Monday, October 30, 2017

How to make your code 80 times faster

I often hear people who are happy because PyPy makes their code 2 times faster or so. Here is a short personal story which shows PyPy can go well beyond that.

DISCLAIMER: this is not a silver bullet or a general recipe: it worked in this particular case, it might not work so well in other cases. But I think it is still an interesting technique. Moreover, the various steps and implementations are showed in the same order as I tried them during the development, so it is a real-life example of how to proceed when optimizing for PyPy.

Some months ago I played a bit with evolutionary algorithms: the ambitious plan was to automatically evolve a logic which could control a (simulated) quadcopter, i.e. a PID controller (spoiler: it doesn't fly).

The idea is to have an initial population of random creatures: at each generation, the ones with the best fitness survive and reproduce with small, random variations.

However, for the scope of this post, the actual task at hand is not so important, so let's jump straight to the code. To drive the quadcopter, a Creature has a run_step method which runs at each delta_t (full code):
class Creature(object):
    INPUTS = 2  # z_setpoint, current z position
    OUTPUTS = 1 # PWM for all 4 motors
    STATE_VARS = 1
    ...

    def run_step(self, inputs):
        # state: [state_vars ... inputs]
        # out_values: [state_vars, ... outputs]
        self.state[self.STATE_VARS:] = inputs
        out_values = np.dot(self.matrix, self.state) + self.constant
        self.state[:self.STATE_VARS] = out_values[:self.STATE_VARS]
        outputs = out_values[self.STATE_VARS:]
        return outputs
  • inputs is a numpy array containing the desired setpoint and the current position on the Z axis;
  • outputs is a numpy array containing the thrust to give to the motors. To start easy, all the 4 motors are constrained to have the same thrust, so that the quadcopter only travels up and down the Z axis;
  • self.state contains arbitrary values of unknown size which are passed from one step to the next;
  • self.matrix and self.constant contains the actual logic. By putting the "right" values there, in theory we could get a perfectly tuned PID controller. These are randomly mutated between generations.
run_step is called at 100Hz (in the virtual time frame of the simulation). At each generation, we test 500 creatures for a total of 12 virtual seconds each. So, we have a total of 600,000 executions of run_step at each generation.

At first, I simply tried to run this code on CPython; here is the result:
$ python -m ev.main
Generation   1: ... [population = 500]  [12.06 secs]
Generation   2: ... [population = 500]  [6.13 secs]
Generation   3: ... [population = 500]  [6.11 secs]
Generation   4: ... [population = 500]  [6.09 secs]
Generation   5: ... [population = 500]  [6.18 secs]
Generation   6: ... [population = 500]  [6.26 secs]
Which means ~6.15 seconds/generation, excluding the first.

Then I tried with PyPy 5.9:
$ pypy -m ev.main
Generation   1: ... [population = 500]  [63.90 secs]
Generation   2: ... [population = 500]  [33.92 secs]
Generation   3: ... [population = 500]  [34.21 secs]
Generation   4: ... [population = 500]  [33.75 secs]
Ouch! We are ~5.5x slower than CPython. This was kind of expected: numpy is based on cpyext, which is infamously slow. (Actually, we are working on that and on the cpyext-avoid-roundtrip branch we are already faster than CPython, but this will be the subject of another blog post.)

So, let's try to avoid cpyext. The first obvious step is to use numpypy instead of numpy (actually, there is a hack to use just the micronumpy part). Let's see if the speed improves:
$ pypy -m ev.main   # using numpypy
Generation   1: ... [population = 500]  [5.60 secs]
Generation   2: ... [population = 500]  [2.90 secs]
Generation   3: ... [population = 500]  [2.78 secs]
Generation   4: ... [population = 500]  [2.69 secs]
Generation   5: ... [population = 500]  [2.72 secs]
Generation   6: ... [population = 500]  [2.73 secs]
So, ~2.7 seconds on average: this is 12x faster than PyPy+numpy, and more than 2x faster than the original CPython. At this point, most people would be happy and go tweeting how PyPy is great.

In general, when talking of CPython vs PyPy, I am rarely satified of a 2x speedup: I know that PyPy can do much better than this, especially if you write code which is specifically optimized for the JIT. For a real-life example, have a look at capnpy benchmarks, in which the PyPy version is ~15x faster than the heavily optimized CPython+Cython version (both have been written by me, and I tried hard to write the fastest code for both implementations).

