Add Hypergeometric 1F0 function and gradients #2962
Conversation
andrjohns
changed the title
| * @return Hypergeometric 1F0 function | ||
| */ | ||
| template <typename Ta, typename Tz, require_all_arithmetic_t<Ta, Tz>* = nullptr> | ||
| auto hypergeometric_1f0(const Ta& a, const Tz& z) { |
There was a problem hiding this comment.
Given the type assumptions being made, is Ta and Tz just double and double only?
If I'm reading this properly, this isn't vectorized or anything, is it?
There was a problem hiding this comment.
This is to allow all combinations of int/double (as well as size_t,float, etc.)
Thanks @syclik! Updated to more explicit typing for readability |
|
Thank you!!! |
| * @return Hypergeometric 1F0 function | ||
| */ | ||
| template <typename Ta, typename Tz, require_all_arithmetic_t<Ta, Tz>* = nullptr> | ||
| double hypergeometric_1f0(const Ta& a, const Tz& z) { |
There was a problem hiding this comment.
Is this the (double, double) -> double full template specification? (sorry... notation is tough: Ta = double, Tz = double)
If so, this implementation doesn't make sense as the function template for hypergeomtric_1f0 and we should probably implement this as a normal function. What do you think?
Based on the other implementations, we're not actually taking the route of defining the function template in prim and writing template specialization in each of the folders. We're using SFINAE to knock out template matches instead. So I think that means we don't have a single function template any more.
Am I thinking about this correctly? Thoughts?
(My concrete suggestion: change this function to not have any template parameters.)
There was a problem hiding this comment.
This templating is to implement the signatures:
(int, int) -> double
(int, double) -> double
(double, int) -> double
(double, double) -> double
So removing the template parameters would mean replacing this function definition with four definitions, which seems a little much.
There was a problem hiding this comment.
Since this is the base template, I'd suggest having the return type be return_type<Ta, Tz> instead of double. I'll add the code suggestion right now.
The reason for that is the base template should be a catch-all for everything.
There was a problem hiding this comment.
| double hypergeometric_1f0(const Ta& a, const Tz& z) { | |
| return_type<Ta, Tz> hypergeometric_1f0(const Ta& a, const Tz& z) { |
There was a problem hiding this comment.
Something feels slightly off: the base template has 3 template parameters, but the specializations have 4? Was that intended? (Did I miss the actual base template declaration somewhere?)
There was a problem hiding this comment.
Something feels slightly off: the base template has 3 template parameters, but the specializations have 4? Was that intended? (Did I miss the actual base template declaration somewhere?)
Sorry you've lost me, all of the implementations here only have two parameters?
There was a problem hiding this comment.
Are you counting the require_* elements here? The number of those don't have to be consistent between the implementations
There was a problem hiding this comment.
@syclik small ping
|
@syclik small ping to have a look at the review responses here |
Jenkins Console Log Machine informationNo LSB modules are available. Distributor ID: Ubuntu Description: Ubuntu 20.04.3 LTS Release: 20.04 Codename: focalCPU: G++: Clang: |
Co-authored-by: Daniel Lee <[email protected]>
Jenkins Console Log Machine informationNo LSB modules are available. Distributor ID: Ubuntu Description: Ubuntu 20.04.3 LTS Release: 20.04 Codename: focalCPU: G++: Clang: |
|
Merging! |
Summary
This PR exposes the
hypergeometric_1f0function from Boost and adds gradients. The gradients for the 1F0 are much simpler than the other hypergeometric series and can be expressed analytically (no need for thegrad_pFqfunction)Tests
primtests for values and throwing behaviour added,mixtests for gradients added via the ad testing frameworkSide Effects
N/A
Release notes
Added the Hypergeometric 1F0 function and gradients
Checklist
Math issue #(issue number)
Copyright holder: Andrew Johnson
The copyright holder is typically you or your assignee, such as a university or company. By submitting this pull request, the copyright holder is agreeing to the license the submitted work under the following licenses:
- Code: BSD 3-clause (https://opensource.org/licenses/BSD-3-Clause)
- Documentation: CC-BY 4.0 (https://creativecommons.org/licenses/by/4.0/)
the basic tests are passing
./runTests.py test/unit)make test-headers)make test-math-dependencies)make doxygen)make cpplint)the code is written in idiomatic C++ and changes are documented in the doxygen
the new changes are tested