Message326332
> When testing what, specifically? And the standard 32-bit FNV multiplier, or the standard 64-bit FNV multiplier?
FNV-1a with the t ^= 2 * t mangling running my new testsuite on either PR 9471 or PR 9534 using the 64-bit FNV multiplier to produce 64-bit hashes. In other words, the code from PR 9534 but just changing the multiplier.
On the full 64-bit range, I got 2 collisions (where statistically 0 would be expected). When truncated to 32-bits, I got about 1700 collisions (where statistically about 15 would be expected). |
|
Date |
User |
Action |
Args |
2018-09-25 08:42:03 | jdemeyer | set | recipients:
+ jdemeyer, tim.peters, rhettinger, mark.dickinson, eric.smith, sir-sigurd |
2018-09-25 08:42:03 | jdemeyer | set | messageid: <1537864923.96.0.545547206417.issue34751@psf.upfronthosting.co.za> |
2018-09-25 08:42:03 | jdemeyer | link | issue34751 messages |
2018-09-25 08:42:03 | jdemeyer | create | |
|