Commit 4a55d9ba9fbd731f42c23fec10c12848345bacaf

Authored by Kevin Greenan
1 parent d8c84dcb
Exists in master and in 1 other branch v2

Added performance numbers and more text to the README

Showing 3 changed files with 325 additions and 0 deletions   Show diff stats
PERF.txt 0 → 100644
... ... @@ -0,0 +1,243 @@
  1 +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
  2 +
  3 +#uname -a
  4 +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
  5 +
  6 +<Arguments to reed_sol_time_gf> <Throughput MB/s>
  7 +
  8 +_12_3_8_128_65536_-m_SPLIT_8_4_-r_SSE_- 2813.34
  9 +_12_3_8_128_65536_-m_COMPOSITE_2_-_-r_ALTMAP_- 2808.39
  10 +_12_3_8_128_65536_- 2797.62
  11 +_12_3_8_128_65536_-m_SPLIT_8_4_- 2793.14
  12 +_12_3_8_128_65536_-m_SPLIT_8_4_-r_SSE_-d_EUCLID_- 2779.97
  13 +_12_3_8_128_65536_-m_SPLIT_8_4_-d_EUCLID_- 2776.50
  14 +_12_3_8_128_65536_-m_SPLIT_8_4_-d_MATRIX_- 2762.82
  15 +_12_3_8_128_65536_-m_SPLIT_8_4_-r_SSE_-d_MATRIX_- 2711.49
  16 +_12_3_16_128_65536_-m_COMPOSITE_2_-_-r_ALTMAP_- 2700.11
  17 +_12_3_16_128_65536_-m_SPLIT_16_4_-r_ALTMAP_- 2367.78
  18 +_12_3_16_128_65536_-m_SPLIT_16_4_-r_SSE_-r_ALTMAP_-d_MATRIX_- 2365.21
  19 +_12_3_16_128_65536_-m_SPLIT_16_4_-r_ALTMAP_-d_EUCLID_- 2364.95
  20 +_12_3_16_128_65536_-m_SPLIT_16_4_-r_SSE_-r_ALTMAP_- 2356.81
  21 +_12_3_16_128_65536_-m_SPLIT_16_4_-r_SSE_-r_ALTMAP_-d_EUCLID_- 2319.16
  22 +_12_3_16_128_65536_-m_SPLIT_16_4_-r_ALTMAP_-d_MATRIX_- 2307.02
  23 +_12_3_16_128_65536_-m_SPLIT_16_4_-r_SSE_-d_EUCLID_- 1879.46
  24 +_12_3_16_128_65536_- 1877.06
  25 +_12_3_16_128_65536_-m_SPLIT_16_4_-d_MATRIX_- 1868.61
  26 +_12_3_16_128_65536_-m_SPLIT_16_4_- 1864.30
  27 +_12_3_16_128_65536_-m_SPLIT_16_4_-r_SSE_-d_MATRIX_- 1861.95
  28 +_12_3_16_128_65536_-m_SPLIT_16_4_-d_EUCLID_- 1854.40
  29 +_12_3_16_128_65536_-m_SPLIT_16_4_-r_SSE_- 1850.96
  30 +_12_3_32_128_65536_-m_COMPOSITE_2_-_-r_ALTMAP_- 1828.50
  31 +_12_3_8_128_65536_-m_TABLE_-r_CAUCHY_- 1534.09
  32 +_12_3_8_128_65536_-m_SPLIT_8_4_-r_CAUCHY_-d_MATRIX_- 1531.78
  33 +_12_3_8_128_65536_-m_LOG_ZERO_EXT_-r_CAUCHY_-d_EUCLID_- 1526.97