So, let's try to do better. As usual, the first thing to do is to profile and see where we spend most of the time. Here is the vmprof profile. We spend a lot of time inside the internals of numpypy, and allocating tons of temporary arrays to store the results of the various operations.

Also, let's look at the jit traces and search for the function run: this is loop in which we spend most of the time, and it is composed of 1796 operations. The operations emitted for the line np.dot(...) + self.constant are listed between lines 1217 and 1456. Here is the excerpt which calls np.dot(...); most of the ops are cheap, but at line 1232 we see a call to the RPython function descr_dot; by looking at the implementation we see that it creates a new W_NDimArray to store the result, which means it has to do a malloc():

The implementation of the + self.constant part is also interesting: contrary the former, the call to W_NDimArray.descr_add has been inlined by the JIT, so we have a better picture of what's happening; in particular, we can see the call to __0_alloc_with_del____ which allocates the W_NDimArray for the result, and the raw_malloc which allocates the actual array. Then we have a long list of 149 simple operations which set the fields of the resulting array, construct an iterator, and finally do a call_assembler: this is the actual logic to do the addition, which was JITtted indipendently; call_assembler is one of the operations to do JIT-to-JIT calls:

All of this is very suboptimal: in this particular case, we know that the shape of self.matrix is always (3, 2): so, we are doing an incredible amount of work, including calling malloc() twice for the temporary arrays, just to call two functions which ultimately do a total of 6 multiplications and 6 additions. Note also that this is not a fault of the JIT: CPython+numpy has to do the same amount of work, just hidden inside C calls.

One possible solution to this nonsense is a well known compiler optimization: loop unrolling. From the compiler point of view, unrolling the loop is always risky because if the matrix is too big you might end up emitting a huge blob of code, possibly uselss if the shape of the matrices change frequently: this is the main reason why the PyPy JIT does not even try to do it in this case.

However, we know that the matrix is small, and always of the same shape. So, let's unroll the loop manually:
class SpecializedCreature(Creature):

    def __init__(self, *args, **kwargs):
        Creature.__init__(self, *args, **kwargs)
        # store the data in a plain Python list
        self.data = list(self.matrix.ravel()) + list(self.constant)
        self.data_state = [0.0]
        assert self.matrix.shape == (2, 3)
        assert len(self.data) == 8

    def run_step(self, inputs):
        # state: [state_vars ... inputs]
        # out_values: [state_vars, ... outputs]
        k0, k1, k2, q0, q1, q2, c0, c1 = self.data
        s0 = self.data_state[0]
        z_sp, z = inputs
        #
        # compute the output
        out0 = s0*k0 + z_sp*k1 + z*k2 + c0
        out1 = s0*q0 + z_sp*q1 + z*q2 + c1
        #
        self.data_state[0] = out0
        outputs = [out1]
        return outputs
In the actual code there is also a sanity check which asserts that the computed output is the very same as the one returned by Creature.run_step.

So, let's try to see how it performs. First, with CPython:
$ python -m ev.main
Generation   1: ... [population = 500]  [7.61 secs]
Generation   2: ... [population = 500]  [3.96 secs]
Generation   3: ... [population = 500]  [3.79 secs]
Generation   4: ... [population = 500]  [3.74 secs]
Generation   5: ... [population = 500]  [3.84 secs]
Generation   6: ... [population = 500]  [3.69 secs]
This looks good: 60% faster than the original CPython+numpy implementation. Let's try on PyPy:
Generation   1: ... [population = 500]  [0.39 secs]
Generation   2: ... [population = 500]  [0.10 secs]
Generation   3: ... [population = 500]  [0.11 secs]
Generation   4: ... [population = 500]  [0.09 secs]
Generation   5: ... [population = 500]  [0.08 secs]
Generation   6: ... [population = 500]  [0.12 secs]
Generation   7: ... [population = 500]  [0.09 secs]
Generation   8: ... [population = 500]  [0.08 secs]
Generation   9: ... [population = 500]  [0.08 secs]
Generation  10: ... [population = 500]  [0.08 secs]
Generation  11: ... [population = 500]  [0.08 secs]
Generation  12: ... [population = 500]  [0.07 secs]
Generation  13: ... [population = 500]  [0.07 secs]
Generation  14: ... [population = 500]  [0.08 secs]
Generation  15: ... [population = 500]  [0.07 secs]
Yes, it's not an error. After a couple of generations, it stabilizes at around ~0.07-0.08 seconds per generation. This is around 80 (eighty) times faster than the original CPython+numpy implementation, and around 35-40x faster than the naive PyPy+numpypy one.

