Findafactor

This release (only) fixes a bug in parallel Gaussian elimination.

**Full Changelog**: https://github.com/vm6502q/FindAFactor/compare/v2.1.1...v2.1.2

sha1sum results:
a2ff3c44bff1e8a87255d39b6a2aba30f8cb74f8 FindAFactor-2.1.2-cp310-cp310-manylinux_2_35_x86_64.whl
38c84fc2fa012e743763f895e9d22abe25099b60 FindAFactor-2.1.2-cp312-cp312-manylinux_2_39_x86_64.whl
a4ca600f969f88932753214fdf15f6e7bb827c61 FindAFactor-2.1.2-cp312-cp312-win_amd64.whl
9216f3c02738b265bf7cff3db8c774a6c8ea998c FindAFactor-2.1.2-cp313-cp313-macosx_13_0_x86_64.whl
aaa3f9a6e557ca07de89d14e0cac4a91c91ccd6a FindAFactor-2.1.2-cp313-cp313-macosx_14_0_arm64.whl
dc07c152b3c3d2b3c7cb28cd2b9e8d94ec43ce8a FindAFactor-2.1.2-cp313-cp313-macosx_15_0_arm64.whl
3a255f2d6788db3af92c68e882ade99fe5525441 FindAFactor-2.1.2-cp38-cp38-manylinux_2_31_x86_64.whl
b87977c04f9e2a32682c30877a6ee4fdb0f069e9 findafactor-2.1.2.tar.gz

A single unified Gaussian elimination matrix is used _per node,_ now, instead of per thread. This significantly reduces overall memory usage and makes `use_congruence_of_squares` more likely to find a congruence in fewer Gaussian elimination iterations at scale.

**Full Changelog**: https://github.com/vm6502q/FindAFactor/compare/v2.1.0...v2.1.1

sha1sum results:
a02c31685f56a9e48b7579b0e9a4dd260d849391 FindAFactor-2.1.1-cp310-cp310-manylinux_2_35_x86_64.whl
932b467eb8db104e77244dc3d2bb2047da999a45 FindAFactor-2.1.1-cp312-cp312-manylinux_2_39_x86_64.whl
4f7d128546294ea0c6f4a4955c4643b840488b02 FindAFactor-2.1.1-cp312-cp312-win_amd64.whl
c0daa613538df962f20b8e6a46a404d5143ef6c0 FindAFactor-2.1.1-cp313-cp313-macosx_13_0_x86_64.whl
3005c17b31285a6d3b7f9bb0585c78b585e33924 FindAFactor-2.1.1-cp313-cp313-macosx_14_0_arm64.whl
141f0bf9726737104e3367722175b25e8f90a5fc FindAFactor-2.1.1-cp313-cp313-macosx_15_0_arm64.whl
b9a73ec9192acb46af03fb8069db4e4c85b861ec FindAFactor-2.1.1-cp38-cp38-manylinux_2_31_x86_64.whl
98adf4ffd052622333569baeb1825772244408e1 findafactor-2.1.1.tar.gz

v.3.0.0
This release removes the `thread_count` option (because it's difficult to respect this option throughout the algorithm, while it has relatively applicability to memory management) and parallelizes Gaussian elimination to about the maximum feasible extent. (Realize that `batch_size_multiple` might need significant user experimentation and tuning.)

**Full Changelog**: https://github.com/vm6502q/FindAFactor/compare/v2.1.2...v.3.0.0

sha1sum results:
2faaccb1a857943e7d75015254db14326277f501 FindAFactor-3.0.0-cp310-cp310-manylinux_2_35_x86_64.whl
91014d3cde3311ed512553d5b7791192c54bb3f1 FindAFactor-3.0.0-cp312-cp312-manylinux_2_39_x86_64.whl
d6f1bfc21ad2183c9c4297a9463169aa1c691f12 FindAFactor-3.0.0-cp312-cp312-win_amd64.whl
05a12745f533424e24a6d2b6027fc9076ec545f0 FindAFactor-3.0.0-cp313-cp313-macosx_13_0_x86_64.whl
64b171e2e68ff4205e244b0b4e3be808e185c602 FindAFactor-3.0.0-cp313-cp313-macosx_14_0_arm64.whl
376d6258a8058519da7cc167ab9ca0a59c0db956 FindAFactor-3.0.0-cp313-cp313-macosx_15_0_arm64.whl
cd71a42060e3f3e00931c4bb9cc55b6ec1270398 FindAFactor-3.0.0-cp38-cp38-manylinux_2_31_x86_64.whl
0e5438943ca1c9c2dc2fb641af184b8e6ad02459 findafactor-3.0.0.tar.gz

batch_size_multiple` is a potential tunable parameter contingency that should be available. However, as for the semantic version change, it seems to make far more sense for it be a multiple of (2 times) the count of surviving entries on the largest factorization wheel (up to gear level). Wheel batches alternate between opposite ends of the iteration range, so 2 wheel entry batches is considered `batch_size_multiplier=1.0`.