  34 +_12_3_8_128_65536_-m_TABLE_-r_CAUCHY_-d_EUCLID_- 1527.21
  35 +_12_3_8_128_65536_-m_LOG_-r_CAUCHY_- 1524.65
  36 +_12_3_8_128_65536_-m_SHIFT_-r_CAUCHY_-d_MATRIX_- 1525.47
  37 +_12_3_8_128_65536_-m_BYTWO_b_-r_CAUCHY_-d_EUCLID_- 1525.53
  38 +_12_3_8_128_65536_-m_LOG_ZERO_EXT_-r_CAUCHY_- 1522.80
  39 +_12_3_8_128_65536_-m_LOG_ZERO_-r_CAUCHY_-d_EUCLID_- 1522.57
  40 +_12_3_8_128_65536_-m_BYTWO_p_-r_CAUCHY_-d_EUCLID_- 1519.20
  41 +_12_3_8_128_65536_-m_BYTWO_p_-r_CAUCHY_-d_MATRIX_- 1517.28
  42 +_12_3_8_128_65536_-m_LOG_-r_CAUCHY_-d_MATRIX_- 1515.57
  43 +_12_3_8_128_65536_-m_TABLE_-r_CAUCHY_-d_MATRIX_- 1516.30
  44 +_12_3_8_128_65536_-m_LOG_ZERO_-r_CAUCHY_- 1516.35
  45 +_12_3_8_128_65536_-m_LOG_ZERO_-r_CAUCHY_-d_MATRIX_- 1515.13
  46 +_12_3_8_128_65536_-m_SHIFT_-r_CAUCHY_-d_EUCLID_- 1513.07
  47 +_12_3_8_128_65536_-m_LOG_ZERO_EXT_-r_CAUCHY_-d_MATRIX_- 1512.42
  48 +_12_3_8_128_65536_-m_LOG_-r_CAUCHY_-d_EUCLID_- 1510.87
  49 +_12_3_8_128_65536_-m_BYTWO_p_-r_CAUCHY_- 1511.34
  50 +_12_3_8_128_65536_-m_BYTWO_b_-r_CAUCHY_- 1508.70
  51 +_12_3_8_128_65536_-m_SPLIT_8_4_-r_CAUCHY_- 1508.28
  52 +_12_3_8_128_65536_-m_SHIFT_-r_CAUCHY_- 1504.25
  53 +_12_3_8_128_65536_-m_SPLIT_8_4_-r_CAUCHY_-d_EUCLID_- 1499.62
  54 +_12_3_8_128_65536_-m_BYTWO_b_-r_CAUCHY_-d_MATRIX_- 1488.90
  55 +_12_3_32_128_65536_-m_SPLIT_32_4_-r_SSE_- 1337.74
  56 +_12_3_32_128_65536_-m_SPLIT_32_4_-r_SSE_-d_EUCLID_- 1334.65
  57 +_12_3_32_128_65536_-m_SPLIT_32_4_-r_SSE_-d_MATRIX_- 1326.72
  58 +_12_3_32_128_65536_-m_SPLIT_32_4_-d_MATRIX_- 1325.45
  59 +_12_3_32_128_65536_- 1325.80
  60 +_12_3_32_128_65536_-m_SPLIT_32_4_-d_EUCLID_- 1325.00
  61 +_12_3_32_128_65536_-m_SPLIT_32_4_- 1300.37
  62 +_12_3_32_128_65536_-m_SPLIT_32_4_-r_ALTMAP_-d_MATRIX_- 1196.01
  63 +_12_3_32_128_65536_-m_SPLIT_32_4_-r_ALTMAP_- 1196.97
  64 +_12_3_32_128_65536_-m_SPLIT_32_4_-r_ALTMAP_-d_EUCLID_- 1193.25
  65 +_12_3_32_128_65536_-m_SPLIT_32_4_-r_SSE_-r_ALTMAP_- 1191.37
  66 +_12_3_32_128_65536_-m_SPLIT_32_4_-r_SSE_-r_ALTMAP_-d_EUCLID_- 1188.98
  67 +_12_3_32_128_65536_-m_SPLIT_32_4_-r_SSE_-r_ALTMAP_-d_MATRIX_- 1187.58
