From 4a55d9ba9fbd731f42c23fec10c12848345bacaf Mon Sep 17 00:00:00 2001 From: Kevin Greenan Date: Sun, 24 Nov 2013 18:40:05 -0800 Subject: [PATCH] Added performance numbers and more text to the README --- PERF.txt | 243 +++++++++++++++++++++++++++++++++++++++++++++++++++++ README.nd | 41 +++++++++ README.txt | 41 +++++++++ 3 files changed, 325 insertions(+) create mode 100644 PERF.txt diff --git a/PERF.txt b/PERF.txt new file mode 100644 index 0000000..ee27826 --- /dev/null +++ b/PERF.txt @@ -0,0 +1,243 @@ +This reflects time_all_gfs_argv_init.sh run on a MacBook Air with 4 GB of memory and a 1.7 GHz Intel Core i5 + +#uname -a +11.4.2 Darwin Kernel Version 11.4.2: Thu Aug 23 16:25:48 PDT 2012; root:xnu-1699.32.7~1/RELEASE_X86_64 x86_64 + + + +_12_3_8_128_65536_-m_SPLIT_8_4_-r_SSE_- 2813.34 +_12_3_8_128_65536_-m_COMPOSITE_2_-_-r_ALTMAP_- 2808.39 +_12_3_8_128_65536_- 2797.62 +_12_3_8_128_65536_-m_SPLIT_8_4_- 2793.14 +_12_3_8_128_65536_-m_SPLIT_8_4_-r_SSE_-d_EUCLID_- 2779.97 +_12_3_8_128_65536_-m_SPLIT_8_4_-d_EUCLID_- 2776.50 +_12_3_8_128_65536_-m_SPLIT_8_4_-d_MATRIX_- 2762.82 +_12_3_8_128_65536_-m_SPLIT_8_4_-r_SSE_-d_MATRIX_- 2711.49 +_12_3_16_128_65536_-m_COMPOSITE_2_-_-r_ALTMAP_- 2700.11 +_12_3_16_128_65536_-m_SPLIT_16_4_-r_ALTMAP_- 2367.78 +_12_3_16_128_65536_-m_SPLIT_16_4_-r_SSE_-r_ALTMAP_-d_MATRIX_- 2365.21 +_12_3_16_128_65536_-m_SPLIT_16_4_-r_ALTMAP_-d_EUCLID_- 2364.95 +_12_3_16_128_65536_-m_SPLIT_16_4_-r_SSE_-r_ALTMAP_- 2356.81 +_12_3_16_128_65536_-m_SPLIT_16_4_-r_SSE_-r_ALTMAP_-d_EUCLID_- 2319.16 +_12_3_16_128_65536_-m_SPLIT_16_4_-r_ALTMAP_-d_MATRIX_- 2307.02 +_12_3_16_128_65536_-m_SPLIT_16_4_-r_SSE_-d_EUCLID_- 1879.46 +_12_3_16_128_65536_- 1877.06 +_12_3_16_128_65536_-m_SPLIT_16_4_-d_MATRIX_- 1868.61 +_12_3_16_128_65536_-m_SPLIT_16_4_- 1864.30 +_12_3_16_128_65536_-m_SPLIT_16_4_-r_SSE_-d_MATRIX_- 1861.95 +_12_3_16_128_65536_-m_SPLIT_16_4_-d_EUCLID_- 1854.40 +_12_3_16_128_65536_-m_SPLIT_16_4_-r_SSE_- 1850.96 +_12_3_32_128_65536_-m_COMPOSITE_2_-_-r_ALTMAP_- 1828.50 +_12_3_8_128_65536_-m_TABLE_-r_CAUCHY_- 1534.09 +_12_3_8_128_65536_-m_SPLIT_8_4_-r_CAUCHY_-d_MATRIX_- 1531.78 +_12_3_8_128_65536_-m_LOG_ZERO_EXT_-r_CAUCHY_-d_EUCLID_- 1526.97 +_12_3_8_128_65536_-m_TABLE_-r_CAUCHY_-d_EUCLID_- 1527.21 +_12_3_8_128_65536_-m_LOG_-r_CAUCHY_- 1524.65 +_12_3_8_128_65536_-m_SHIFT_-r_CAUCHY_-d_MATRIX_- 1525.47 +_12_3_8_128_65536_-m_BYTWO_b_-r_CAUCHY_-d_EUCLID_- 1525.53 +_12_3_8_128_65536_-m_LOG_ZERO_EXT_-r_CAUCHY_- 1522.80 +_12_3_8_128_65536_-m_LOG_ZERO_-r_CAUCHY_-d_EUCLID_- 1522.57 +_12_3_8_128_65536_-m_BYTWO_p_-r_CAUCHY_-d_EUCLID_- 1519.20 +_12_3_8_128_65536_-m_BYTWO_p_-r_CAUCHY_-d_MATRIX_- 1517.28 +_12_3_8_128_65536_-m_LOG_-r_CAUCHY_-d_MATRIX_- 1515.57 +_12_3_8_128_65536_-m_TABLE_-r_CAUCHY_-d_MATRIX_- 1516.30 +_12_3_8_128_65536_-m_LOG_ZERO_-r_CAUCHY_- 1516.35 +_12_3_8_128_65536_-m_LOG_ZERO_-r_CAUCHY_-d_MATRIX_- 1515.13 +_12_3_8_128_65536_-m_SHIFT_-r_CAUCHY_-d_EUCLID_- 1513.07 +_12_3_8_128_65536_-m_LOG_ZERO_EXT_-r_CAUCHY_-d_MATRIX_- 1512.42 +_12_3_8_128_65536_-m_LOG_-r_CAUCHY_-d_EUCLID_- 1510.87 +_12_3_8_128_65536_-m_BYTWO_p_-r_CAUCHY_- 1511.34 +_12_3_8_128_65536_-m_BYTWO_b_-r_CAUCHY_- 1508.70 +_12_3_8_128_65536_-m_SPLIT_8_4_-r_CAUCHY_- 1508.28 +_12_3_8_128_65536_-m_SHIFT_-r_CAUCHY_- 1504.25 +_12_3_8_128_65536_-m_SPLIT_8_4_-r_CAUCHY_-d_EUCLID_- 1499.62 +_12_3_8_128_65536_-m_BYTWO_b_-r_CAUCHY_-d_MATRIX_- 1488.90 +_12_3_32_128_65536_-m_SPLIT_32_4_-r_SSE_- 1337.74 +_12_3_32_128_65536_-m_SPLIT_32_4_-r_SSE_-d_EUCLID_- 1334.65 +_12_3_32_128_65536_-m_SPLIT_32_4_-r_SSE_-d_MATRIX_- 1326.72 +_12_3_32_128_65536_-m_SPLIT_32_4_-d_MATRIX_- 1325.45 +_12_3_32_128_65536_- 1325.80 +_12_3_32_128_65536_-m_SPLIT_32_4_-d_EUCLID_- 1325.00 +_12_3_32_128_65536_-m_SPLIT_32_4_- 1300.37 +_12_3_32_128_65536_-m_SPLIT_32_4_-r_ALTMAP_-d_MATRIX_- 1196.01 +_12_3_32_128_65536_-m_SPLIT_32_4_-r_ALTMAP_- 1196.97 +_12_3_32_128_65536_-m_SPLIT_32_4_-r_ALTMAP_-d_EUCLID_- 1193.25 +_12_3_32_128_65536_-m_SPLIT_32_4_-r_SSE_-r_ALTMAP_- 1191.37 +_12_3_32_128_65536_-m_SPLIT_32_4_-r_SSE_-r_ALTMAP_-d_EUCLID_- 1188.98 +_12_3_32_128_65536_-m_SPLIT_32_4_-r_SSE_-r_ALTMAP_-d_MATRIX_- 1187.58 +_12_3_8_128_65536_-m_TABLE_-r_DOUBLE_- 1015.70 +_12_3_8_128_65536_-m_TABLE_-r_DOUBLE_-d_EUCLID_- 999.25 +_12_3_8_128_65536_-m_TABLE_-r_DOUBLE_-d_MATRIX_- 996.35 +_12_3_16_128_65536_-m_SPLIT_16_4_-r_CAUCHY_-d_EUCLID_- 971.19 +_12_3_16_128_65536_-m_LOG_-r_CAUCHY_-d_MATRIX_- 972.08 +_12_3_16_128_65536_-m_LOG_ZERO_-r_CAUCHY_- 967.87 +_12_3_16_128_65536_-m_LOG_ZERO_-r_CAUCHY_-d_EUCLID_- 965.51 +_12_3_16_128_65536_-m_SPLIT_16_8_-r_CAUCHY_-d_MATRIX_- 965.75 +_12_3_16_128_65536_-m_LOG_ZERO_-r_CAUCHY_-d_MATRIX_- 965.41 +_12_3_16_128_65536_-m_SPLIT_16_4_-r_CAUCHY_- 966.20 +_12_3_16_128_65536_-m_SPLIT_16_8_-r_CAUCHY_-d_EUCLID_- 964.21 +_12_3_16_128_65536_-m_LOG_-r_CAUCHY_- 962.11 +_12_3_16_128_65536_-m_SPLIT_8_8_-r_CAUCHY_-d_MATRIX_- 