
======================== Announcing NumExpr 2.7.2 ======================== Hi everyone, It's been awhile since the last update to NumExpr, mostly as the existing scientific Python tool chain for building wheels on PyPi became defunct and we have had to redevelop a new one based on `cibuildwheel` and GitHub Actions. This release also brings us support (and wheels for) Python 3.9. There have been a number of changes to enhance how NumExpr works when NumPy uses MKL as a backend. Project documentation is available at: http://numexpr.readthedocs.io/ Changes from 2.7.1 to 2.7.2 --------------------------- - Support for Python 2.7 and 3.5 is deprecated and will be discontinued when `cibuildwheels` and/or GitHub Actions no longer support these versions. - Wheels are now provided for Python 3.7, 3.5, 3.6, 3.7, 3.8, and 3.9 via GitHub Actions. - The block size is now exported into the namespace as `numexpr.__BLOCK_SIZE1__` as a read-only value. - If using MKL, the number of threads for VML is no longer forced to 1 on loading the module. Testing has shown that VML never runs in multi-threaded mode for the default BLOCKSIZE1 of 1024 elements, and forcing to 1 can have deleterious effects on NumPy functions when built with MKL. See issue #355 for details. - Use of `ndarray.tostring()` in tests has been switch to `ndarray.tobytes()` for future-proofing deprecation of `.tostring()`, if the version of NumPy is greater than 1.9. - Added a utility method `get_num_threads` that returns the (maximum) number of threads currently in use by the virtual machine. The functionality of `set_num_threads` whereby it returns the previous value has been deprecated and will be removed in 2.8.X. What's Numexpr? --------------- Numexpr is a fast numerical expression evaluator for NumPy. With it, expressions that operate on arrays (like "3*a+4*b") are accelerated and use less memory than doing the same calculation in Python. It has multi-threaded capabilities, as well as support for Intel's MKL (Math Kernel Library), which allows an extremely fast evaluation of transcendental functions (sin, cos, tan, exp, log...) while squeezing the last drop of performance out of your multi-core processors. Look here for a some benchmarks of numexpr using MKL: https://github.com/pydata/numexpr/wiki/NumexprMKL Its only dependency is NumPy (MKL is optional), so it works well as an easy-to-deploy, easy-to-use, computational engine for projects that don't want to adopt other solutions requiring more heavy dependencies. Where I can find Numexpr? ------------------------- The project is hosted at GitHub in: https://github.com/pydata/numexpr You can get the packages from PyPI as well (but not for RC releases): http://pypi.python.org/pypi/numexpr Documentation is hosted at: http://numexpr.readthedocs.io/en/latest/ Share your experience --------------------- Let us know of any bugs, suggestions, gripes, kudos, etc. you may have. Enjoy data! -- Robert McLeod robbmcleod@gmail.com robert.mcleod@hitachi-hhtc.ca
participants (1)
-
Robert McLeod