 andrew
|
447da11d59
|
Remove obsolete optimizations
These are a relic of when we used forEachInRange for checkMaxBetween
|
2024-06-29 22:47:42 -07:00 |
|
|
andrew
|
daa8e02d4f
|
Fixes from testing on an avx512f-capable machine
|
2024-06-29 22:41:39 -07:00 |
|
|
andrew
|
fd3ea2c2a8
|
clang-format fixes
|
2024-06-29 22:21:50 -07:00 |
|
|
andrew
|
0b839b9d7e
|
Fixes for symbol multi-versioning with avx512f
|
2024-06-29 22:20:50 -07:00 |
|
|
andrew
|
11a022dcf7
|
Attempt at avx512f 32bit compare
|
2024-06-29 21:56:21 -07:00 |
|
|
andrew
|
94da4c72a5
|
Fix clang-format
|
2024-06-29 15:11:40 -07:00 |
|
|
andrew
|
461e07822a
|
32-bit x86 simd for the other scan16 too
|
2024-06-29 15:10:36 -07:00 |
|
|
andrew
|
75499543e7
|
Fix clang-format
|
2024-06-29 15:03:44 -07:00 |
|
|
andrew
|
81f44d352f
|
SIMD scan16 for x86 + 32-bit versions
|
2024-06-29 15:01:18 -07:00 |
|
|
andrew
|
45da8fb996
|
Use the faster unvectorized implementation for Node3
|
2024-06-28 22:52:39 -07:00 |
|
|
andrew
|
4958a4cced
|
Make always_inline function inline
Try to fix warning in jenkins
|
2024-06-28 22:42:01 -07:00 |
|
|
andrew
|
648b0b9238
|
Add an always_inline, with explanatory comment
|
2024-06-28 19:55:33 -07:00 |
|
|
andrew
|
d3f4afa167
|
More SIMD for scanning Node256 with 32-bit versions
|
2024-06-28 19:48:06 -07:00 |
|
|
andrew
|
f762add4d6
|
Write vectorized 32-bit compare by hand for arm in scan16
|
2024-06-28 17:28:59 -07:00 |
|
|
andrew
|
b311e5f1f0
|
Add an experimental, disabled 32 bit internal version
I think it's only missing detection for full-precision versions more
than 2e9 apart
|
2024-06-28 15:53:35 -07:00 |
|
|
andrew
|
ff81890921
|
Rename MaxVersionT to InternalVersionT
|
2024-06-28 14:34:48 -07:00 |
|
|
andrew
|
0e96177f5c
|
Allow to easily experiment with 32 bit "max version" type
|
2024-06-28 13:54:12 -07:00 |
|
|
andrew
|
efb0e52a0a
|
SIMD implementation of scan16 for x86
Closes #29
|
2024-06-27 22:21:41 -07:00 |
|
|
andrew
|
2df7000090
|
Remove switch on phase from Stepwise left/right step
|
2024-06-27 20:51:35 -07:00 |
|
|
andrew
|
5378a06c39
|
Vectorize all bounds checks for Node256 scan
|
2024-06-27 17:40:23 -07:00 |
|
|
andrew
|
12c6ed2568
|
Reorganize to prepare for better vectorized first/last page
|
2024-06-27 17:21:41 -07:00 |
|
|
andrew
|
c065b185ae
|
Vectorize inner page check for Node256
|
2024-06-27 17:09:45 -07:00 |
|
|
andrew
|
639518bed4
|
Share "scan16" between Node16 and Node48
|
2024-06-27 13:22:51 -07:00 |
|
|
andrew
|
7de983cc15
|
Simd bounds checking for scan for Node16
|
2024-06-27 13:08:12 -07:00 |
|
|
andrew
|
1b4b61ddc6
|
Write Node16 scan in a "more vectorized" style
|
2024-06-27 12:07:38 -07:00 |
|
|
andrew
|
bff7b85de2
|
Remove "Child" struct
|
2024-06-27 10:03:14 -07:00 |
|
|
andrew
|
9108ee209a
|
SoA instead of AoS for child, maxVersion
|
2024-06-27 09:57:54 -07:00 |
|
|
andrew
|
f8bf1c6eb4
|
Remove unreachable code
|
2024-06-26 22:14:27 -07:00 |
|
|
andrew
|
4da2a01614
|
Use single &, to show branch-free intent
|
2024-06-26 22:14:05 -07:00 |
|
|
andrew
|
bb0e654040
|
Fix missed update for Node48::maxOfMax
|
2024-06-26 22:11:33 -07:00 |
|
|
andrew
|
02866a8cae
|
Save some bounds checking for scanning Node256
|
2024-06-26 20:55:18 -07:00 |
|
|
andrew
|
fa86d3e707
|
"max of max" for Node48 again, but physical instead of logical
|
2024-06-26 20:41:27 -07:00 |
|
|
andrew
|
7d1d1d7b2a
|
Maintain childMaxVersion == 0 for unused children in Node48
|
2024-06-26 20:16:50 -07:00 |
|
|
andrew
|
789ecc29b3
|
Use unsigned compare trick to check in bounds
|
2024-06-26 19:41:25 -07:00 |
|
|
andrew
|
08f2998a85
|
Use 8 byte pages for "max of max"
This seems to benchmark better
|
2024-06-26 19:18:38 -07:00 |
|
|
andrew
|
c882d7663d
|
Maintain "reverseIndex" in Node48
|
2024-06-26 19:11:34 -07:00 |
|
|
andrew
|
bfea4384ba
|
Branchless inner page check for Node256
|
2024-06-26 18:28:41 -07:00 |
|
|
andrew
|
6520e3d734
|
"max of max" for Node48
|
2024-06-26 17:54:03 -07:00 |
|
|
andrew
|
22e4ab01a1
|
Track "max of max" versions in Node256
|
2024-06-26 16:28:24 -07:00 |
|
|
andrew
|
b3aeed0caa
|
Warning: interface change! Require versions >= 0
|
2024-06-26 15:46:36 -07:00 |
|
|
andrew
|
5f3833e965
|
Change maxVersion to return by value, and add setMaxVersion
|
2024-06-26 15:33:15 -07:00 |
|
|
andrew
|
8b1cd9c052
|
Minor improvements to checkMaxBetweenExclusive
|
2024-06-26 15:06:50 -07:00 |
|
|
andrew
|
488c723726
|
Improve worst-case radix tree checkRangeRead
|
2024-06-25 21:22:55 -07:00 |
|
|
andrew
|
2eb461b8ea
|
Fix build for llvm 18
|
2024-06-11 11:38:55 -07:00 |
|
|
andrew
|
6e63fd5126
|
Add internal entry points, with test coverage
Closes #25
|
2024-04-19 11:23:25 -07:00 |
|
|
andrew
|
d2e1863593
|
Account for every uncovered line in the implementation
Closes #23
|
2024-04-18 12:43:48 -07:00 |
|
|
andrew
|
c15d296432
|
Exercise copyChildrenAndKeyFrom for Node{48,256} to itself
|
2024-04-17 17:13:22 -07:00 |
|
|
andrew
|
309d315956
|
Add DebugConflictSet, which asserts using skip list as a reference
CC #23
|
2024-04-17 12:08:39 -07:00 |
|
|
andrew
|
85db1a8786
|
Allow to choose implementation in python wrapper
And fix a few minor bugs to make the python tests pass for skip_list.
CC #23
|
2024-04-17 12:08:28 -07:00 |
|
|
andrew
|
fd93300ce8
|
Tweaks and add more planned sections
|
2024-04-16 17:36:29 -07:00 |
|