GPU Batch 7 #1187
Merged
GPU Batch 7 #1187
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Fix igamma_large support on device Add GPU support to toms748 Add GPU support to igamma_inv Add GPU markers to gamma_inva Add GPU Markers to lgamma_small Remove STL usage from gamma Remove NVRTC workaround Fix fraction use of STL headers Mark gamma functions in fwd Disable declval on all GPU platforms Disable more unneeded code on device Add forward decl for NVRTC tgamma Disable unneeded items for all GPU Change workaround for missing overloads Rearrange definition location Add include path to cuda now that workaround is removed Fix NVRTC incompatibility with recursion and forward decls Add tgamma_ratio CUDA and NVRTC testing Fix NVRTC handling of gamma_p_derivative Add gamma_p_derivative CUDA and NVRTC testing Remove recursion from gamma_incomplete_imp Add SYCL testing of igamma, igamma_inv, and igamma_inva Ignore literal-range warnings Remove use of static const char* for function name Fix missing CUDA header Remove calls under NVRTC to fwd decl Add more nvrtc workarounds Use builtin erfc instead of header cycle Add CUDA and NVRTC testing of gamma_p_inv Adjust tolerances Add GPU support to chi squared dist Fix static local variable Add chi squared dist SYCL testing Add chi squared dist CUDA testing Add chi squared dist NVRTC testing Add GPU support to weibull dist Add weibull dist SYCL testing Add weibull dist CUDA testing Add weibull dist NVRTC testing
Codecov ReportAttention: Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## develop #1187 +/- ##
===========================================
- Coverage 94.08% 94.06% -0.02%
===========================================
Files 780 780
Lines 65815 65848 +33
===========================================
+ Hits 61922 61941 +19
- Misses 3893 3907 +14
... and 1 file with indirect coverage changes Continue to review full report in Codecov by Sentry.
|
Merged
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This begins to remove earlier workarounds for NVRTC now that support is more robust. In this PR workarounds were removed from gamma so now instead of having just
tgamma, andlgammaon device we have the whole gamma family including:tgamma_ratio,tgamma_delta_ratio,gamma_p_inv,gamma_p_inva,gamma_q_inv,gamma_q_inva,gamma_p_derivative,igamma,igamma_inv,igamma_inva. This additional support allows us to add GPU support to the chi squared distribution and weibull distributions. We also gain root finding support with toms748 now being device capable.This PR is a bit uglier than other because NVRTC does not really support forward declarations.
gamma.hppused these extensively so we either call the CUDA standard library where we can, or apply the correct pre-processing to call detail functions. A little bit of recursion was eliminated at the cost of additional dispatching functions.Completed device CI runs can be found here: cppalliance/cuda-math#18.
CC: @steppi, @dschmitz89, @izaid.