  68 +_12_3_8_128_65536_-m_TABLE_-r_DOUBLE_- 1015.70
  69 +_12_3_8_128_65536_-m_TABLE_-r_DOUBLE_-d_EUCLID_- 999.25
  70 +_12_3_8_128_65536_-m_TABLE_-r_DOUBLE_-d_MATRIX_- 996.35
  71 +_12_3_16_128_65536_-m_SPLIT_16_4_-r_CAUCHY_-d_EUCLID_- 971.19
  72 +_12_3_16_128_65536_-m_LOG_-r_CAUCHY_-d_MATRIX_- 972.08
  73 +_12_3_16_128_65536_-m_LOG_ZERO_-r_CAUCHY_- 967.87
  74 +_12_3_16_128_65536_-m_LOG_ZERO_-r_CAUCHY_-d_EUCLID_- 965.51
  75 +_12_3_16_128_65536_-m_SPLIT_16_8_-r_CAUCHY_-d_MATRIX_- 965.75
  76 +_12_3_16_128_65536_-m_LOG_ZERO_-r_CAUCHY_-d_MATRIX_- 965.41
  77 +_12_3_16_128_65536_-m_SPLIT_16_4_-r_CAUCHY_- 966.20
  78 +_12_3_16_128_65536_-m_SPLIT_16_8_-r_CAUCHY_-d_EUCLID_- 964.21
  79 +_12_3_16_128_65536_-m_LOG_-r_CAUCHY_- 962.11
  80 +_12_3_16_128_65536_-m_SPLIT_8_8_-r_CAUCHY_-d_MATRIX_- 959.53
  81 +_12_3_16_128_65536_-m_LOG_-r_CAUCHY_-d_EUCLID_- 959.75
  82 +_12_3_16_128_65536_-m_SPLIT_16_8_-r_CAUCHY_- 958.98
  83 +_12_3_16_128_65536_-m_BYTWO_p_-r_CAUCHY_-d_MATRIX_- 957.03
  84 +_12_3_16_128_65536_-m_BYTWO_b_-r_CAUCHY_-d_MATRIX_- 955.82
  85 +_12_3_16_128_65536_-m_GROUP_4_4_-r_CAUCHY_-d_EUCLID_- 956.46
  86 +_12_3_16_128_65536_-m_GROUP_4_4_-r_CAUCHY_- 955.41
  87 +_12_3_16_128_65536_-m_GROUP_4_4_-r_CAUCHY_-d_MATRIX_- 955.45
  88 +_12_3_16_128_65536_-m_SPLIT_8_8_-r_CAUCHY_-d_EUCLID_- 955.93
  89 +_12_3_16_128_65536_-m_BYTWO_p_-r_CAUCHY_-d_EUCLID_- 952.43
  90 +_12_3_16_128_65536_-m_BYTWO_b_-r_CAUCHY_-d_EUCLID_- 951.90
  91 +_12_3_16_128_65536_-m_BYTWO_p_-r_CAUCHY_- 945.16
  92 +_12_3_16_128_65536_-m_BYTWO_b_-r_CAUCHY_- 945.30
  93 +_12_3_16_128_65536_-m_SHIFT_-r_CAUCHY_- 944.22
  94 +_12_3_16_128_65536_-m_SHIFT_-r_CAUCHY_-d_MATRIX_- 939.66
  95 +_12_3_16_128_65536_-m_SHIFT_-r_CAUCHY_-d_EUCLID_- 934.56
  96 +_12_3_16_128_65536_-m_SPLIT_16_4_-r_CAUCHY_-d_MATRIX_- 917.26
  97 +_12_3_16_128_65536_-m_SPLIT_8_8_-r_CAUCHY_- 831.63
  98 +_12_3_16_128_65536_-m_SPLIT_16_8_- 621.13
  99 +_12_3_16_128_65536_-m_SPLIT_16_8_-d_MATRIX_- 619.98
  100 +_12_3_16_128_65536_-m_SPLIT_8_8_- 614.67
  101 +_12_3_16_128_65536_-m_SPLIT_16_8_-d_EUCLID_- 603.05
  102 +_12_3_16_128_65536_-m_SPLIT_8_8_-d_MATRIX_- 553.45
