Allow Separable Growth Functions to compute Isotropized Power Spectra #1216

paulrogozenski · 2024-12-03T18:53:26Z

A large bottleneck in the computation time of covariances is the calculation of isotropized power spectra. The current implementation calculates this at each scale factor considered in the computation. By separating the growth factor from this computation, one can save computation time and maintain accuracy to a good approximation.

The implementation includes optional flags when calculating the 2h_22, 3h, and 4h Trispectrum terms in pk_4pt.py and the corresponding functions for wrappers.

…ions

coveralls · 2024-12-03T19:07:22Z

Pull Request Test Coverage Report for Build 12161562031

Warning: This coverage report may be inaccurate.

This pull request's base commit is no longer the HEAD commit of its target branch. This means it includes changes from outside the original pull request, including, potentially, unrelated coverage changes.

For more information on this, see Tracking coverage changes with pull request builds.
To avoid this issue with future PRs, see these Recommended CI Configurations.
For a quick fix, rebase this PR at GitHub. Your next report should be accurate.

Details

17 of 17 (100.0%) changed or added relevant lines in 1 file are covered.
36 unchanged lines in 7 files lost coverage.
Overall coverage increased (+0.04%) to 97.455%

Files with Coverage Reduction	New Missed Lines	%
pyccl/baryons/baccoemu_baryons.py	1	98.72%
pyccl/cells.py	1	97.92%
pyccl/nl_pt/ept.py	1	99.63%
pyccl/_nonlimber_FKEM.py	2	96.12%
pyccl/pk2d.py	3	98.71%
pyccl/tracers.py	12	95.07%
pyccl/halos/pk_4pt.py	16	96.25%

Totals
Change from base Build 11958349869:	0.04%
Covered Lines:	6548
Relevant Lines:	6719

💛 - Coveralls

rreischke · 2024-12-04T01:26:45Z

pyccl/halos/pk_4pt.py

@@ -669,6 +670,9 @@ def halomod_trispectrum_2h_22(cosmo, hmc, k, a, prof, *, prof2=None,
        p_of_k_a (:class:`~pyccl.pk2d.Pk2D`): a `Pk2D` object to
            be used as the linear matter power spectrum. If `None`, the power
            spectrum stored within `cosmo` will be used.
+        separable_growth (Boolean): Indeicates whether a separable


Typo: Indeicates -> Indicates and in all following occurrences.

and Boolean -> bool for consistency (and so the docs look nice)

rreischke · 2024-12-04T01:27:56Z

pyccl/halos/pk_4pt.py

@@ -976,6 +985,9 @@ def halomod_trispectrum_3h(cosmo, hmc, k, a, prof, *, prof2=None,
        p_of_k_a (:class:`~pyccl.pk2d.Pk2D`): a `Pk2D` object to
            be used as the linear matter power spectrum. If `None`, the power
            spectrum stored within `cosmo` will be used.
+        separable_growth (Boolean): Indeicates whether a separable


Typo: Indeicates -> Indicates

Boolean -> bool

damonge

Looks great @paulrogozenski ! Just a few typos (thanks @rreischke )

damonge · 2024-12-04T08:37:13Z

pyccl/halos/pk_4pt.py

@@ -1303,6 +1335,9 @@ def halomod_Tk3D_2h(cosmo, hmc,
        use_log (bool): if `True`, the trispectrum will be
            interpolated in log-space (unless negative or
            zero values are found).
+        separable_growth (Boolean): Indeicates whether a separable
+            growth function approximation can be used to calculate
+            the isotropized power spectrum.


Indeicates -> Indicates
Boolean -> bool

damonge · 2024-12-04T08:38:32Z

pyccl/halos/pk_4pt.py

@@ -1396,6 +1432,9 @@ def halomod_Tk3D_3h(cosmo, hmc,
        use_log (bool): if `True`, the trispectrum will be
            interpolated in log-space (unless negative or
            zero values are found).
+        separable_growth (Boolean): Indeicates whether a separable
+            growth function approximation can be used to calculate
+            the isotropized power spectrum.


Indicates, bool

damonge · 2024-12-04T08:39:02Z

pyccl/halos/pk_4pt.py

@@ -1469,6 +1509,9 @@ def halomod_Tk3D_4h(cosmo, hmc,
        use_log (bool): if `True`, the trispectrum will be
            interpolated in log-space (unless negative or
            zero values are found).
+        separable_growth (Boolean): Indeicates whether a separable
+            growth function approximation can be used to calculate
+            the isotropized power spectrum.