959.53 +_12_3_16_128_65536_-m_LOG_-r_CAUCHY_-d_EUCLID_- 959.75 +_12_3_16_128_65536_-m_SPLIT_16_8_-r_CAUCHY_- 958.98 +_12_3_16_128_65536_-m_BYTWO_p_-r_CAUCHY_-d_MATRIX_- 957.03 +_12_3_16_128_65536_-m_BYTWO_b_-r_CAUCHY_-d_MATRIX_- 955.82 +_12_3_16_128_65536_-m_GROUP_4_4_-r_CAUCHY_-d_EUCLID_- 956.46 +_12_3_16_128_65536_-m_GROUP_4_4_-r_CAUCHY_- 955.41 +_12_3_16_128_65536_-m_GROUP_4_4_-r_CAUCHY_-d_MATRIX_- 955.45 +_12_3_16_128_65536_-m_SPLIT_8_8_-r_CAUCHY_-d_EUCLID_- 955.93 +_12_3_16_128_65536_-m_BYTWO_p_-r_CAUCHY_-d_EUCLID_- 952.43 +_12_3_16_128_65536_-m_BYTWO_b_-r_CAUCHY_-d_EUCLID_- 951.90 +_12_3_16_128_65536_-m_BYTWO_p_-r_CAUCHY_- 945.16 +_12_3_16_128_65536_-m_BYTWO_b_-r_CAUCHY_- 945.30 +_12_3_16_128_65536_-m_SHIFT_-r_CAUCHY_- 944.22 +_12_3_16_128_65536_-m_SHIFT_-r_CAUCHY_-d_MATRIX_- 939.66 +_12_3_16_128_65536_-m_SHIFT_-r_CAUCHY_-d_EUCLID_- 934.56 +_12_3_16_128_65536_-m_SPLIT_16_4_-r_CAUCHY_-d_MATRIX_- 917.26 +_12_3_16_128_65536_-m_SPLIT_8_8_-r_CAUCHY_- 831.63 +_12_3_16_128_65536_-m_SPLIT_16_8_- 621.13 +_12_3_16_128_65536_-m_SPLIT_16_8_-d_MATRIX_- 619.98 +_12_3_16_128_65536_-m_SPLIT_8_8_- 614.67 +_12_3_16_128_65536_-m_SPLIT_16_8_-d_EUCLID_- 603.05 +_12_3_16_128_65536_-m_SPLIT_8_8_-d_MATRIX_- 553.45 +_12_3_16_128_65536_-m_SPLIT_8_8_-d_EUCLID_- 543.90 +_12_3_8_128_65536_-m_BYTWO_b_-d_EUCLID_- 522.34 +_12_3_8_128_65536_-m_BYTWO_b_-r_SSE_- 520.97 +_12_3_8_128_65536_-m_BYTWO_b_-r_SSE_-d_MATRIX_- 520.49 +_12_3_8_128_65536_-m_BYTWO_b_-d_MATRIX_- 518.74 +_12_3_8_128_65536_-m_BYTWO_b_-r_SSE_-d_EUCLID_- 518.95 +_12_3_8_128_65536_-m_BYTWO_b_- 519.19 +_12_3_16_128_65536_-m_LOG_-d_MATRIX_- 485.44 +_12_3_16_128_65536_-m_LOG_- 483.58 +_12_3_16_128_65536_-m_LOG_-d_EUCLID_- 480.18 +_12_3_16_128_65536_-m_LOG_ZERO_-d_EUCLID_- 441.75 +_12_3_16_128_65536_-m_LOG_ZERO_-d_MATRIX_- 427.39 +_12_3_16_128_65536_-m_LOG_ZERO_- 419.07 +_12_3_32_128_65536_-m_SPLIT_8_8_-d_EUCLID_- 415.80 +_12_3_32_128_65536_-m_SPLIT_8_8_- 415.31 +_12_3_32_128_65536_-m_SPLIT_32_8_-d_EUCLID_- 416.06 +_12_3_32_128_65536_-m_BYTWO_p_-r_CAUCHY_-d_EUCLID_- 415.84 +_12_3_8_128_65536_-m_BYTWO_p_-d_MATRIX_- 416.50 +_12_3_8_128_65536_-m_BYTWO_p_- 416.57 +_12_3_32_128_65536_-m_SPLIT_32_8_- 416.36 +_12_3_8_128_65536_-m_BYTWO_p_-r_SSE_- 414.77 +_12_3_8_128_65536_-m_BYTWO_p_-r_SSE_-d_EUCLID_- 414.64 +_12_3_32_128_65536_-m_BYTWO_b_-r_CAUCHY_-d_EUCLID_- 415.14 +_12_3_32_128_65536_-m_BYTWO_b_-r_CAUCHY_- 413.58 +_12_3_32_128_65536_-m_GROUP_4_8_-r_CAUCHY_-d_EUCLID_- 