_png_decode_filter_4_paeth/clang9     127MB/s ± 0%   138MB/s ± 0%   +9.01%  (p=0.008 n=5+5)

wuffs_png_decode_19k_8bpp/gcc10            100MB/s ± 1%   100MB/s ± 0%     ~     (p=0.690 n=5+5)
wuffs_png_decode_40k_24bpp/gcc10           114MB/s ± 0%   118MB/s ± 0%   +3.21%  (p=0.008 n=5+5)
wuffs_png_decode_77k_8bpp/gcc10            323MB/s ± 2%   325MB/s ± 0%     ~     (p=0.841 n=5+5)
wuffs_png_decode_552k_32bpp/gcc10          159MB/s ± 0%   172MB/s ± 0%   +8.38%  (p=0.008 n=5+5)
wuffs_png_decode_4002k_24bpp/gcc10         118MB/s ± 0%   118MB/s ± 0%     ~     (p=0.056 n=5+5)

wuffs_png_decode_filter_1_sub/gcc10       1.84GB/s ± 1%  1.85GB/s ± 0%   +0.22%  (p=0.008 n=5+5)
wuffs_png_decode_filter_2_up/gcc10        10.5GB/s ± 0%  10.5GB/s ± 0%     ~     (p=0.310 n=5+5)
wuffs_png_decode_filter_3_average/gcc10    743MB/s ± 0%  1008MB/s ± 0%  +35.66%  (p=0.008 n=5+5)
wuffs_png_decode_filter_4_paeth/gcc10     83.1MB/s ± 0%  94.1MB/s ± 0%  +13.15%  (p=0.008 n=5+5)
2021-01-07 23:40:50 +11:00
Nigel Tao d842a20e43 Let iterate parameters be non-powers-of-2 2021-01-07 22:48:31 +11:00
Nigel Tao beaf3ef11f Add choose
wuffs_png_decode_filter_1_sub/clang9       905MB/s ± 0%  1339MB/s ± 0%  +48.05%  (p=0.008 n=5+5)
wuffs_png_decode_filter_1_sub/gcc10       1.12GB/s ± 0%  1.85GB/s ± 0%  +64.80%  (p=0.016 n=5+4)
2021-01-06 23:16:16 +11:00
Nigel Tao 364c396ae8 Have cgen.go call aNode.Public() more often 2021-01-04 23:26:32 +11:00
Nigel Tao 81c6437e09 Add choosy 2021-01-04 23:20:41 +11:00
Nigel Tao b4dcc4e5ed Support swizzle from PIXEL_FORMAT__Y to same 2020-12-26 22:11:11 +11:00
Nigel Tao 43a74d6b81 Apply io_limit to original not just derived vars 2020-12-24 22:45:31 +11:00
Nigel Tao b81999f9fc Restrict an io_bind io arg to local variables only 2020-12-24 14:11:48 +11:00
Nigel Tao 78c7087eee Implement io_limit 2020-12-24 14:01:48 +11:00
Nigel Tao 332fb46b5c Remove std/nie decoder.swizzle derived var hack 2020-12-23 15:34:42 +11:00
Nigel Tao f6d11c78da Add ast.Expr IsArgsDotFoo and IsMethodCall methods 2020-12-23 15:13:14 +11:00
Nigel Tao 5a387fe090 Swizzler support for src_pixfmt = RGB|A_NONPREMUL
Binary size, before:
21520 gen/lib/c/clang-9-dynamic/wuffs-base-pixconv.lo
22536 gen/lib/c/clang-9-static/wuffs-base-pixconv.o
34488 gen/lib/c/gcc-dynamic/wuffs-base-pixconv.lo
34408 gen/lib/c/gcc-static/wuffs-base-pixconv.o

After:
25712 gen/lib/c/clang-9-dynamic/wuffs-base-pixconv.lo
26800 gen/lib/c/clang-9-static/wuffs-base-pixconv.o
42816 gen/lib/c/gcc-dynamic/wuffs-base-pixconv.lo
42768 gen/lib/c/gcc-static/wuffs-base-pixconv.o
2020-12-20 21:50:37 +11:00
Nigel Tao c79129a853 Restore deprecated {reader,writer}_io_position 2020-12-15 16:23:02 +11:00
Nigel Tao 8a12a64754 Add limited_swizzle_u32_interleaved_from_reader 2020-12-12 00:28:16 +11:00
Nigel Tao 8ee76df4f5 Add swizzle_interleaved_transparent_black decl 2020-12-11 14:22:53 +11:00
Nigel Tao 0ccdece50a Add swizzle_interleaved_transparent_black 2020-12-04 22:53:59 +11:00
Nigel Tao 0c481fe7c6 Improve cgen for constant-indexed slices of arrays 2020-12-02 22:39:34 +11:00
Nigel Tao 8482161c91 Add base.pixel_buffer.palette_or_else 2020-12-02 22:13:35 +11:00
Nigel Tao 50a0e767d4 Fix -Werror=conversion in base/image-public.h 2020-11-24 09:02:43 +11:00
Nigel Tao ac930b801b Swizzler support for src_pixfmt = BGRX
Binary size, before:
18520 gen/lib/c/clang-9-dynamic/wuffs-base-pixconv.lo
19200 gen/lib/c/clang-9-static/wuffs-base-pixconv.o
30256 gen/lib/c/gcc-dynamic/wuffs-base-pixconv.lo
30192 gen/lib/c/gcc-static/wuffs-base-pixconv.o

After:
20440 gen/lib/c/clang-9-dynamic/wuffs-base-pixconv.lo
21200 gen/lib/c/clang-9-static/wuffs-base-pixconv.o
33424 gen/lib/c/gcc-dynamic/wuffs-base-pixconv.lo
33384 gen/lib/c/gcc-static/wuffs-base-pixconv.o
2020-11-16 21:51:42 +11:00
Nigel Tao 6b69425247 Don't unroll pixel_swizzler 4x8 4x16le conversion
Binary size, before:
18712 gen/lib/c/clang-9-dynamic/wuffs-base-pixconv.lo
19392 gen/lib/c/clang-9-static/wuffs-base-pixconv.o
36880 gen/lib/c/gcc-dynamic/wuffs-base-pixconv.lo
36824 gen/lib/c/gcc-static/wuffs-base-pixconv.o

After:
18520 gen/lib/c/clang-9-dynamic/wuffs-base-pixconv.lo
19200 gen/lib/c/clang-9-static/wuffs-base-pixconv.o
30256 gen/lib/c/gcc-dynamic/wuffs-base-pixconv.lo
30192 gen/lib/c/gcc-static/wuffs-base-pixconv.o
2020-11-16 18:59:52 +11:00