Indicates, bool

damonge · 2024-12-04T08:39:39Z

pyccl/halos/pk_4pt.py

@@ -669,6 +670,9 @@ def halomod_trispectrum_2h_22(cosmo, hmc, k, a, prof, *, prof2=None,
        p_of_k_a (:class:`~pyccl.pk2d.Pk2D`): a `Pk2D` object to
            be used as the linear matter power spectrum. If `None`, the power
            spectrum stored within `cosmo` will be used.
+        separable_growth (Boolean): Indeicates whether a separable


and Boolean -> bool for consistency (and so the docs look nice)

damonge · 2024-12-04T08:39:50Z

pyccl/halos/pk_4pt.py

@@ -976,6 +985,9 @@ def halomod_trispectrum_3h(cosmo, hmc, k, a, prof, *, prof2=None,
        p_of_k_a (:class:`~pyccl.pk2d.Pk2D`): a `Pk2D` object to
            be used as the linear matter power spectrum. If `None`, the power
            spectrum stored within `cosmo` will be used.
+        separable_growth (Boolean): Indeicates whether a separable


Boolean -> bool

damonge · 2024-12-04T08:39:56Z

pyccl/halos/pk_4pt.py

@@ -1140,6 +1159,9 @@ def halomod_trispectrum_4h(cosmo, hmc, k, a, prof, prof2=None, prof3=None,
        p_of_k_a (:class:`~pyccl.pk2d.Pk2D`): a `Pk2D` object to
            be used as the linear matter power spectrum. If `None`, the power
            spectrum stored within `cosmo` will be used.
+        separable_growth (Boolean): Indeicates whether a separable


Boolean -> bool

damonge · 2024-12-04T08:40:59Z

pyccl/halos/pk_4pt.py

@@ -1553,6 +1598,9 @@ def halomod_Tk3D_cNG(cosmo, hmc, prof, prof2=None, prof3=None, prof4=None,
        use_log (bool): if `True`, the trispectrum will be
            interpolated in log-space (unless negative or
            zero values are found).
+        separable_growth (Boolean): Indeicates whether a separable
+            growth function approximation can be used to calculate
+            the isotropized power spectrum.


Indicates, bool

damonge · 2024-12-04T08:42:35Z

Also, we need to add a unit test for this. Checking that you get the same result for some of this functions with or without scaling would be enough.

rreischke

Looks good to me, apart from the typo (as noted by @damonge ).

If the results agree, we can merge.

rreischke

Looks great, thanks Paul!

damonge

LGTM!

initial commit for allowing separable growth function in pk_4pt funct…

97e4498

…ions

paulrogozenski requested review from carlosggarcia and rreischke December 3, 2024 18:53

lint cleanup

6b6f01c

rreischke reviewed Dec 4, 2024

View reviewed changes

damonge requested changes Dec 4, 2024

View reviewed changes

rreischke requested changes Dec 4, 2024

View reviewed changes

paulrogozenski added 2 commits December 4, 2024 09:05

added tests for separable growth approximation; fixed typos

c74b4d0

added tests for separable growth approximation; fixed typos

b70bdb1

paulrogozenski requested review from rreischke and damonge December 4, 2024 14:41

rreischke approved these changes Dec 4, 2024

View reviewed changes

paulrogozenski removed the request for review from carlosggarcia December 4, 2024 14:57

damonge approved these changes Dec 4, 2024

View reviewed changes

damonge merged commit 406b548 into master Dec 4, 2024
4 checks passed

damonge deleted the cov_separable_growth branch December 4, 2024 15:07

paulrogozenski mentioned this pull request Dec 13, 2024

nGC covariance LSSTDESC/TJPCov#114

Draft

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Allow Separable Growth Functions to compute Isotropized Power Spectra #1216

Allow Separable Growth Functions to compute Isotropized Power Spectra #1216

paulrogozenski commented Dec 3, 2024

coveralls commented Dec 3, 2024 •

edited

Loading

rreischke Dec 4, 2024 •

edited

Loading

damonge Dec 4, 2024

rreischke Dec 4, 2024

damonge Dec 4, 2024

damonge left a comment

damonge Dec 4, 2024

damonge Dec 4, 2024

damonge Dec 4, 2024

damonge Dec 4, 2024

damonge Dec 4, 2024

damonge Dec 4, 2024

damonge Dec 4, 2024

damonge commented Dec 4, 2024

rreischke left a comment

rreischke left a comment

damonge left a comment

Allow Separable Growth Functions to compute Isotropized Power Spectra #1216

Allow Separable Growth Functions to compute Isotropized Power Spectra #1216

Conversation

paulrogozenski commented Dec 3, 2024

coveralls commented Dec 3, 2024 • edited Loading

Pull Request Test Coverage Report for Build 12161562031

Warning: This coverage report may be inaccurate.

Details

💛 - Coveralls

rreischke Dec 4, 2024 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

damonge left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

damonge commented Dec 4, 2024

rreischke left a comment

Choose a reason for hiding this comment

rreischke left a comment

Choose a reason for hiding this comment

damonge left a comment

Choose a reason for hiding this comment

coveralls commented Dec 3, 2024 •

edited

Loading

rreischke Dec 4, 2024 •

edited

Loading