  103 +_12_3_16_128_65536_-m_SPLIT_8_8_-d_EUCLID_- 543.90
  104 +_12_3_8_128_65536_-m_BYTWO_b_-d_EUCLID_- 522.34
  105 +_12_3_8_128_65536_-m_BYTWO_b_-r_SSE_- 520.97
  106 +_12_3_8_128_65536_-m_BYTWO_b_-r_SSE_-d_MATRIX_- 520.49
  107 +_12_3_8_128_65536_-m_BYTWO_b_-d_MATRIX_- 518.74
  108 +_12_3_8_128_65536_-m_BYTWO_b_-r_SSE_-d_EUCLID_- 518.95
  109 +_12_3_8_128_65536_-m_BYTWO_b_- 519.19
  110 +_12_3_16_128_65536_-m_LOG_-d_MATRIX_- 485.44
  111 +_12_3_16_128_65536_-m_LOG_- 483.58
  112 +_12_3_16_128_65536_-m_LOG_-d_EUCLID_- 480.18
  113 +_12_3_16_128_65536_-m_LOG_ZERO_-d_EUCLID_- 441.75
  114 +_12_3_16_128_65536_-m_LOG_ZERO_-d_MATRIX_- 427.39
  115 +_12_3_16_128_65536_-m_LOG_ZERO_- 419.07
  116 +_12_3_32_128_65536_-m_SPLIT_8_8_-d_EUCLID_- 415.80
  117 +_12_3_32_128_65536_-m_SPLIT_8_8_- 415.31
  118 +_12_3_32_128_65536_-m_SPLIT_32_8_-d_EUCLID_- 416.06
  119 +_12_3_32_128_65536_-m_BYTWO_p_-r_CAUCHY_-d_EUCLID_- 415.84
  120 +_12_3_8_128_65536_-m_BYTWO_p_-d_MATRIX_- 416.50
  121 +_12_3_8_128_65536_-m_BYTWO_p_- 416.57
  122 +_12_3_32_128_65536_-m_SPLIT_32_8_- 416.36
  123 +_12_3_8_128_65536_-m_BYTWO_p_-r_SSE_- 414.77
  124 +_12_3_8_128_65536_-m_BYTWO_p_-r_SSE_-d_EUCLID_- 414.64
  125 +_12_3_32_128_65536_-m_BYTWO_b_-r_CAUCHY_-d_EUCLID_- 415.14
  126 +_12_3_32_128_65536_-m_BYTWO_b_-r_CAUCHY_- 413.58
  127 +_12_3_32_128_65536_-m_GROUP_4_8_-r_CAUCHY_-d_EUCLID_- 413.55
  128 +_12_3_32_128_65536_-m_SPLIT_8_8_-r_CAUCHY_-d_EUCLID_- 413.76
  129 +_12_3_32_128_65536_-m_SPLIT_32_8_-d_MATRIX_- 413.83
  130 +_12_3_32_128_65536_-m_BYTWO_p_-r_CAUCHY_-d_MATRIX_- 412.09
  131 +_12_3_32_128_65536_-m_SPLIT_32_16_-r_CAUCHY_- 413.06
  132 +_12_3_32_128_65536_-m_GROUP_4_4_-r_CAUCHY_-d_EUCLID_- 413.62
  133 +_12_3_32_128_65536_-m_SPLIT_8_8_-r_CAUCHY_-d_MATRIX_- 411.67
  134 +_12_3_32_128_65536_-m_GROUP_4_4_-r_CAUCHY_- 412.35
  135 +_12_3_32_128_65536_-m_SPLIT_32_4_-r_CAUCHY_-d_MATRIX_- 412.30
  136 +_12_3_32_128_65536_-m_SPLIT_32_4_-r_CAUCHY_-d_EUCLID_- 411.08
  137 +_12_3_32_128_65536_-m_BYTWO_p_-r_CAUCHY_- 411.89
  138 +_12_3_32_128_65536_-m_SPLIT_8_8_-r_CAUCHY_- 412.23
