Commit Graph
3885 Commits
Author SHA1 Message Date
Mateusz Pusz e1b0fb7e6e refactor: first_100_primes refactored to use a varaible template for caching the result 2024-11-16 22:00:21 +01:00
Mateusz Pusz 658c0950f8 refactor: dimension_symbol and unit_symbol refactored to use a variable template for caching result 2024-11-16 21:59:40 +01:00
Mateusz Pusz daf4b1cc5c refactor: SymbolicConstant concept refactored 2024-11-16 11:53:05 +01:00
Mateusz Pusz 7b64b4b650 refactor: 💥 from now unit_symbol and dimension_symbol always returns std::string_view 2024-11-16 09:57:53 +01:00
Mateusz Pusz 2c24c61da5 perf: first_100_primes moved to the static storage 2024-11-16 09:11:52 +01:00
Mateusz Pusz 4f8e959b5c fix(test): shadowing warnings on gcc fixed 2024-11-16 08:28:39 +01:00
Mateusz Pusz 4aaebb9426 Merge branch 'master' of github.com:mpusz/units 2024-11-16 08:07:33 +01:00
Mateusz Pusz 9d6e195ed8 Merge pull request #641 from chiphogg/chiphogg/baillie-psw#509
Use Baillie-PSW prime test to factor numbers
2024-11-15 22:50:25 +01:00
Chip Hogg a93a5d6d2e Add std:: prefix 2024-11-15 13:38:48 -05:00
Chip Hogg de68b3a9f1 Replace old factoring with Baillie-PSW 2024-11-15 13:30:05 -05:00
Chip Hogg 3b586a685f Implement Baillie-PSW 2024-11-15 13:30:05 -05:00
Mateusz Pusz ec914986d1 Merge pull request #640 from chiphogg/chiphogg/strong-lucas#509
Add the Strong Lucas Probable Prime test
2024-11-15 19:28:16 +01:00
Chip Hogg 82040c1788 Handle edge case
Subtracting from `n` doesn't work in literally every case: it fails for
`0` input.
2024-11-15 12:58:22 -05:00
Chip Hogg d963ce8f93 Apply mod to D
Looks like we were violating a precondition.
2024-11-15 11:58:49 -05:00
Chip Hogg 346a001283 Add more missing std:: prefixes 2024-11-15 10:56:53 -05:00
Chip Hogg 9e8dfec265 Add the Strong Lucas Probable Prime test
This is more involved than the Miller-Rabin test, but we can tame the
complexity by breaking it down into helper functions, performing tasks
such as:

- Increment the index of the (U_k, V_k) sequence elements by one.
- Double the index of the (U_k, V_k) sequence elements.
- Find an appropriate D parameter.

etc.

With these helpers, the algorithm becomes straightforward (see, for
instance,
https://en.wikipedia.org/wiki/Lucas_pseudoprime#Strong_Lucas_pseudoprimes).
We start by ruling out perfect squares (because if we don't, then the
search for `D` will never terminate).  Then we find our `D`, and
decompose `n + 1` into `s` and `d` parameters (exactly as we did for
Miller-Rabin, except there we used `n - 1`).  At this point, the strong
test is easy: check whether `U_d` is 0, then check `V_d`, as well as `V`
for all successive doublings of the index less than `s`.

A similar testing strategy as for the Miller Rabin gives us sufficient
confidence.

1. Test that we get small primes right.
2. Test that we get known pseudoprimes "correctly wrong".
3. Test some really big primes.

(Remember, a probable prime test must mark every actual prime as
"probably prime".)

Helps #509.
2024-11-15 10:53:26 -05:00
Mateusz Pusz c901b73ee5 Merge pull request #639 from chiphogg/chiphogg/jacobi#509
Add utilities to make Strong Lucas tests easier
2024-11-15 15:57:09 +01:00
Chip Hogg 26639dab34 Add some more 2024-11-15 09:39:39 -05:00
Chip Hogg b49a4b711b Add missing std:: prefixes 2024-11-15 09:17:40 -05:00
Mateusz Pusz 08e94171cb test: vector quantities unit tests addded 2024-11-15 12:59:18 +01:00
Chip Hogg d56ffc08b8 Add utilities to make Strong Lucas tests easier
The Strong Lucas test coming in the next PR will already be complicated
enough.  It'll be convenient, and less distracting, if we already have
functions for certain operations we'll need.

One thing we'll need to do is detect inputs that are perfect squares.
Fortunately, this is pretty easy to do robustly and quickly, with
Newton's method.  We _don't_ want to use `std::sqrt`, because that takes
us into the floating point domain for no good reason, which could give
us wrong answers for larger integers.