Let's look at the trace again: it no longer contains expensive calls, and certainly no more temporary malloc() s. The core of the logic is between lines 386-416, where we can see that it does fast C-level multiplications and additions: float_mul and float_add are translated straight into mulsd and addsd x86 instructions.

As I said before, this is a very particular example, and the techniques described here do not always apply: it is not realistic to expect an 80x speedup on arbitrary code, unfortunately. However, it clearly shows the potential of PyPy when it comes to high-speed computing. And most importantly, it's not a toy benchmark which was designed specifically to have good performance on PyPy: it's a real world example, albeit small.

You might be also interested in the talk I gave at last EuroPython, in which I talk about a similar topic: "The Joy of PyPy JIT: abstractions for free" (abstract, slides and video).

How to reproduce the results

$ git clone https://github.com/antocuni/evolvingcopter
$ cd evolvingcopter
$ {python,pypy} -m ev.main --no-specialized --no-numpypy
$ {python,pypy} -m ev.main --no-specialized
$ {python,pypy} -m ev.main

Wednesday, October 18, 2017

(Cape of) Good Hope for PyPy


Hello from the other side of the world (for most of you)!

With the excuse of coming to PyCon ZA during the last two weeks Armin, Ronan, Antonio and sometimes Maciek had a very nice and productive sprint in Cape Town, as pictures show :). We would like to say a big thank you to Kiwi.com, which sponsored part of the travel costs via its awesome Sourcelift program to help Open Source projects.

Armin, Anto and Ronan at Cape Point

Armin, Ronan and Anto spent most of the time hacking at cpyext, our CPython C-API compatibility layer: during the last years, the focus was to make it working and compatible with CPython, in order to run existing libraries such as numpy and pandas. However, we never paid too much attention to performance, so the net result is that with the latest released version of PyPy, C extensions generally work but their speed ranges from "slow" to "horribly slow".

For example, these very simple microbenchmarks measure the speed of calling (empty) C functions, i.e. the time you spend to "cross the border" between RPython and C. (Note: this includes the time spent doing the loop in regular Python code.) These are the results on CPython, on PyPy 5.8, and on our newest in-progress version:

$ python bench.py     # CPython
noargs      : 0.41 secs
onearg(None): 0.44 secs
onearg(i)   : 0.44 secs
varargs     : 0.58 secs

$ pypy-5.8 bench.py   # PyPy 5.8
noargs      : 1.01 secs
onearg(None): 1.31 secs
onearg(i)   : 2.57 secs
varargs     : 2.79 secs

$ pypy bench.py       # cpyext-refactor-methodobject branch
noargs      : 0.17 secs
onearg(None): 0.21 secs
onearg(i)   : 0.22 secs
varargs     : 0.47 secs



So yes: before the sprint, we were ~2-6x slower than CPython. Now, we are
faster than it!
To reach this result, we did various improvements, such as:

  1. teach the JIT how to look (a bit) inside the cpyext module;
  2. write specialized code for calling METH_NOARGS, METH_O and METH_VARARGS functions; previously, we always used a very general and slow logic;
  3. implement freelists to allocate the cpyext versions of int and tuple objects, as CPython does;
  4. the cpyext-avoid-roundtrip branch: crossing the RPython/C border is slowish, but the real problem was (and still is for many cases) we often cross it many times for no good reason. So, depending on the actual API call, you might end up in the C land, which calls back into the RPython land, which goes to C, etc. etc. (ad libitum).
The branch tries to fix such nonsense: so far, we fixed only some cases, which are enough to speed up the benchmarks shown above. But most importantly, we now have a clear path and an actual plan to improve cpyext more and more. Ideally, we would like to reach a point in which cpyext-intensive programs run at worst at the same speed of CPython.

The other big topic of the sprint was Armin and Maciej doing a lot of work on the unicode-utf8 branch: the goal of the branch is to always use UTF-8 as the internal representation of unicode strings. The advantages are various:
  • decoding a UTF-8 stream is super fast, as you just need to check that the stream is valid;
  • encoding to UTF-8 is almost a no-op;
  • UTF-8 is always more compact representation than the currently used UCS-4. It's also almost always more compact than CPython 3.5 latin1/UCS2/UCS4 combo;
  • smaller representation means everything becomes quite a bit faster due to lower cache pressure.
Before you ask: yes, this branch contains special logic to ensure that random access of single unicode chars is still O(1), as it is on both CPython and the current PyPy.
We also plan to improve the speed of decoding even more by using modern processor features, like SSE and AVX. Preliminary results show that decoding can be done 100x faster than the current setup.