**Full Changelog**: https://github.com/vm6502q/FindAFactor/compare/v2.0.0...v2.1.0

sha1sum results:
ba6f9167f34b3cbb25650c86eca8ac855827b76b FindAFactor-2.1.0-cp310-cp310-manylinux_2_35_x86_64.whl
ed4941bc8be3a550625d81a43816e2bbceadab96 FindAFactor-2.1.0-cp312-cp312-manylinux_2_39_x86_64.whl
43ad4e9fac9dd31de0cfffe9a2d91036dcd781a7 FindAFactor-2.1.0-cp312-cp312-win_amd64.whl
81adfbbcb8c8740b1f3a23dfdfa0f1725e0d9032 FindAFactor-2.1.0-cp313-cp313-macosx_13_0_x86_64.whl
9a4b07889630e360a72535fdd2102059e578d503 FindAFactor-2.1.0-cp313-cp313-macosx_14_0_arm64.whl
78f5bb375ee7ccef960cfe9f7a28b604d352c0fe FindAFactor-2.1.0-cp313-cp313-macosx_15_0_arm64.whl
db39659bff0bfe69027d16a53d37ab4d86216e25 FindAFactor-2.1.0-cp38-cp38-manylinux_2_31_x86_64.whl
9038fa9407ac6d67dce9df9664939bb87a2f6760 findafactor-2.1.0.tar.gz

This release drops the `batch_size` argument as unnecessary while debugging and improving Gaussian elimination. (A simple 56 bit test with `use_congruence_of_squares` is not slow.)

**Full Changelog**: https://github.com/vm6502q/FindAFactor/compare/v1.11.2...v2.0.0

sha1sum results:
442efe4ea94a1ebb4a03df4759edfdb333823903 FindAFactor-2.0.0-cp310-cp310-manylinux_2_35_x86_64.whl
1f8e7084d17feac776e0217656865977b7b11a9f FindAFactor-2.0.0-cp312-cp312-manylinux_2_39_x86_64.whl
d429d1700a0131109bb2c3a1be2089bcf344addd FindAFactor-2.0.0-cp312-cp312-win_amd64.whl
95ef42323119418589b338d9361914652096991a FindAFactor-2.0.0-cp313-cp313-macosx_13_0_x86_64.whl
1bccf6d74d31c923ea7cd01bfbc1ad166299b976 FindAFactor-2.0.0-cp313-cp313-macosx_14_0_arm64.whl
3ce493b03a5ac441e320921b0c76f8efe40bd3c8 FindAFactor-2.0.0-cp313-cp313-macosx_15_0_arm64.whl
b778a8cdc3889019533fea9600a22909e2c1dc76 FindAFactor-2.0.0-cp38-cp38-manylinux_2_31_x86_64.whl
140e37776e84a3f39d9047536b075df981eae9b3 findafactor-2.0.0.tar.gz

This release adds Mac and Windows support. (Windows users might _still_ prefer WSL for performance.)

**Full Changelog**: https://github.com/vm6502q/FindAFactor/compare/v1.11.1...v1.11.2

sha1sum results:
fc81597d04ae1db394586bc0c159b307e678457f FindAFactor-1.11.2-cp310-cp310-macosx_13_0_x86_64.whl
82f6c8759c7f84080bff6e5fb5af29b51115b204 FindAFactor-1.11.2-cp310-cp310-macosx_14_0_arm64.whl
ae6094ed44b9824fddea1074eba5bf912e76e88d FindAFactor-1.11.2-cp310-cp310-macosx_15_0_arm64.whl
cb33bf0b08aab73474a82e30e43b99aa2e3dd566 FindAFactor-1.11.2-cp310-cp310-manylinux_2_35_x86_64.whl
690c61ffc0c907332600ddb080e8e1a7a22aaf2b FindAFactor-1.11.2-cp312-cp312-manylinux_2_39_x86_64.whl
968e5ddac7c22c0e00d1874db094270f0a7a1ec3 FindAFactor-1.11.2-cp312-cp312-win_amd64.whl
c46c18840ed8fd55e98e4c585bc629ffe65723a7 FindAFactor-1.11.2-cp38-cp38-manylinux_2_31_x86_64.whl
ae08de8183bdb99fbb811132861b47af4dbed831 findafactor-1.11.2.tar.gz

This release fixes `getNextAltBatch()` and uses only this increment in general, between both congruence of squares and brute force.

**Full Changelog**: https://github.com/vm6502q/FindAFactor/compare/v1.11.0...v1.11.1

sha1sum results:
89dc69688b23c57b2581e8b5f7b407672de4da5a FindAFactor-1.11.1-cp310-cp310-manylinux_2_35_x86_64.whl
4e55973b9c533eb1f220292874e4860d2351afee FindAFactor-1.11.1-cp312-cp312-manylinux_2_39_x86_64.whl
9647f8a77f84c77aebdd06299fd8c2697a0e4f3f FindAFactor-1.11.1-cp38-cp38-manylinux_2_31_x86_64.whl
1095d7aedc01b52b7152e1c7fa78be3f4b01d8fc findafactor-1.11.1.tar.gz