  139 +_12_3_8_128_65536_-m_BYTWO_p_-r_SSE_-d_MATRIX_- 413.10
  140 +_12_3_32_128_65536_-m_SPLIT_32_4_-r_CAUCHY_- 411.24
  141 +_12_3_32_128_65536_-m_SPLIT_32_16_-r_CAUCHY_-d_EUCLID_- 411.23
  142 +_12_3_32_128_65536_-m_SPLIT_32_16_-r_CAUCHY_-d_MATRIX_- 411.49
  143 +_12_3_32_128_65536_-m_GROUP_4_4_-r_CAUCHY_-d_MATRIX_- 410.80
  144 +_12_3_32_128_65536_-m_SPLIT_32_8_-r_CAUCHY_-d_EUCLID_- 409.80
  145 +_12_3_32_128_65536_-m_SPLIT_32_8_-r_CAUCHY_- 408.72
  146 +_12_3_32_128_65536_-m_SPLIT_32_8_-r_CAUCHY_-d_MATRIX_- 409.51
  147 +_12_3_32_128_65536_-m_GROUP_4_8_-r_CAUCHY_- 409.34
  148 +_12_3_8_128_65536_-m_BYTWO_p_-d_EUCLID_- 405.75
  149 +_12_3_32_128_65536_-m_BYTWO_b_-r_CAUCHY_-d_MATRIX_- 406.53
  150 +_12_3_32_128_65536_-m_SPLIT_8_8_-d_MATRIX_- 405.91
  151 +_12_3_32_128_65536_-m_SHIFT_-r_CAUCHY_- 403.98
  152 +_12_3_32_128_65536_-m_SHIFT_-r_CAUCHY_-d_EUCLID_- 404.79
  153 +_12_3_32_128_65536_-m_SHIFT_-r_CAUCHY_-d_MATRIX_- 401.29
  154 +_12_3_8_128_65536_-m_TABLE_-r_DOUBLE_-r_LAZY_-d_EUCLID_- 384.38
  155 +_12_3_8_128_65536_-m_TABLE_-r_DOUBLE_-r_LAZY_- 381.47
  156 +_12_3_8_128_65536_-m_TABLE_-r_DOUBLE_-r_LAZY_-d_MATRIX_- 381.49
  157 +_12_3_32_128_65536_-m_GROUP_4_8_-r_CAUCHY_-d_MATRIX_- 374.88
  158 +_12_3_8_128_65536_-m_LOG_ZERO_-d_MATRIX_- 349.17
  159 +_12_3_8_128_65536_-m_LOG_ZERO_- 349.34
  160 +_12_3_8_128_65536_-m_LOG_ZERO_EXT_-d_MATRIX_- 349.43
  161 +_12_3_8_128_65536_-m_LOG_ZERO_-d_EUCLID_- 349.61
  162 +_12_3_8_128_65536_-m_LOG_ZERO_EXT_- 349.71
  163 +_12_3_8_128_65536_-m_TABLE_-d_EUCLID_- 343.28
  164 +_12_3_8_128_65536_-m_LOG_ZERO_EXT_-d_EUCLID_- 341.73
  165 +_12_3_8_128_65536_-m_TABLE_-d_MATRIX_- 329.42
  166 +_12_3_16_128_65536_-m_SPLIT_16_4_-r_NOSSE_- 327.32
  167 +_12_3_16_128_65536_-m_SPLIT_16_4_-r_NOSSE_-d_EUCLID_- 318.34
  168 +_12_3_16_128_65536_-m_SPLIT_16_4_-r_NOSSE_-d_MATRIX_- 317.40
  169 +_12_3_8_128_65536_-m_TABLE_- 316.84
  170 +_12_3_32_128_65536_-m_COMPOSITE_2_-_- 301.69
  171 +_12_3_8_128_65536_-m_LOG_-d_MATRIX_- 281.59
  172 +_12_3_8_128_65536_-m_LOG_-d_EUCLID_- 281.28
  173 +_12_3_8_128_65536_-m_LOG_- 279.75
