reikdas / SABLE

1 stars 0 forks source link

Same dense blocks for different percentage of zeros per block #31

Closed adhithadias closed 7 months ago

adhithadias commented 7 months ago

The same blocks for different numbers of zero percentages. Output from running the shell script. For equal values of r 1000, 1000, 50, 50 and number of dense blocks. We get the same dense block pattern.

Namespace(num_rows=1000, num_cols=1000, partition_type=<PartitionType.uniform: 'uniform'>, row_split=50, col_split=50, percentage_of_blocks=20, percentage_of_zeros=40, operation=<Operation.vbr: 'vbr'>) Generating VBR matrix with 1000 rows, 1000 columns, 50 row splits, 50 column splits, 500 dense blocks, 40% zeros, uniform partitioning, 500 dense blocks. dense blocks = [0, 4, 6, 9, 13, 32, 36, 39, 41, 50, 54, 57, 58, 60, 62, 66, 72, 76, 83, 88, 89, 93, 96, 109, 113, 123, 129, 135, 137, 145, 146, 147, 149, 154, 158, 159, 161, 165, 166, 167, 168, 172, 173, 185, 187, 191, 197, 204, 207, 215, 227, 237, 250, 254, 255, 257, 261, 268, 269, 271, 282, 286, 290, 294, 302, 303, 310, 312, 323, 328, 329, 330, 333, 337, 341, 342, 346, 356, 367, 371, 373, 374, 375, 382, 383, 388, 399, 400, 401, 404, 407, 410, 412, 413, 415, 428, 429, 445, 446, 470, 471, 472, 474, 475, 479, 480, 487, 490, 507, 509, 511, 521, 531, 533, 534, 542, 546, 552, 567, 570, 572, 577, 583, 599, 601, 607, 609, 612, 618, 619, 628, 632, 637, 638, 645, 648, 655, 658, 663, 670, 671, 678, 682, 693, 698, 709, 716, 722, 730, 735, 738, 745, 746, 753, 757, 764, 770, 772, 775, 779, 780, 782, 785, 786, 787, 791, 797, 810, 811, 813, 816, 826, 827, 832, 833, 837, 842, 844, 848, 851, 862, 874, 881, 884, 889, 894, 896, 901, 903, 904, 905, 908, 913, 921, 939, 954, 956, 964, 977, 978, 980, 981, 983, 991, 995, 999, 1004, 1007, 1026, 1027, 1042, 1050, 1052, 1060, 1061, 1063, 1064, 1065, 1066, 1068, 1073, 1104, 1110, 1111, 1124, 1125, 1127, 1128, 1140, 1146, 1150, 1154, 1175, 1178, 1179, 1185, 1186, 1187, 1189, 1192, 1201, 1205, 1225, 1230, 1232, 1233, 1234, 1236, 1238, 1239, 1240, 1242, 1249, 1252, 1261, 1264, 1270, 1275, 1278, 1289, 1290, 1295, 1298, 1303, 1310, 1318, 1320, 1328, 1329, 1332, 1335, 1345, 1352, 1354, 1362, 1364, 1365, 1368, 1369, 1372, 1373, 1375, 1379, 1380, 1382, 1385, 1393, 1402, 1407, 1421, 1430, 1433, 1434, 1435, 1442, 1449, 1462, 1463, 1466, 1468, 1471, 1478, 1481, 1502, 1504, 1508, 1516, 1520, 1523, 1525, 1538, 1548, 1550, 1552, 1565, 1571, 1573, 1576, 1577, 1589, 1590, 1591, 1592, 1593, 1602, 1604, 1613, 1617, 1624, 1630, 1633, 1645, 1647, 1651, 1658, 1659, 1662, 1671, 1678, 1697, 1698, 1700, 1710, 1711, 1713, 1717, 1722, 1732, 1733, 1743, 1745, 1746, 1747, 1766, 1768, 1778, 1783, 1786, 1804, 1811, 1813, 1822, 1824, 1828, 1834, 1835, 1837, 1845, 1849, 1855, 1875, 1877, 1878, 1887, 1912, 1913, 1926, 1931, 1933, 1941, 1948, 1951, 1952, 1953, 1958, 1962, 1970, 1971, 1973, 1975, 1990, 1992, 1995, 1998, 2001, 2004, 2017, 2020, 2021, 2025, 2029, 2036, 2038, 2041, 2042, 2051, 2057, 2062, 2067, 2068, 2071, 2074, 2080, 2083, 2094, 2095, 2098, 2118, 2125, 2133, 2135, 2137, 2140, 2142, 2148, 2151, 2160, 2161, 2164, 2165, 2168, 2171, 2181, 2182, 2188, 2203, 2213, 2214, 2222, 2224, 2226, 2227, 2229, 2232, 2236, 2241, 2242, 2246, 2247, 2249, 2257, 2261, 2263, 2270, 2272, 2278, 2292, 2295, 2296, 2315, 2324, 2328, 2333, 2334, 2335, 2340, 2343, 2350, 2354, 2356, 2357, 2358, 2359, 2364, 2367, 2371, 2374, 2377, 2378, 2382, 2388, 2389, 2403, 2406, 2407, 2408, 2411, 2416, 2425, 2427, 2433, 2436, 2456, 2461, 2462, 2465, 2469, 2472, 2478, 2482] Namespace(num_rows=1000, num_cols=1000, partition_type=<PartitionType.uniform: 'uniform'>, row_split=50, col_split=50, percentage_of_blocks=20, percentage_of_zeros=10, operation=<Operation.vbr: 'vbr'>) Generating VBR matrix with 1000 rows, 1000 columns, 50 row splits, 50 column splits, 500 dense blocks, 10% zeros, uniform partitioning, 500 dense blocks. dense blocks = [0, 4, 6, 9, 13, 32, 36, 39, 41, 50, 54, 57, 58, 60, 62, 66, 72, 76, 83, 88, 89, 93, 96, 109, 113, 123, 129, 135, 137, 145, 146, 147, 149, 154, 158, 159, 161, 165, 166, 167, 168, 172, 173, 185, 187, 191, 197, 204, 207, 215, 227, 237, 250, 254, 255, 257, 261, 268, 269, 271, 282, 286, 290, 294, 302, 303, 310, 312, 323, 328, 329, 330, 333, 337, 341, 342, 346, 356, 367, 371, 373, 374, 375, 382, 383, 388, 399, 400, 401, 404, 407, 410, 412, 413, 415, 428, 429, 445, 446, 470, 471, 472, 474, 475, 479, 480, 487, 490, 507, 509, 511, 521, 531, 533, 534, 542, 546, 552, 567, 570, 572, 577, 583, 599, 601, 607, 609, 612, 618, 619, 628, 632, 637, 638, 645, 648, 655, 658, 663, 670, 671, 678, 682, 693, 698, 709, 716, 722, 730, 735, 738, 745, 746, 753, 757, 764, 770, 772, 775, 779, 780, 782, 785, 786, 787, 791, 797, 810, 811, 813, 816, 826, 827, 832, 833, 837, 842, 844, 848, 851, 862, 874, 881, 884, 889, 894, 896, 901, 903, 904, 905, 908, 913, 921, 939, 954, 956, 964, 977, 978, 980, 981, 983, 991, 995, 999, 1004, 1007, 1026, 1027, 1042, 1050, 1052, 1060, 1061, 1063, 1064, 1065, 1066, 1068, 1073, 1104, 1110, 1111, 1124, 1125, 1127, 1128, 1140, 1146, 1150, 1154, 1175, 1178, 1179, 1185, 1186, 1187, 1189, 1192, 1201, 1205, 1225, 1230, 1232, 1233, 1234, 1236, 1238, 1239, 1240, 1242, 1249, 1252, 1261, 1264, 1270, 1275, 1278, 1289, 1290, 1295, 1298, 1303, 1310, 1318, 1320, 1328, 1329, 1332, 1335, 1345, 1352, 1354, 1362, 1364, 1365, 1368, 1369, 1372, 1373, 1375, 1379, 1380, 1382, 1385, 1393, 1402, 1407, 1421, 1430, 1433, 1434, 1435, 1442, 1449, 1462, 1463, 1466, 1468, 1471, 1478, 1481, 1502, 1504, 1508, 1516, 1520, 1523, 1525, 1538, 1548, 1550, 1552, 1565, 1571, 1573, 1576, 1577, 1589, 1590, 1591, 1592, 1593, 1602, 1604, 1613, 1617, 1624, 1630, 1633, 1645, 1647, 1651, 1658, 1659, 1662, 1671, 1678, 1697, 1698, 1700, 1710, 1711, 1713, 1717, 1722, 1732, 1733, 1743, 1745, 1746, 1747, 1766, 1768, 1778, 1783, 1786, 1804, 1811, 1813, 1822, 1824, 1828, 1834, 1835, 1837, 1845, 1849, 1855, 1875, 1877, 1878, 1887, 1912, 1913, 1926, 1931, 1933, 1941, 1948, 1951, 1952, 1953, 1958, 1962, 1970, 1971, 1973, 1975, 1990, 1992, 1995, 1998, 2001, 2004, 2017, 2020, 2021, 2025, 2029, 2036, 2038, 2041, 2042, 2051, 2057, 2062, 2067, 2068, 2071, 2074, 2080, 2083, 2094, 2095, 2098, 2118, 2125, 2133, 2135, 2137, 2140, 2142, 2148, 2151, 2160, 2161, 2164, 2165, 2168, 2171, 2181, 2182, 2188, 2203, 2213, 2214, 2222, 2224, 2226, 2227, 2229, 2232, 2236, 2241, 2242, 2246, 2247, 2249, 2257, 2261, 2263, 2270, 2272, 2278, 2292, 2295, 2296, 2315, 2324, 2328, 2333, 2334, 2335, 2340, 2343, 2350, 2354, 2356, 2357, 2358, 2359, 2364, 2367, 2371, 2374, 2377, 2378, 2382, 2388, 2389, 2403, 2406, 2407, 2408, 2411, 2416, 2425, 2427, 2433, 2436, 2456, 2461, 2462, 2465, 2469, 2472, 2478, 2482]

ulysses4ever commented 7 months ago

Sorry, what is this PR?

adhithadias commented 7 months ago

Sorry, what is this PR?

It is supposed to be a replacement for https://github.com/reikdas/VBR-SpMV/pull/28/ Pratyush is saying that this is different from https://github.com/reikdas/VBR-SpMV/pull/28/. I will get clarification from him tomorrow when we meet to discuss the PLDI SpMM code repo.

ulysses4ever commented 7 months ago

It is supposed to be a replacement for https://github.com/reikdas/VBR-SpMV/pull/28

Cool. Just not sure why the commit with GNU parallel got here ..

adhithadias commented 7 months ago

It is supposed to be a replacement for #28

Cool. Just not sure why the commit with GNU parallel got here ..

Oh, those are just previous commits on the branch. Ideally, I should have created a new branch. But I just pushed the change to the same branch because it's just a one-line change.

ulysses4ever commented 7 months ago

Now you know that saving yourself some seconds on creating a new branch can leave people confused 🙃