In summary, this was a long and profitable sprint, in which we achieved lots of interesting results. However, what we liked even more was the privilege of doing commits from awesome places such as the top of Table Mountain:


The panorama we looked at instead of staring at cpyext code

Thursday, October 5, 2017

PyPy v5.9 Released, Now Supports Pandas, NumPy

The PyPy team is proud to release both PyPy3.5 v5.9 (a beta-quality interpreter for Python 3.5 syntax) and PyPy2.7 v5.9 (an interpreter supporting Python 2.7 syntax).

  • NumPy and Pandas now work on PyPy2.7 (together with Cython 0.27.1). Many other modules based on C-API extensions work on PyPy as well.

  • Cython 0.27.1 (released very recently) supports more projects with PyPy, both on PyPy2.7 and PyPy3.5 beta. Note version 0.27.1 is now the minimum version that supports this version of PyPy, due to some interactions with updated C-API interface code.

  • We optimized the JSON parser for recurring string keys, which should decrease memory use by up to 50% and increase parsing speed by up to 15% for large JSON files with many repeating dictionary keys (which is quite common).

  • CFFI, which is part of the PyPy release, has been updated to 1.11.1, improving an already great package for interfacing with C. CFFI now supports complex arguments in API mode, as well as char16_t and char32_t and has improved support for callbacks.

  • Issues in the C-API compatibility layer that appeared as excessive memory use were cleared up and other incompatibilities were resolved. The C-API compatibility layer does slow down code which crosses the python-c interface often. Some fixes are in the pipelines for some of the performance issues, and we still recommend using pure python on PyPy or interfacing via CFFI

Please let us know if your use case is slow, we have ideas how to make things faster but need real-world examples (not micro-benchmarks) of problematic code.

Work sponsored by a Mozilla grant continues on PyPy3.5; we continue on the path to the goal of a complete python 3.5 implementation. Of course the bug fixes and performance enhancements mentioned above are part of both PyPy2.7 and PyPy3.5 beta.

As always, this release fixed many other issues and bugs raised by the growing community of PyPy users. We strongly recommend updating.

You can download the v5.9 releases here (note that we provide PyPy3.5 binaries for only Linux 64bit for now):

We would like to thank our donors and contributors, and encourage new people to join the project. PyPy has many layers and we need help with all of them: PyPy and RPython documentation improvements, tweaking popular modules to run on PyPy, or general help with making RPython’s JIT even better.

What is PyPy?

PyPy is a very compliant Python interpreter, almost a drop-in replacement for CPython 2.7 (stdlib version 2.7.13), and CPython 3.5 (stdlib version 3.5.3). It’s fast (PyPy and CPython 2.7.x performance comparison) due to its integrated tracing JIT compiler.

We also welcome developers of other dynamic languages to see what RPython can do for them.

The PyPy 2.7 release supports:
  • x86 machines on most common operating systems (Linux 32/64 bits, Mac OS X 64 bits, Windows 32 bits, OpenBSD, FreeBSD)
  • newer ARM hardware (ARMv6 or ARMv7, with VFPv3) running Linux,
  • big- and little-endian variants of PPC64 running Linux,
  • s390x running Linux

What else is new?

PyPy 5.8 was released in June, 2017.
There are many incremental improvements to RPython and PyPy, the complete listing is here.
 
Please update, and continue to help us make PyPy better.

Cheers, The PyPy team

Monday, August 14, 2017

Let's remove the Global Interpreter Lock

Hello everyone

The Python community has been discussing removing the Global Interpreter Lock for a long time. There have been various attempts at removing it: Jython or IronPython successfully removed it with the help of the underlying platform, and some have yet to bear fruit, like gilectomy. Since our February sprint in Leysin, we have experimented with the topic of GIL removal in the PyPy project. We believe that the work done in IronPython or Jython can be reproduced with only a bit more effort in PyPy. Compared to that, removing the GIL in CPython is a much harder topic, since it also requires tackling the problem of multi-threaded reference counting. See the section below for further details.

As we announced at EuroPython, what we have so far is a GIL-less PyPy which can run very simple multi-threaded, nicely parallelized, programs. At the moment, more complicated programs probably segfault. The remaining 90% (and another 90%) of work is with putting locks in strategic places so PyPy does not segfault during concurrent accesses to data structures.

