Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Improve payne hanek #210

Merged
merged 7 commits into from
Aug 16, 2018
Merged

Improve payne hanek #210

merged 7 commits into from
Aug 16, 2018

Conversation

shibatch
Copy link
Owner

This patch improves the Payne-Hanek reduction algorithm.
The restriction on the input domain of trig functions is all removed.

#define CONFIG 1
#include "helperpurec.h"
#include "norename.h"
#ifdef ENABLE_AVX512FNOFMA
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The changes in this file seem to be unnecessary. If I understand correctly, these are the "nofma" and "pure_c*_scalar" target that are needed for the "determinism" patch, and are unrelated to the improvements of the payne hanek algorithm.

In fact, "rename*nofma.h" is not included in this patch.

May I ask you to revert the changes in this file?

@@ -126,6 +112,18 @@ typedef Sleef_vector_double_2 vdouble2;
typedef Sleef_vector_float_2 vfloat2;
#endif

#ifdef ENABLE_VECEXT
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Please revert the change in line 115-126.

@@ -111,11 +99,9 @@ typedef Sleef_float32x4_t_2 vfloat2;
#ifdef ENABLE_SVE
#define CONFIG 1
#include "helpersve.h"
#ifdef DORENAME
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Why remove this?

Copy link
Owner Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Because DORENAME is always defined.

@@ -127,6 +121,12 @@ typedef Sleef_vector_double_2 vdouble2;
typedef Sleef_vector_float_2 vfloat2;
#endif

#ifdef ENABLE_PUREC
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Please restore line 124-129 in the original position.

@@ -2473,6 +2463,7 @@ EXPORT CONST vfloat __tgammaf_u1_finite(vfloat) __attribute__((weak, ali
// gcc -DENABLE_MAIN -Wno-attributes -I../common -I../arch -DENABLE_AVX2 -mavx2 -mfma sleefsimdsp.c ../common/common.c -lm
#include <stdio.h>
#include <stdlib.h>
#include <math.h>
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Spurious include?

@shibatch shibatch merged commit 86d5c2d into master Aug 16, 2018
@shibatch shibatch deleted the Improve_Payne_Hanek branch August 16, 2018 03:28
shibatch added a commit that referenced this pull request Aug 19, 2018
This patch improves the Payne-Hanek reduction algorithm.
The restriction on the input domain of trig functions is all removed.
@shibatch shibatch mentioned this pull request Aug 19, 2018
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants