36c1052183
This patch is a code optimisation, trading text bytes for speed. On
pyboard it's an increase of 0.06% in code size for a gain (in pystone
performance) of roughly 6.5%.
The patch optimises load/store/delete of attributes in user defined classes
by not looking up special accessors (@property, __get__, __delete__,
__set__, __setattr__ and __getattr_) if they are guaranteed not to exist in
the class.
Currently, if you do my_obj.foo() then the runtime has to do a few checks
to see if foo is a property or has __get__, and if so delegate the call.
And for stores things like my_obj.foo = 1 has to first check if foo is a
property or has __set__ defined on it.
Doing all those checks each and every time the attribute is accessed has a
performance penalty. This patch eliminates all those checks for cases when
it's guaranteed that the checks will always fail, ie no attributes are
properties nor have any special accessor methods defined on them.
To make this guarantee it checks all attributes of a user-defined class
when it is first created. If any of the attributes of the user class are
properties or have special accessors, or any of the base classes of the
user class have them, then it sets a flag in the class to indicate that
special accessors must be checked for. Then in the load/store/delete code
it checks this flag to see if it can take the shortcut and optimise the
lookup.
It's an optimisation that's pretty widely applicable because it improves
lookup performance for all methods of user defined classes, and stores of
attributes, at least for those that don't have special accessors. And, it
allows to enable descriptors with minimal additional runtime overhead if
they are not used for a particular user class.
There is one restriction on dynamic class creation that has been introduced
by this patch: a user-defined class cannot go from zero special accessors
to one special accessor (or more) after that class has been subclassed. If
the script attempts this an AttributeError is raised (see addition to
tests/misc/non_compliant.py for an example of this case).
The cost in code space bytes for the optimisation in this patch is:
unix x64: +528
unix nanbox: +508
stm32: +192
cc3200: +200
esp8266: +332
esp32: +244
Performance tests that were done:
- on unix x86-64, pystone improved by about 5%
- on pyboard, pystone improved by about 6.5%, from 1683 up to 1794
- on pyboard, bm_chaos (from CPython benchmark suite) improved by about 5%
- on esp32, pystone improved by about 30% (but there are caching effects)
- on esp32, bm_chaos improved by about 11%
|
||
|---|---|---|
| docs | ||
| drivers | ||
| examples | ||
| extmod | ||
| lib | ||
| logo | ||
| mpy-cross | ||
| ports | ||
| py | ||
| tests | ||
| tools | ||
| .gitattributes | ||
| .gitignore | ||
| .gitmodules | ||
| .travis.yml | ||
| ACKNOWLEDGEMENTS | ||
| CODECONVENTIONS.md | ||
| CONTRIBUTING.md | ||
| LICENSE | ||
| README.md | ||
README.md
The MicroPython project
This is the MicroPython project, which aims to put an implementation of Python 3.x on microcontrollers and small embedded systems. You can find the official website at micropython.org.
WARNING: this project is in beta stage and is subject to changes of the code-base, including project-wide name changes and API changes.
MicroPython implements the entire Python 3.4 syntax (including exceptions,
with, yield from, etc., and additionally async/await keywords from
Python 3.5). The following core datatypes are provided: str (including
basic Unicode support), bytes, bytearray, tuple, list, dict, set,
frozenset, array.array, collections.namedtuple, classes and instances.
Builtin modules include sys, time, and struct, etc. Select ports have
support for _thread module (multithreading). Note that only a subset of
Python 3 functionality is implemented for the data types and modules.
MicroPython can execute scripts in textual source form or from precompiled bytecode, in both cases either from an on-device filesystem or "frozen" into the MicroPython executable.
See the repository http://github.com/micropython/pyboard for the MicroPython board (PyBoard), the officially supported reference electronic circuit board.
Major components in this repository:
- py/ -- the core Python implementation, including compiler, runtime, and core library.
- mpy-cross/ -- the MicroPython cross-compiler which is used to turn scripts into precompiled bytecode.
- ports/unix/ -- a version of MicroPython that runs on Unix.
- ports/stm32/ -- a version of MicroPython that runs on the PyBoard and similar STM32 boards (using ST's Cube HAL drivers).
- ports/minimal/ -- a minimal MicroPython port. Start with this if you want to port MicroPython to another microcontroller.