Since such work would complicate the PyPy code base and our day-to-day work, we would like to judge the interest of the community and the commercial partners to make it happen (we are not looking for individual donations at this point). We estimate a total cost of $50k, out of which we already have backing for about 1/3 (with a possible 1/3 extra from the STM money, see below). This would give us a good shot at delivering a good proof-of-concept working PyPy with no GIL. If we can get a $100k contract, we will deliver a fully working PyPy interpreter with no GIL as a release, possibly separate from the default PyPy release.

People asked several questions, so I'll try to answer the technical parts here.

What would the plan entail?

We've already done the work on the Garbage Collector to allow doing multi- threaded programs in RPython. "All" that is left is adding locks on mutable data structures everywhere in the PyPy codebase. Since it would significantly complicate our workflow, we require real interest in that topic, backed up by commercial contracts in order to justify the added maintenance burden.

Why did the STM effort not work out?

STM was a research project that proved that the idea is possible. However, the amount of user effort that is required to make programs run in a parallelizable way is significant, and we never managed to develop tools that would help in doing so. At the moment we're not sure if more work spent on tooling would improve the situation or if the whole idea is really doomed. The approach also ended up adding significant overhead on single threaded programs, so in the end it is very easy to make your programs slower. (We have some money left in the donation pot for STM which we are not using; according to the rules, we could declare the STM attempt failed and channel that money towards the present GIL removal proposal.)

Wouldn't subinterpreters be a better idea?

Python is a very mutable language - there are tons of mutable state and basic objects (classes, functions,...) that are compile-time in other language but runtime and fully mutable in Python. In the end, sharing things between subinterpreters would be restricted to basic immutable data structures, which defeats the point. Subinterpreters suffers from the same problems as multiprocessing with no additional benefits. We believe that reducing mutability to implement subinterpreters is not viable without seriously impacting the semantics of the language (a conclusion which applies to many other approaches too).

Why is it easier to do in PyPy than CPython?

Removing the GIL in CPython has two problems:

  • how do we guard access to mutable data structures with locks and
  • what to do with reference counting that needs to be guarded.

PyPy only has the former problem; the latter doesn't exist, due to a different garbage collector approach. Of course the first problem is a mess too, but at least we are already half-way there. Compared to Jython or IronPython, PyPy lacks some data structures that are provided by JVM or .NET, which we would need to implement, hence the problem is a little harder than on an existing multithreaded platform. However, there is good research and we know how that problem can be solved.

Best regards,
Maciej Fijalkowski


Wednesday, July 26, 2017

Binary wheels for PyPy

Hi,

this is a short blog post, just to announce the existence of this Github repository, which contains binary PyPy wheels for some selected packages. The availability of binary wheels means that you can install the packages much more quickly, without having to wait for compilation.

At the moment of writing, these packages are available:

  • numpy
  • scipy
  • pandas
  • psutil
  • netifaces

For now, we provide only wheels built on Ubuntu, compiled for PyPy 5.8.
In particular, it is worth noting that they are not manylinux1 wheels, which means they could not work on other Linux distributions. For more information, see the explanation in the README of the above repo.

Moreover, the existence of the wheels does not guarantee that they work correctly 100% of the time. they still depend on cpyext, our C-API emulation layer, which is still work-in-progress, although it has become better and better during the last months. Again, the wheels are there only to save compilation time.

To install a package from the wheel repository, you can invoke pip like this:

$ pip install --extra-index https://antocuni.github.io/pypy-wheels/ubuntu numpy

Happy installing!

Friday, June 9, 2017

PyPy v5.8 released

The PyPy team is proud to release both PyPy2.7 v5.8 (an interpreter supporting Python 2.7 syntax), and a beta-quality PyPy3.5 v5.8 (an interpreter for Python 3.5 syntax). The two releases are both based on much the same codebase, thus the dual release. Note that PyPy3.5 supports Linux 64bit only for now.

This new PyPy2.7 release includes the upstream stdlib version 2.7.13, and PyPy3.5 includes the upstream stdlib version 3.5.3.

We fixed critical bugs in the shadowstack rootfinder garbage collector strategy that crashed multithreaded programs and very rarely showed up even in single threaded programs.

We added native PyPy support to profile frames in the vmprof statistical profiler.

The struct module functions pack* and unpack* are now much faster, especially on raw buffers and bytearrays. Microbenchmarks show a 2x to 10x speedup. Thanks to Gambit Research for sponsoring this work.

This release adds (but disables by default) link-time optimization and profile guided optimization of the base interpreter, which may make unjitted code run faster. To use these, translate with appropriate options. Be aware of issues with gcc toolchains, though.