413.55 +_12_3_32_128_65536_-m_SPLIT_8_8_-r_CAUCHY_-d_EUCLID_- 413.76 +_12_3_32_128_65536_-m_SPLIT_32_8_-d_MATRIX_- 413.83 +_12_3_32_128_65536_-m_BYTWO_p_-r_CAUCHY_-d_MATRIX_- 412.09 +_12_3_32_128_65536_-m_SPLIT_32_16_-r_CAUCHY_- 413.06 +_12_3_32_128_65536_-m_GROUP_4_4_-r_CAUCHY_-d_EUCLID_- 413.62 +_12_3_32_128_65536_-m_SPLIT_8_8_-r_CAUCHY_-d_MATRIX_- 411.67 +_12_3_32_128_65536_-m_GROUP_4_4_-r_CAUCHY_- 412.35 +_12_3_32_128_65536_-m_SPLIT_32_4_-r_CAUCHY_-d_MATRIX_- 412.30 +_12_3_32_128_65536_-m_SPLIT_32_4_-r_CAUCHY_-d_EUCLID_- 411.08 +_12_3_32_128_65536_-m_BYTWO_p_-r_CAUCHY_- 411.89 +_12_3_32_128_65536_-m_SPLIT_8_8_-r_CAUCHY_- 412.23 +_12_3_8_128_65536_-m_BYTWO_p_-r_SSE_-d_MATRIX_- 413.10 +_12_3_32_128_65536_-m_SPLIT_32_4_-r_CAUCHY_- 411.24 +_12_3_32_128_65536_-m_SPLIT_32_16_-r_CAUCHY_-d_EUCLID_- 411.23 +_12_3_32_128_65536_-m_SPLIT_32_16_-r_CAUCHY_-d_MATRIX_- 411.49 +_12_3_32_128_65536_-m_GROUP_4_4_-r_CAUCHY_-d_MATRIX_- 410.80 +_12_3_32_128_65536_-m_SPLIT_32_8_-r_CAUCHY_-d_EUCLID_- 409.80 +_12_3_32_128_65536_-m_SPLIT_32_8_-r_CAUCHY_- 408.72 +_12_3_32_128_65536_-m_SPLIT_32_8_-r_CAUCHY_-d_MATRIX_- 409.51 +_12_3_32_128_65536_-m_GROUP_4_8_-r_CAUCHY_- 409.34 +_12_3_8_128_65536_-m_BYTWO_p_-d_EUCLID_- 405.75 +_12_3_32_128_65536_-m_BYTWO_b_-r_CAUCHY_-d_MATRIX_- 406.53 +_12_3_32_128_65536_-m_SPLIT_8_8_-d_MATRIX_- 405.91 +_12_3_32_128_65536_-m_SHIFT_-r_CAUCHY_- 403.98 +_12_3_32_128_65536_-m_SHIFT_-r_CAUCHY_-d_EUCLID_- 404.79 +_12_3_32_128_65536_-m_SHIFT_-r_CAUCHY_-d_MATRIX_- 401.29 +_12_3_8_128_65536_-m_TABLE_-r_DOUBLE_-r_LAZY_-d_EUCLID_- 384.38 +_12_3_8_128_65536_-m_TABLE_-r_DOUBLE_-r_LAZY_- 381.47 +_12_3_8_128_65536_-m_TABLE_-r_DOUBLE_-r_LAZY_-d_MATRIX_- 381.49 +_12_3_32_128_65536_-m_GROUP_4_8_-r_CAUCHY_-d_MATRIX_- 374.88 +_12_3_8_128_65536_-m_LOG_ZERO_-d_MATRIX_- 349.17 +_12_3_8_128_65536_-m_LOG_ZERO_- 349.34 +_12_3_8_128_65536_-m_LOG_ZERO_EXT_-d_MATRIX_- 349.43 +_12_3_8_128_65536_-m_LOG_ZERO_-d_EUCLID_- 349.61 +_12_3_8_128_65536_-m_LOG_ZERO_EXT_- 349.71 +_12_3_8_128_65536_-m_TABLE_-d_EUCLID_- 343.28 +_12_3_8_128_65536_-m_LOG_ZERO_EXT_-d_EUCLID_- 341.73 +_12_3_8_128_65536_-m_TABLE_-d_MATRIX_- 329.42 +_12_3_16_128_65536_-m_SPLIT_16_4_-r_NOSSE_- 327.32 +_12_3_16_128_65536_-m_SPLIT_16_4_-r_NOSSE_-d_EUCLID_- 318.34 +_12_3_16_128_65536_-m_SPLIT_16_4_-r_NOSSE_-d_MATRIX_- 317.40 +_12_3_8_128_65536_-m_TABLE_- 316.84 +_12_3_32_128_65536_-m_COMPOSITE_2_-_- 301.69 +_12_3_8_128_65536_-m_LOG_-d_MATRIX_- 281.59 +_12_3_8_128_65536_-m_LOG_-d_EUCLID_- 281.28 +_12_3_8_128_65536_-m_LOG_- 279.75 +_12_3_16_128_65536_-m_SPLIT_16_4_-r_NOSSE_-r_ALTMAP_- 275.37 +_12_3_16_128_65536_-m_SPLIT_16_4_-r_NOSSE_-r_ALTMAP_-d_EUCLID_- 276.07 +_12_3_16_128_65536_-m_SPLIT_16_4_-r_NOSSE_-r_ALTMAP_-d_MATRIX_- 269.77 +_12_3_8_128_65536_-m_BYTWO_b_-r_NOSSE_-d_MATRIX_- 257.86 +_12_3_8_128_65536_-m_BYTWO_b_-r_NOSSE_-d_EUCLID_- 256.55 +_12_3_8_128_65536_-m_SPLIT_8_4_-r_NOSSE_-d_EUCLID_- 236.07 +_12_3_8_128_65536_-m_SPLIT_8_4_-r_NOSSE_-d_MATRIX_- 236.76 +_12_3_8_128_65536_-m_SPLIT_8_4_-r_NOSSE_- 236.71 +_12_3_16_128_65536_-m_BYTWO_b_-d_EUCLID_- 217.68 +_12_3_16_128_65536_-m_BYTWO_b_- 217.87 +_12_3_16_128_65536_-m_BYTWO_b_-r_SSE_- 217.10 +_12_3_16_128_65536_-m_BYTWO_b_-r_SSE_-d_MATRIX_- 216.91 +_12_3_16_128_65536_-m_BYTWO_b_-r_SSE_-d_EUCLID_- 217.27 +_12_3_16_128_65536_-m_BYTWO_b_-d_MATRIX_- 215.01 +_12_3_8_128_65536_-m_BYTWO_p_-r_NOSSE_- 206.00 +_12_3_8_128_65536_-m_BYTWO_p_-r_NOSSE_-d_MATRIX_- 205.66 +_12_3_8_128_65536_-m_BYTWO_p_-r_NOSSE_-d_EUCLID_- 204.44 +_12_3_32_128_65536_-m_SPLIT_32_4_-r_NOSSE_-d_MATRIX_- 199.64 +_12_3_32_128_65536_-m_SPLIT_32_4_-r_NOSSE_-d_EUCLID_- 198.20 +_12_3_16_128_65536_-m_COMPOSITE_2_-_- 182.98 +_12_3_16_128_65536_-m_BYTWO_p_-r_SSE_- 183.34 +_12_3_16_128_65536_-m_BYTWO_p_-r_SSE_-d_EUCLID_- 182.87 +_12_3_16_128_65536_-m_BYTWO_p_-d_EUCLID_- 183.57 +_12_3_16_128_65536_-m_BYTWO_p_-d_MATRIX_- 183.95 +_12_3_16_128_65536_-m_BYTWO_p_-r_SSE_-d_MATRIX_- 179.94 +_12_3_16_128_65536_-m_GROUP_4_4_- 166.90 +_12_3_32_128_65536_-m_SPLIT_32_16_- 167.55 +_12_3_32_128_65536_-m_SPLIT_32_16_-d_MATRIX_- 167.51 +_12_3_32_128_65536_-m_SPLIT_32_16_-d_EUCLID_- 167.70 +_12_3_16_128_65536_-m_GROUP_4_4_-d_EUCLID_- 166.56 +_12_3_16_128_65536_-m_GROUP_4_4_-d_MATRIX_- 167.07 +_12_3_16_128_65536_-m_BYTWO_b_-r_NOSSE_-d_EUCLID_- 110.64 +_12_3_16_128_65536_-m_BYTWO_b_-r_NOSSE_- 110.76 +_12_3_16_128_65536_-m_BYTWO_b_-r_NOSSE_-d_MATRIX_- 111.17 +_12_3_16_128_65536_-m_BYTWO_p_-r_NOSSE_- 100.39 +_12_3_16_128_65536_-m_BYTWO_p_-r_NOSSE_-d_EUCLID_- 100.27 +_12_3_16_128_65536_-m_BYTWO_p_-r_NOSSE_-d_MATRIX_- 100.18 +_12_3_32_128_65536_-m_BYTWO_b_-r_SSE_- 96.85 +_12_3_32_128_65536_-m_BYTWO_b_-d_MATRIX_- 97.76 +_12_3_32_128_65536_-m_BYTWO_b_-r_SSE_-d_EUCLID_- 97.69 +_12_3_32_128_65536_-m_BYTWO_b_-r_SSE_-d_MATRIX_- 97.48 +_12_3_32_128_65536_-m_BYTWO_b_-d_EUCLID_- 97.42 +_12_3_32_128_65536_-m_BYTWO_b_- 