Keys: WordData.txt 111444 strings Hash fn load avg worst % < log N --------------------------------------------------- elfhash 1.0 1.5416 9 100% 1.2 1.4554 (1.4240) 8 (7) 100% 1.4 1.3630 (1.3682) 8 (7) 100% 1.6 1.3228 7 100% 1.8 1.2822 7 100% 2.0 1.3215 7 100% --------------------------------------------------- sfold 1.0 1.5078 7 100% 1.2 1.4578 (1.4225) 9 (7) 100% 1.4 1.3637 (1.3698) 7 100% 1.6 1.3195 7 100% 1.8 1.2881 6 100% 2.0 1.2569 6 100% --------------------------------------------------- FNV 1.0 1.4995 7 100% 1.2 1.4172 (1.4147) 7 100% 1.4 1.3584 (1.3566) 7 (6) 100% 1.6 1.3137 6 100% 1.8 1.2817 6 100% 2.0 1.2484 5 100% --------------------------------------------------- DEK 1.0 1.5046 8 100% 1.2 1.4213 (1.4156) 8 (7) 100% 1.4 1.3544 (1.3568) 7 (8) 100% 1.6 1.3122 6 100% 1.8 1.2784 8 100% 2.0 1.2534 7 100% Keys: CA_FeatureNames.txt 97497 strings Hash fn load avg worst % < log N --------------------------------------------------- elfhash 1.0 1.5054 8 100% 1.2 1.4486 7 100% 1.4 1.3564 (1.5534) 6 (8) 100% 1.6 1.3120 6 100% 1.8 1.2795 (1.2794) 6 100% 2.0 1.2524 6 100% --------------------------------------------------- sfold 1.0 1.5033 8 100% 1.2 1.4137 7 100% 1.4 1.3594 (1.3579) 6 (7) 100% 1.6 1.3105 7 100% 1.8 1.2773 (1.2766) 6 100% 2.0 1.2503 7 100% --------------------------------------------------- FNV 1.0 1.4985 7 100% 1.2 1.4161 7 100% 1.4 1.3553 (1.3570) 6 100% 1.6 1.3118 6 100% 1.8 1.2760 (1.2759) 6 100% 2.0 1.2482 6 100% --------------------------------------------------- DEK 1.0 1.4975 8 100% 1.2 1.4140 7 100% 1.4 1.3590 (1.3588) 8 (7) 100% 1.6 1.3115 6 100% 1.8 1.2752 (1.2777) 6 (7) 100% 2.0 1.2509 6 100% I computed my solution using integer truncation when I calculated the table size; the values in parentheses are from a second run using rounding when computing the table size. Either would be acceptable.