Please let us know if your use case is slow, we have ideas how to make things faster but need real-world examples (not micro-benchmarks) of problematic code.

Work sponsored by a Mozilla grant continues on PyPy3.5; numerous fixes from CPython were ported to PyPy and PEP 489 was fully implemented. Of course the bug fixes and performance enhancements mentioned above are part of both PyPy 2.7 and PyPy 3.5.

CFFI, which is part of the PyPy release, has been updated to an unreleased 1.10.1, improving an already great package for interfacing with C.

Anyone using NumPy 1.13.0, must upgrade PyPy to this release since we implemented some previously missing C-API functionality. Many other c-extension modules now work with PyPy, let us know if yours does not.

As always, this release fixed many issues and bugs raised by the growing community of PyPy users. We strongly recommend updating.

You can download the v5.8 release here:
We would like to thank our donors and contributors, and encourage new people to join the project. PyPy has many layers and we need help with all of them: PyPy and RPython documentation improvements, tweaking popular modules to run on PyPy, or general help with making RPython’s JIT even better.

What is PyPy?

PyPy is a very compliant Python interpreter, almost a drop-in replacement for CPython 2.7 and CPython 3.5. It’s fast (PyPy and CPython 2.7.x performance comparison) due to its integrated tracing JIT compiler.
We also welcome developers of other dynamic languages to see what RPython can do for them.
The PyPy 2.7 release supports:
  • x86 machines on most common operating systems (Linux 32/64 bits, Mac OS X 64 bits, Windows 32 bits, OpenBSD, FreeBSD)
  • newer ARM hardware (ARMv6 or ARMv7, with VFPv3) running Linux,
  • big- and little-endian variants of PPC64 running Linux,
  • s390x running Linux

What else is new?

PyPy 5.7 was released in March, 2017.
There are many incremental improvements to RPython and PyPy, the complete listing is here.
 
Please update, and continue to help us make PyPy better.

Cheers, The PyPy team

Monday, April 3, 2017

PyPy 5.7.1 bugfix released

We have released a bugfix PyPy2.7-v5.7.1 and PyPy3.5-v5.7.1 beta (Linux 64bit), due to the following issues:
  • correctly handle an edge case in dict.pop (issue 2508)
  • fix a regression to correctly handle multiple inheritance in a C-API type where the second base is an app-level class with a __new__ function
  • fix a regression to fill a C-API type’s tp_getattr slot from a __getattr__ method (issue 2523)
Thanks to those who reported issues and helped test out the fixes

You can download the v5.7.1 release here:

What is PyPy?

PyPy is a very compliant Python interpreter, almost a drop-in replacement for CPython 2.7 and CPython 3.5. It’s fast (PyPy and CPython 2.7.x performance comparison) due to its integrated tracing JIT compiler.
We also welcome developers of other dynamic languages to see what RPython can do for them.
The PyPy 2.7 release supports:
  • x86 machines on most common operating systems (Linux 32/64 bits, Mac OS X 64 bits, Windows 32 bits, OpenBSD, FreeBSD)
  • newer ARM hardware (ARMv6 or ARMv7, with VFPv3) running Linux,
  • big- and little-endian variants of PPC64 running Linux,
  • s390x running Linux
Please update, and continue to help us make PyPy better.

Cheers, The PyPy team

Saturday, April 1, 2017

Native profiling in VMProf

We are happy to announce a new release for the PyPI package vmprof.
It is now able to capture native stack frames on Linux and Mac OS X to show you bottle necks in compiled code (such as CFFI modules, Cython or C Python extensions). It supports PyPy, CPython versions 2.7, 3.4, 3.5 and 3.6. Special thanks to Jetbrains for funding the native profiling support.

vmprof logo

What is vmprof?

If you have already worked with vmprof you can skip the next two section. If not, here is a short introduction:

The goal of vmprof package is to give you more insight into your program. It is a statistical profiler. Another prominent profiler you might already have worked with is cProfile. It is bundled with the Python standard library.

vmprof's distinct feature (from most other profilers) is that it does not significantly slow down your program execution. The employed strategy is statistical, rather than deterministic. Not every function call is intercepted, but it samples stack traces and memory usage at a configured sample rate (usually around 100hz). You can imagine that this creates a lot less contention than doing work before and after each function call.

As mentioned earlier cProfile gives you a complete profile, but it needs to intercept every function call (it is a deterministic profiler). Usually this means that you have to capture and record every function call, but this takes an significant amount time.