97.54 +_12_3_32_128_65536_-m_BYTWO_p_- 86.61 +_12_3_32_128_65536_-m_BYTWO_p_-d_MATRIX_- 86.81 +_12_3_32_128_65536_-m_BYTWO_p_-r_SSE_-d_MATRIX_- 86.31 +_12_3_32_128_65536_-m_BYTWO_p_-d_EUCLID_- 86.70 +_12_3_32_128_65536_-m_BYTWO_p_-r_SSE_- 86.69 +_12_3_32_128_65536_-m_BYTWO_p_-r_SSE_-d_EUCLID_- 86.80 +_12_3_8_128_65536_-m_COMPOSITE_2_-_- 76.58 +_12_3_32_128_65536_-m_GROUP_4_8_-d_EUCLID_- 57.06 +_12_3_32_128_65536_-m_GROUP_4_8_-d_MATRIX_- 57.08 +_12_3_32_128_65536_-m_GROUP_4_4_-d_EUCLID_- 56.59 +_12_3_32_128_65536_-m_GROUP_4_4_-d_MATRIX_- 56.91 +_12_3_32_128_65536_-m_GROUP_4_8_- 54.68 +_12_3_32_128_65536_-m_BYTWO_b_-r_NOSSE_-d_EUCLID_- 50.64 +_12_3_32_128_65536_-m_BYTWO_b_-r_NOSSE_-d_MATRIX_- 50.55 +_12_3_32_128_65536_-m_BYTWO_b_-r_NOSSE_- 50.75 +_12_3_32_128_65536_-m_BYTWO_p_-r_NOSSE_-d_EUCLID_- 49.50 +_12_3_32_128_65536_-m_BYTWO_p_-r_NOSSE_- 49.14 +_12_3_32_128_65536_-m_BYTWO_p_-r_NOSSE_-d_MATRIX_- 47.39 +_12_3_8_128_65536_-m_SHIFT_-d_EUCLID_- 14.08 +_12_3_8_128_65536_-m_SHIFT_-d_MATRIX_- 14.39 +_12_3_8_128_65536_-m_SHIFT_- 14.48 +_12_3_16_128_65536_-m_SHIFT_-d_EUCLID_- 12.75 +_12_3_16_128_65536_-m_SHIFT_-d_MATRIX_- 12.74 +_12_3_16_128_65536_-m_SHIFT_- 12.77 +_12_3_32_128_65536_-m_SHIFT_-d_MATRIX_- 12.51 +_12_3_32_128_65536_-m_SHIFT_-d_EUCLID_- 12.50 +_12_3_32_128_65536_-m_SHIFT_- 12.46 diff --git a/README.nd b/README.nd index e1c7628..d501aec 100644 --- a/README.nd +++ b/README.nd @@ -20,6 +20,26 @@ The Examples directory contains the example programs. The makefile assumes that Examples is a subdirectory of the home directory. +Installing: + +1.) Install GF-Complete + +2.) 'cd' to root directory of Jerasure + +3.) Set PREFIX in ./makefile to appropriate directory (/usr/local by default) + +4.) make + +5.) make install + +Running tests and examples: + +1.) 'cd' into Examples + +2.) make + +See individual source file to determine what the examples do. + Inclusion of GF-Complete: As long as GF-Complete is installed, Jerasure 2.0 can be used just as previous @@ -138,3 +158,24 @@ For more information on how to change the backing fields for Jerasure, please re for a backing GF (uses galois_init_field and galois_init_composite_field to get gf_t pointer) +Performance: + +There are two performance-based test scripts: time_all_gfs_argv_init.sh and +time_all_gfs_hard_init.sh. Both scripts run the same tests, but initialize the +underlying GF fields in different ways (*argv* uses reed_sol_time_gf and *hard* +uses reed_sol_hard_time_gf.c). + +You can run 'time_all_gfs_argv_init.sh' to time *all* possible GF +implementations on your computer. This script requires the 'gf_methods' +utility from GF-Complete to be in your PATH. + +time_all_gfs_argv_init.sh was run on a MacBook Air and the distilled numbers +are given in ./PERF.txt. The results are sorted by encoding throughput. The +format of each entry reflects the arguments given to reed_sol_time_gf. For example, +the test run of + +'Examples/reed_sol_time_gf 12 3 8 128 65536 -m SPLIT 8 4 -r SSE' + +is recorded as this in PERF.txt: + +_12_3_8_128_65536_-m_SPLIT_8_4_-r_SSE_- 2813.34 diff --git a/README.txt b/README.txt index e1c7628..d501aec 100644 --- a/README.txt +++ b/README.txt @@ -20,6 +20,26 @@ The Examples directory contains the example programs. The makefile assumes that Examples is a subdirectory of the home directory. +Installing: + +1.) Install GF-Complete + +2.) 'cd' to root directory of Jerasure + +3.) Set PREFIX in ./makefile to appropriate directory (/usr/local by default) + +4.) make + +5.) make install + +Running tests and examples: + +1.) 'cd' into Examples + +2.) make + +See individual source file to determine what the examples do. + Inclusion of GF-Complete: As long as GF-Complete is installed, Jerasure 2.0 can be used just as previous @@ -138,3 +158,24 @@ For more information on how to change the backing fields for Jerasure, please re for a backing GF (uses galois_init_field and galois_init_composite_field to get gf_t pointer) +Performance: + +There are two performance-based test scripts: time_all_gfs_argv_init.sh and +time_all_gfs_hard_init.sh. Both scripts run the same tests, but initialize the +underlying GF fields in different ways (*argv* uses reed_sol_time_gf and *hard* +uses reed_sol_hard_time_gf.c). + +You can run 'time_all_gfs_argv_init.sh' to time *all* possible GF +implementations on your computer. This script requires the 'gf_methods' +utility from GF-Complete to be in your PATH. + +time_all_gfs_argv_init.sh was run on a MacBook Air and the distilled numbers +are given in ./PERF.txt. The results are sorted by encoding throughput. The +format of each entry reflects the arguments given to reed_sol_time_gf. For example, +the test run of + +'Examples/reed_sol_time_gf 12 3 8 128 65536 -m SPLIT 8 4 -r SSE' + +is recorded as this in PERF.txt: + +_12_3_8_128_65536_-m_SPLIT_8_4_-r_SSE_- 2813.34