The other thing we need is Jacobi symbols.  These are a lot more
obscure, but thankfully, still resonably straightforward to compute.
The Wikipedia page (https://en.wikipedia.org/wiki/Jacobi_symbol) has a
good explanation, and in particular, good instructions for computing
values.

With these utilities in place, the Strong Lucas code should be easier to
review.
2024-11-14 20:14:20 -05:00
Mateusz Pusz 1581fa59ad test: extend concepts tests with cartesian_vector as a representation 2024-11-14 20:42:21 +01:00
Mateusz Pusz 2cf64f20c5 fix: cartesian_vector convertibility tests fixed 2024-11-14 20:40:08 +01:00
Mateusz Pusz a7cf015a3e feat: add compound assignment operators to cartesian_vector 2024-11-14 20:38:50 +01:00
Mateusz Pusz 4b5d37e83e feat: constructors of cartesian_vector refactored 2024-11-14 20:30:12 +01:00
Mateusz Pusz 7e1baf65fa Merge pull request #638 from chiphogg/chiphogg/miller-rabin#509
Add Miller-Rabin probable prime test
2024-11-14 08:27:31 +01:00
Chip Hogg cfd9ddb675 Add Miller-Rabin probable prime test
This can mark a number as either "probably prime", or "definitely
composite".  The first parameter is the base, and the second is the
number to test.

Future PRs will build up the Strong Lucas test which complements this,
and then form the Baillie-PSW test by composing the two.

Helps #506.
2024-11-13 19:56:05 -05:00
Mateusz Pusz e4bccad75c ci: names of jobs reordered to better reflect a new matrix sorting algorithm 2024-11-13 23:09:00 +01:00
Mateusz Pusz f5b9288825 fix: MP_UNITS_NONCONST_TYPE disabled for gcc-12 2024-11-13 22:48:19 +01:00
Mateusz Pusz 2590aa4b80 refactor: MP_UNITS_NONCONST_TYPE introduced to benefit from the C++23 feature 2024-11-13 21:13:22 +01:00
Mateusz Pusz 5cd07bc621 Merge pull request #634 from chiphogg/chiphogg/mod#509
Add helpers for modular arithmetic
2024-11-13 20:32:28 +01:00
Chip Hogg b99faf051b Merge branch 'master' into chiphogg/mod#509 2024-11-13 08:41:52 -05:00
Mateusz Pusz 9dd59e845c ci: import_std now checks if at least C++23 is being used 2024-11-13 12:34:13 +01:00
Mateusz Pusz 6bcb46cc4e Merge pull request #635 from burnpanck/feature/faster-CI
Faster CI: Sparse run matrix, cancel previous jobs on the same branch
2024-11-13 07:26:47 +01:00
Yves Delley b35e241a98 fix clang on macos-14 (arm64) 2024-11-12 22:03:43 +01:00
Yves Delley ff118784fa exclude clang18+debug from freestanding again 2024-11-12 21:48:59 +01:00
Yves Delley 5713243d07 fix clang 16 2024-11-12 21:41:09 +01:00
Yves Delley 01f44c66fa heed github warning, and use output file instead of set-output command; also, fix freestanding 2024-11-12 21:29:42 +01:00
Yves Delley f4c8e901ac fix boolean conan options in dynamic CI matrix 2024-11-12 21:18:36 +01:00
Yves Delley a51462cc65 missing checkout in generate-matrix step 2024-11-12 19:09:12 +01:00
Yves Delley cc9ea9dd1a add cancel-in-progress to all workflows 2024-11-12 19:05:04 +01:00
Yves Delley e464677200 don't test Clang 19 just yet; fix cancel-in-progres 2024-11-12 19:02:21 +01:00
Yves Delley 7fa15d224d fix formatting 2024-11-12 18:43:50 +01:00
Yves Delley 329b9f585c Merge branch 'master' into feature/faster-CI 2024-11-12 18:20:20 +01:00
Mateusz Pusz 420ffef6c5 fix(test): missing header files added 2024-11-12 18:12:36 +01:00
Yves Delley 35ed472975 first attempt at generating sparse CI run matrix in python; also, cancel previous runs on the same branch 2024-11-12 18:07:52 +01:00
Mateusz Pusz 1ee824423e refactor: usage of v for the function argument name changed to prevent shadowing errors 2024-11-12 16:22:14 +01:00
Mateusz Pusz 73ad1f08d4 fix: fmt_test fixed to use delta to create a quantity of thermodynamic_temperature 2024-11-12 16:20:51 +01:00
Mateusz Pusz 520610ab2f test: cartesian_vector used in cgs_test 2024-11-12 14:39:04 +01:00
Mateusz Pusz 5a206c3ef1 test: cartesian_vector used in fmt_test 2024-11-12 14:38:51 +01:00