  174 +_12_3_16_128_65536_-m_SPLIT_16_4_-r_NOSSE_-r_ALTMAP_- 275.37
  175 +_12_3_16_128_65536_-m_SPLIT_16_4_-r_NOSSE_-r_ALTMAP_-d_EUCLID_- 276.07
  176 +_12_3_16_128_65536_-m_SPLIT_16_4_-r_NOSSE_-r_ALTMAP_-d_MATRIX_- 269.77
  177 +_12_3_8_128_65536_-m_BYTWO_b_-r_NOSSE_-d_MATRIX_- 257.86
  178 +_12_3_8_128_65536_-m_BYTWO_b_-r_NOSSE_-d_EUCLID_- 256.55
  179 +_12_3_8_128_65536_-m_SPLIT_8_4_-r_NOSSE_-d_EUCLID_- 236.07
  180 +_12_3_8_128_65536_-m_SPLIT_8_4_-r_NOSSE_-d_MATRIX_- 236.76
  181 +_12_3_8_128_65536_-m_SPLIT_8_4_-r_NOSSE_- 236.71
  182 +_12_3_16_128_65536_-m_BYTWO_b_-d_EUCLID_- 217.68
  183 +_12_3_16_128_65536_-m_BYTWO_b_- 217.87
  184 +_12_3_16_128_65536_-m_BYTWO_b_-r_SSE_- 217.10
  185 +_12_3_16_128_65536_-m_BYTWO_b_-r_SSE_-d_MATRIX_- 216.91
  186 +_12_3_16_128_65536_-m_BYTWO_b_-r_SSE_-d_EUCLID_- 217.27
  187 +_12_3_16_128_65536_-m_BYTWO_b_-d_MATRIX_- 215.01
  188 +_12_3_8_128_65536_-m_BYTWO_p_-r_NOSSE_- 206.00
  189 +_12_3_8_128_65536_-m_BYTWO_p_-r_NOSSE_-d_MATRIX_- 205.66
  190 +_12_3_8_128_65536_-m_BYTWO_p_-r_NOSSE_-d_EUCLID_- 204.44
  191 +_12_3_32_128_65536_-m_SPLIT_32_4_-r_NOSSE_-d_MATRIX_- 199.64
  192 +_12_3_32_128_65536_-m_SPLIT_32_4_-r_NOSSE_-d_EUCLID_- 198.20
  193 +_12_3_16_128_65536_-m_COMPOSITE_2_-_- 182.98
  194 +_12_3_16_128_65536_-m_BYTWO_p_-r_SSE_- 183.34
  195 +_12_3_16_128_65536_-m_BYTWO_p_-r_SSE_-d_EUCLID_- 182.87
  196 +_12_3_16_128_65536_-m_BYTWO_p_-d_EUCLID_- 183.57
  197 +_12_3_16_128_65536_-m_BYTWO_p_-d_MATRIX_- 183.95
  198 +_12_3_16_128_65536_-m_BYTWO_p_-r_SSE_-d_MATRIX_- 179.94
  199 +_12_3_16_128_65536_-m_GROUP_4_4_- 166.90
  200 +_12_3_32_128_65536_-m_SPLIT_32_16_- 167.55
  201 +_12_3_32_128_65536_-m_SPLIT_32_16_-d_MATRIX_- 167.51
  202 +_12_3_32_128_65536_-m_SPLIT_32_16_-d_EUCLID_- 167.70
  203 +_12_3_16_128_65536_-m_GROUP_4_4_-d_EUCLID_- 166.56
  204 +_12_3_16_128_65536_-m_GROUP_4_4_-d_MATRIX_- 167.07
  205 +_12_3_16_128_65536_-m_BYTWO_b_-r_NOSSE_-d_EUCLID_- 110.64
  206 +_12_3_16_128_65536_-m_BYTWO_b_-r_NOSSE_- 110.76
  207 +_12_3_16_128_65536_-m_BYTWO_b_-r_NOSSE_-d_MATRIX_- 111.17
  208 +_12_3_16_128_65536_-m_BYTWO_p_-r_NOSSE_- 100.39