The overhead vmprof consumes is roughly 3-4% of your total program runtime or even less if you reduce the sampling frequency. Indeed it lets you sample and inspect much larger programs. If you failed to profile a large application with cProfile, please give vmprof a shot.

vmprof.com or PyCharm

There are two major alternatives to the command-line tools shipped with vmprof:
  • A web service on vmprof.com
  • PyCharm Professional Edition
While the command line tool is only good for quick inspections, vmprof.com and PyCharm compliment each other providing deeper insight into your program. With PyCharm you can view the per-line profiling results inside the editor. With the vmprof.com you get a handy visualization of the profiling results as a flame chart and memory usage graph.

Since the PyPy Team runs and maintains the service on vmprof.com (which is by the way free and open-source), I’ll explain some more details here. On vmprof.com you can inspect the generated profile interactively instead of looking at console output. What is sent to vmprof.com? You can find details here.

Flamegraph: Accumulates and displays the most frequent codepaths. It allows you to quickly and accurately identify hot spots in your code. The flame graph below is a very short run of richards.py (Thus it shows a lot of time spent in PyPy's JIT compiler).



List all functions (optionally sorted): the equivalent of the vmprof command line output in the web.


 Memory curve: A line plot that shows how how many MBytes have been consumed over the lifetime of your program (see more info in the section below).

Native programs

The new feature introduced in vmprof 0.4.x allows you to look beyond the Python level. As you might know, Python maintains a stack of frames to save the execution. Up to now the vmprof profiles only contained that level of information. But what if you program jumps to native code (such as calling gzip compression on a large file)? Up to now you would not see that information.

Many packages make use of the CPython C API (which we discurage, please lookup cffi for a better way to call C). Have you ever had the issue that you know that your performance problems reach down to, but you could not profile it properly? Now you can!

Let's inspect a very simple Python program to find out why a program is significantly slower on Linux than on Mac:

import numpy as np
n = 1000
a = np.random.random((n, n))
b = np.random.random((n, n))
c = np.dot(np.abs(a), b)



Take two NxN random matrix objects and create a dot product. The first argument to the dot product provides the absolute value of the random matrix.

RunPythonNumPyOSn=... Took
[1]CPython 3.5.2NumPy 1.12.1Mac OS X, 10.12.3n=5000~9 sec
[2]CPython 3.6.0NumPy 1.12.1Linux 64, Kernel 4.9.14n=1000~26 sec

Note that the Linux machine operates on a 5 times smaller matrix, still it takes much longer. What is wrong? Is Linux slow? CPython 3.6.0? Well no, lets inspect and [1] and [2] (shown below in that order).

[2] runs on Linux, spends nearly all of the time in PyArray_MatrixProduct2, if you compare to [1] on Mac OS X, you'll see that a lot of time is spent in generating the random numbers and the rest in cblas_matrixproduct.

Blas has a very efficient implementation so you can achieve the same on Linux if you install a blas implementation (such as openblas).

Usually you can spot potential program source locations that take a lot of time and might be the first starting point to resolve performance issues.

Beyond Python programs

It is not unthinkable that the strategy can be reused for native programs. Indeed this can already be done by creating a small cffi wrapper around an entry point of a compiled C program. It would even work for programs compiled from other languages (e.g. C++ or Fortran). The resulting function names are the full symbol name embedded into either the executable symboltable or extracted from the dwarf debugging information. Most of those will be compiler specific and contain some cryptic information.

Memory profiling
We thankfully received a code contribution from the company Blue Yonder. They have built a memory profiler (for Linux and Mac OS X) on top of vmprof.com that displays the memory consumption for the runtime of your process.

You can run it the following way:

$ python -m vmprof --mem --web script.py

By adding --mem, vmprof will capture memory information and display it in the dedicated view on vmprof.com. You can tha view by by clicking the 'Memory' switch in the flamegraph view.

There is more

Some more minor highlights contained in 0.4.x:
  • VMProf support for Windows 64 bit (No native profiling)
  • VMProf can read profiles generated by another host system
  • VMProf is now bundled in several binary wheel for fast and easy installation (Mac OS X, Linux 32/64 for CPython 2.7, 3.4, 3.5, 3.6)
Future plans - Profile Streaming

vmprof has not reached the end of development. There are many features we could implement. But there is one feature that could be a great asset to many Python developers.

Continuous delivery of your statistical profile, or in short, profile streaming. One of the great strengths of vmprof is that is consumes very little overhead. It is not a crazy idea to run this in production.

