3
I Use This!
Low Activity

Commits : Listings

Analyzed about 8 hours ago. based on code collected about 9 hours ago.
Oct 09, 2025 — Oct 09, 2026
Commit Message Contributor Files Modified Lines Added Lines Removed Code Location Date
Added more info about preferred parallelizer in Numexpr. More... about 16 years ago
Better function description. More... about 16 years ago
Heating engine for 1.4 release. More... about 16 years ago
Applied the patch for adding where_bbbb and cast_ib opcodes. Fixes #28. More... about 16 years ago
Merged branch multithreaded into trunk. More... about 16 years ago
pthread library is not needed to be explicitly put for MKL. More... about 16 years ago
Changed size_t to intp as per discussion in the NumPy mailing list. More... about 16 years ago
Added a note about the size_t use for arrays lengths. More... about 16 years ago
Fixed a couple of errors introduced in the last big changeset. Also, the size_t typesize has been used so as to deal with array lengths. More... about 16 years ago
Added mkl_solver_ilp64 to the list of suggested MKL libraries. Fixes #27. More... about 16 years ago
Added an example on how to link with VML in Windows 64-bit. More... over 16 years ago
Fix #26. I neglected to rename the opcodes OP_ONES_LIKE_II to OP_ONE_LIKE_I in interpreter.c when I renamed them in opcodes.inc. More... over 16 years ago
Better nose compability for the tests. - Use a generator to create the expressions test cases. Create a containing class within the suite() function for them for when using unittest only. - Mark the test() function as being not a test itself. More... over 16 years ago
Explicitly number the opcodes. More... over 16 years ago
Do the same thing with functions as with the opcodes. Construct a "database" of the functions in functions.inc, which is queried using C preprocessor magic. More... over 16 years ago
Actually add in the list of opcodes... More... over 16 years ago
Replace multiple lists of opcodes with a "database" opcodes.inc The enums, signatures, and exported names are selected from this by C preprocessor magic. Also replace OP_FUNC_FF (and similiar) with OP_FUNC_FFN, making explicit in the name the last argument which specifies the function. More... over 16 years ago
Add note to RELEASE_NOTES.txt about fixing #25 More... over 16 years ago
[ISSUE 25] OP_NEG_LL treats the argument as an int, not a long long. Added fix and test case. More... over 16 years ago
Specify that set_vml_accuracy_mode() returns the previous mode. More... almost 17 years ago
Merged in r147 from 1.3 branch (Small output beautification). More... over 17 years ago
unaligned-bench.py renamed to unaligned-simple.py More... over 17 years ago
When on AMD/Intel platforms, copies for unaligned arrays are disabled. This is because AMD/Intel processors can deal very efficiently with unaligned data, so avoiding the copy is generally a good thing (up to 2x speed-ups wrt doing a copy can easily be seen). More... over 17 years ago
Merged in r140 from 1.3 branch (Put in sync with 1.3.1). More... over 17 years ago
Release set to false again. More... over 17 years ago
Merged in r130:138 from 1.3 branch (Improvements and fixes for 1.3.1). More... over 17 years ago
Added a new dependency of msvc_function_stubs.inc. More... over 17 years ago
Some rephrasing in announcement. More... over 17 years ago
More additions to the releasing procedure. More... over 17 years ago
Ops, Forgot to increase the version number in the announcement. More... over 17 years ago