  209 +_12_3_16_128_65536_-m_BYTWO_p_-r_NOSSE_-d_EUCLID_- 100.27
  210 +_12_3_16_128_65536_-m_BYTWO_p_-r_NOSSE_-d_MATRIX_- 100.18
  211 +_12_3_32_128_65536_-m_BYTWO_b_-r_SSE_- 96.85
  212 +_12_3_32_128_65536_-m_BYTWO_b_-d_MATRIX_- 97.76
  213 +_12_3_32_128_65536_-m_BYTWO_b_-r_SSE_-d_EUCLID_- 97.69
  214 +_12_3_32_128_65536_-m_BYTWO_b_-r_SSE_-d_MATRIX_- 97.48
  215 +_12_3_32_128_65536_-m_BYTWO_b_-d_EUCLID_- 97.42
  216 +_12_3_32_128_65536_-m_BYTWO_b_- 97.54
  217 +_12_3_32_128_65536_-m_BYTWO_p_- 86.61
  218 +_12_3_32_128_65536_-m_BYTWO_p_-d_MATRIX_- 86.81
  219 +_12_3_32_128_65536_-m_BYTWO_p_-r_SSE_-d_MATRIX_- 86.31
  220 +_12_3_32_128_65536_-m_BYTWO_p_-d_EUCLID_- 86.70
  221 +_12_3_32_128_65536_-m_BYTWO_p_-r_SSE_- 86.69
  222 +_12_3_32_128_65536_-m_BYTWO_p_-r_SSE_-d_EUCLID_- 86.80
  223 +_12_3_8_128_65536_-m_COMPOSITE_2_-_- 76.58
  224 +_12_3_32_128_65536_-m_GROUP_4_8_-d_EUCLID_- 57.06
  225 +_12_3_32_128_65536_-m_GROUP_4_8_-d_MATRIX_- 57.08
  226 +_12_3_32_128_65536_-m_GROUP_4_4_-d_EUCLID_- 56.59
  227 +_12_3_32_128_65536_-m_GROUP_4_4_-d_MATRIX_- 56.91
  228 +_12_3_32_128_65536_-m_GROUP_4_8_- 54.68
  229 +_12_3_32_128_65536_-m_BYTWO_b_-r_NOSSE_-d_EUCLID_- 50.64
  230 +_12_3_32_128_65536_-m_BYTWO_b_-r_NOSSE_-d_MATRIX_- 50.55
  231 +_12_3_32_128_65536_-m_BYTWO_b_-r_NOSSE_- 50.75
  232 +_12_3_32_128_65536_-m_BYTWO_p_-r_NOSSE_-d_EUCLID_- 49.50
  233 +_12_3_32_128_65536_-m_BYTWO_p_-r_NOSSE_- 49.14
  234 +_12_3_32_128_65536_-m_BYTWO_p_-r_NOSSE_-d_MATRIX_- 47.39
  235 +_12_3_8_128_65536_-m_SHIFT_-d_EUCLID_- 14.08
  236 +_12_3_8_128_65536_-m_SHIFT_-d_MATRIX_- 14.39
  237 +_12_3_8_128_65536_-m_SHIFT_- 14.48
  238 +_12_3_16_128_65536_-m_SHIFT_-d_EUCLID_- 12.75
  239 +_12_3_16_128_65536_-m_SHIFT_-d_MATRIX_- 12.74
  240 +_12_3_16_128_65536_-m_SHIFT_- 12.77
  241 +_12_3_32_128_65536_-m_SHIFT_-d_MATRIX_- 12.51
  242 +_12_3_32_128_65536_-m_SHIFT_-d_EUCLID_- 12.50
  243 +_12_3_32_128_65536_-m_SHIFT_- 12.46
... ...
README.nd
... ... @@ -20,6 +20,26 @@ The Examples directory contains the example programs.
20 20  
21 21 The makefile assumes that Examples is a subdirectory of the home directory.