- tests/ -- test framework and test scripts.
- docs/ -- user documentation in Sphinx reStructuredText format. Rendered HTML documentation is available at http://docs.micropython.org (be sure to select needed board/port at the bottom left corner).
Additional components:
- ports/bare-arm/ -- a bare minimum version of MicroPython for ARM MCUs. Used mostly to control code size.
- ports/teensy/ -- a version of MicroPython that runs on the Teensy 3.1 (preliminary but functional).
- ports/pic16bit/ -- a version of MicroPython for 16-bit PIC microcontrollers.
- ports/cc3200/ -- a version of MicroPython that runs on the CC3200 from TI.
- ports/esp8266/ -- an experimental port for ESP8266 WiFi modules.
- extmod/ -- additional (non-core) modules implemented in C.
- tools/ -- various tools, including the pyboard.py module.
- examples/ -- a few example Python scripts.
The subdirectories above may include READMEs with additional info.
"make" is used to build the components, or "gmake" on BSD-based systems. You will also need bash, gcc, and Python (at least 2.7 or 3.3).
The Unix version
The "unix" port requires a standard Unix environment with gcc and GNU make. x86 and x64 architectures are supported (i.e. x86 32- and 64-bit), as well as ARM and MIPS. Making full-featured port to another architecture requires writing some assembly code for the exception handling and garbage collection. Alternatively, fallback implementation based on setjmp/longjmp can be used.
To build (see section below for required dependencies):
$ git submodule update --init
$ cd ports/unix
$ make axtls
$ make
Then to give it a try:
$ ./micropython
>>> list(5 * x + y for x in range(10) for y in [4, 2, 1])
Use CTRL-D (i.e. EOF) to exit the shell.
Learn about command-line options (in particular, how to increase heap size
which may be needed for larger applications):
$ ./micropython --help
Run complete testsuite:
$ make test
Unix version comes with a builtin package manager called upip, e.g.:
$ ./micropython -m upip install micropython-pystone
$ ./micropython -m pystone
Browse available modules on PyPI. Standard library modules come from micropython-lib project.
External dependencies
Building MicroPython ports may require some dependencies installed.
For Unix port, libffi library and pkg-config tool are required. On
Debian/Ubuntu/Mint derivative Linux distros, install build-essential
(includes toolchain and make), libffi-dev, and pkg-config packages.
Other dependencies can be built together with MicroPython. This may be required to enable extra features or capabilities, and in recent versions of MicroPython, these may be enabled by default. To build these additional dependencies, first fetch git submodules for them:
$ git submodule update --init
Use the same command to get the latest versions of dependencies, as
they are updated from time to time. After that, in the port directory
(e.g. ports/unix/), execute:
$ make deplibs
This will build all available dependencies (regardless whether they
are used or not). If you intend to build MicroPython with additional
options (like cross-compiling), the same set of options should be passed
to make deplibs. To actually enable/disable use of dependencies, edit
ports/unix/mpconfigport.mk file, which has inline descriptions of the options.
For example, to build SSL module (required for upip tool described above,
and so enabled by dfeault), MICROPY_PY_USSL should be set to 1.
For some ports, building required dependences is transparent, and happens automatically. They still need to be fetched with the git submodule command above.
The STM32 version
The "stm32" port requires an ARM compiler, arm-none-eabi-gcc, and associated bin-utils. For those using Arch Linux, you need arm-none-eabi-binutils, arm-none-eabi-gcc and arm-none-eabi-newlib packages. Otherwise, try here: https://launchpad.net/gcc-arm-embedded
To build:
$ git submodule update --init
$ cd ports/stm32
$ make
You then need to get your board into DFU mode. On the pyboard, connect the 3V3 pin to the P1/DFU pin with a wire (on PYBv1.0 they are next to each other on the bottom left of the board, second row from the bottom).
Then to flash the code via USB DFU to your device:
$ make deploy
This will use the included tools/pydfu.py script. If flashing the firmware
does not work it may be because you don't have the correct permissions, and
need to use sudo make deploy.
See the README.md file in the ports/stm32/ directory for further details.
Contributing
MicroPython is an open-source project and welcomes contributions. To be productive, please be sure to follow the Contributors' Guidelines and the Code Conventions. Note that MicroPython is licenced under the MIT license, and all contributions should follow this license.