Reznov

Array#lbound and Array#lbound_index

1

I think that ```ruby array.lbound x ``` is easier to use than ```ruby array.bsearch { _1 >= x } ```

for -e 1: group tiles by type, then decode using SIMD

1

Maybe other JXL-guys can answer? https://github.com/veluca93/fpnge/issues/26

Tested: https://atcoder.jp/contests/practice2/submissions/45976337

[CRBTreeMap] Segmentation fault at Leetcode #1429

4

Only 1 line is different: OK: https://leetcode.com/submissions/detail/1075862918/ `@t.delete p if @t[p]` Not OK: https://leetcode.com/submissions/detail/1075860617/ `@t.delete p`

BIT O(n) init

1

Tested: https://atcoder.jp/contests/practice2/submissions/45974331

When needed a list of unique divisors (2nd solutinon [here](https://leetcode.com/problems/split-the-array-to-make-coprime-products/solutions/3261267/1984-ms/)), creating (and destructing) a 2D array in `prime_division` is unnecessary. Creating an 1D array should be less expensive.

Performance on .fjxl

2

Test image: https://stsci-opo.org/STScI-01GA76Q01D09HFEV174SVMQDMV.png ``` $ time ./dj40 w.fjxl 14560x8418 frame read and discarded. real 0m16,854s user 0m16,325s sys 0m0,549s ``` 8 MPx/s, PNG decoding speed on the same CPU is...

Edge cases suitable for counting sort

1

Look at this approach: https://github.com/alantudyk/ruby/commit/04831f6536cbf8691c1f4adb82c98e0092b25b5e

SIMD-accelerated Heaps

4

Quick search in your blog doesn't show posts about the subject. 3 hours ago, I got the heap that is 2x faster than `qsort()` as a sorter: https://github.com/alantudyk/simd_x/tree/master/16-ary_heap I think...

`_mm256_bslli_epi128` compiles to the same assembler but the name is more descriptive than `_mm256_slli_si256`: https://github.com/algorithmica-org/algorithmica/blob/ed1945c3d2de8548b1c744d3161eb668703db808/content/english/hpc/algorithms/prefix.md?plain=1#L103