You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: CHANGELOG
+16-8
Original file line number
Diff line number
Diff line change
@@ -5,19 +5,27 @@ V 1.2 (June 2020)
5
5
* a major change to code, internally, and major improvements to operation and
6
6
language interfaces.
7
7
WARNING: Here are all the interface changes from 1.1.2:
8
-
- opts is always passed as a pointer in C++/C, not pass-by-reference
9
-
as in v1.1.2 or earlier.
8
+
- opts (nufft_opts) is always passed as a pointer in C++/C, not
9
+
pass-by-reference as in v1.1.2 or earlier.
10
10
- Fortran simple calls are now finufft?d?(..) not finufft?d?_f(..), and
11
11
they add a penultimate opts argument.
12
-
-
13
-
These changes were deemed necessary to rationalize and improve FINUFFT.
12
+
- ier=1 is now a warning not an error; this indicates requested tol
13
+
was too small, but that a transform *was* done at the best possible
14
+
accuracy.
15
+
- opts.nthreads was removed in language interfaces. Control of the
16
+
number of available OpenMP threads is up to the user.
17
+
- opts.fftw directly controls the FFTW plan mode consistently in all
18
+
language interfaces (this changes the meaning of fftw=0 in MATLAB).
19
+
These changes were deemed necessary to rationalize and improve FINUFFT
20
+
for the long term.
14
21
15
22
...
16
23
* completely new Fortran interface, allowing >2^31 sized (int64) arrays,
17
-
all simple, many-vector and guru interface, with options control.
18
-
* all simple and many-vector interfaces call guru interface.
19
-
* new guru interface, by Andrea Malleo and Alex Barnett, allowing easier
20
-
language wrapping and control of point-setting, reuse of sorting and FFTW plans. This finally bypasses the 0.1ms/thread cost of FFTW looking up previous wisdom, which slowed down performance for many small problems.
24
+
all simple, many-vector and guru interface, with full options control.
25
+
* all simple and many-vector interfaces now call guru interface, for much
26
+
better maintainability and less code repetition.
27
+
* new guru interface, by Andrea Malleo and Alex Barnett, allowing easier language wrapping and control of point-setting, reuse of sorting and FFTW plans. This finally bypasses the 0.1ms/thread cost of FFTW looking up previous wisdom, which slowed down performance for many small problems.
0 commit comments