22 22  
  23 +Installing:
  24 +
  25 +1.) Install GF-Complete
  26 +
  27 +2.) 'cd' to root directory of Jerasure
  28 +
  29 +3.) Set PREFIX in ./makefile to appropriate directory (/usr/local by default)
  30 +
  31 +4.) make
  32 +
  33 +5.) make install
  34 +
  35 +Running tests and examples:
  36 +
  37 +1.) 'cd' into Examples
  38 +
  39 +2.) make
  40 +
  41 +See individual source file to determine what the examples do.
  42 +
23 43 Inclusion of GF-Complete:
24 44  
25 45 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
138 158 for a backing GF (uses galois_init_field and galois_init_composite_field to get gf_t
139 159 pointer)
140 160  
  161 +Performance:
  162 +
  163 +There are two performance-based test scripts: time_all_gfs_argv_init.sh and
  164 +time_all_gfs_hard_init.sh. Both scripts run the same tests, but initialize the
  165 +underlying GF fields in different ways (*argv* uses reed_sol_time_gf and *hard*
  166 +uses reed_sol_hard_time_gf.c).
  167 +
  168 +You can run 'time_all_gfs_argv_init.sh' to time *all* possible GF
  169 +implementations on your computer. This script requires the 'gf_methods'
  170 +utility from GF-Complete to be in your PATH.
  171 +
  172 +time_all_gfs_argv_init.sh was run on a MacBook Air and the distilled numbers
  173 +are given in ./PERF.txt. The results are sorted by encoding throughput. The
  174 +format of each entry reflects the arguments given to reed_sol_time_gf. For example,
  175 +the test run of
  176 +
  177 +'Examples/reed_sol_time_gf 12 3 8 128 65536 -m SPLIT 8 4 -r SSE'
  178 +
  179 +is recorded as this in PERF.txt:
  180 +
  181 +_12_3_8_128_65536_-m_SPLIT_8_4_-r_SSE_- 2813.34
... ...
README.txt
... ... @@ -20,6 +20,26 @@ The Examples directory contains the example programs.
20 20  
21 21 The makefile assumes that Examples is a subdirectory of the home directory.
22 22  
  23 +Installing:
  24 +
  25 +1.) Install GF-Complete
  26 +
  27 +2.) 'cd' to root directory of Jerasure
  28 +
  29 +3.) Set PREFIX in ./makefile to appropriate directory (/usr/local by default)
  30 +
  31 +4.) make
  32 +
  33 +5.) make install
  34 +
  35 +Running tests and examples:
  36 +
  37 +1.) 'cd' into Examples
  38 +
  39 +2.) make
  40 +
  41 +See individual source file to determine what the examples do.
  42 +
23 43 Inclusion of GF-Complete:
24 44  
25 45 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
138 158 for a backing GF (uses galois_init_field and galois_init_composite_field to get gf_t
139 159 pointer)
140 160  
  161 +Performance:
  162 +
  163 +There are two performance-based test scripts: time_all_gfs_argv_init.sh and
  164 +time_all_gfs_hard_init.sh. Both scripts run the same tests, but initialize the
  165 +underlying GF fields in different ways (*argv* uses reed_sol_time_gf and *hard*
  166 +uses reed_sol_hard_time_gf.c).
  167 +
  168 +You can run 'time_all_gfs_argv_init.sh' to time *all* possible GF
  169 +implementations on your computer. This script requires the 'gf_methods'
  170 +utility from GF-Complete to be in your PATH.
  171 +
  172 +time_all_gfs_argv_init.sh was run on a MacBook Air and the distilled numbers
  173 +are given in ./PERF.txt. The results are sorted by encoding throughput. The
  174 +format of each entry reflects the arguments given to reed_sol_time_gf. For example,
  175 +the test run of
  176 +
  177 +'Examples/reed_sol_time_gf 12 3 8 128 65536 -m SPLIT 8 4 -r SSE'
  178 +
  179 +is recorded as this in PERF.txt:
  180 +
  181 +_12_3_8_128_65536_-m_SPLIT_8_4_-r_SSE_- 2813.34
... ...