It would require a smart way to stream the profile in the background to vmprof.com and new visualizations to look at much more data your Python service produces.

If that sounds like a solid vmprof improvement, don't hesitate to get in touch with us (e.g. IRC #pypy, mailing list pypy-dev, or comment below)

You can help!

There are some immediate things other people could help with. Either by donating time or money (yes we have occasional contributors which is great)!
  • We gladly received code contribution for the memory profiler. But it was not enough time to finish the migration completely. Sadly it is a bit brittle right now.
  • We would like to spend more time on other visualizations. This should include to give a much better user experience on vmprof.com (like a tutorial that explains the visualization that we already have). 
  • Build Windows 32/64 bit wheels (for all CPython versions we currently support)
We are also happy to accept google summer of code projects on vmprof for new visualizations and other improvements. If you qualify and are interested, don't hesitate to ask!

Richard Plangger (plan_rich) and the PyPy Team

[1] Mac OS X http://vmprof.com/#/567aa150-5927-4867-b22d-dbb67ac824ac
[2] Linux64 http://vmprof.com/#/097fded2-b350-4d68-ae93-7956cd10150c

Tuesday, March 21, 2017

PyPy2.7 and PyPy3.5 v5.7 - two in one release

The PyPy team is proud to release both PyPy2.7 v5.7 (an interpreter supporting Python v2.7 syntax), and a beta-quality PyPy3.5 v5.7 (an interpreter for Python v3.5 syntax). The two releases are both based on much the same codebase, thus the dual release. Note that PyPy3.5 only supports Linux 64bit for now.

This new PyPy2.7 release includes the upstream stdlib version 2.7.13, and PyPy3.5 (our first in the 3.5 series) includes the upstream stdlib version 3.5.3.

We continue to make incremental improvements to our C-API compatibility layer (cpyext). PyPy2 can now import and run many C-extension packages, among the most notable are Numpy, Cython, and Pandas. Performance may be slower than CPython, especially for frequently-called short C functions. Please let us know if your use case is slow, we have ideas how to make things faster but need real-world examples (not micro-benchmarks) of problematic code.

Work proceeds at a good pace on the PyPy3.5 version due to a grant from the Mozilla Foundation, hence our first 3.5.3 beta release. Thanks Mozilla !!! While we do not pass all tests yet, asyncio works and as these benchmarks show it already gives a nice speed bump. We also backported the f"" formatting from 3.6 (as an exception; otherwise “PyPy3.5” supports the Python 3.5 language).

CFFI has been updated to 1.10, improving an already great package for interfacing with C.

We now use shadowstack as our default gcrootfinder even on Linux. The alternative, asmgcc, will be deprecated at some future point. While about 3% slower, shadowstack is much more easily maintained and debuggable. Also, the performance of shadowstack has been improved in general: this should close the speed gap between other platforms and Linux.

As always, this release fixed many issues and bugs raised by the growing community of PyPy users. We strongly recommend updating.

You can download the v5.7 release here:
We would like to thank our donors for the continued support of the PyPy project.
We would also like to thank our contributors and encourage new people to join the project. PyPy has many layers and we need help with all of them: PyPy and RPython documentation improvements, tweaking popular modules to run on pypy, or general help with making RPython’s JIT even better.

 

What is PyPy?

PyPy is a very compliant Python interpreter, almost a drop-in replacement for CPython 2.7 and CPython 3.5. It’s fast (PyPy and CPython 2.7.x performance comparison) due to its integrated tracing JIT compiler.
We also welcome developers of other dynamic languages to see what RPython can do for them.
The PyPy 2.7 release supports:
  • x86 machines on most common operating systems (Linux 32/64 bits, Mac OS X 64 bits, Windows 32 bits, OpenBSD, FreeBSD)
  • newer ARM hardware (ARMv6 or ARMv7, with VFPv3) running Linux,
  • big- and little-endian variants of PPC64 running Linux,
  • s390x running Linux

 

What else is new?

(since the releases of PyPy 2.7 and 3.3 at the end of 2016)
There are many incremental improvements to RPython and PyPy, the complete listing is here.
 
Please update, and continue to help us make PyPy better.

Cheers, The PyPy team

Saturday, March 4, 2017

Leysin Winter Sprint Summary

Today is the last day of our yearly sprint event in Leysin. We had lots of ideas on how to enhance the current state of PyPy, we went skiing and had interesting discussions around virtual machines, the Python ecosystem, and other real world problems.
 

Why don't you join us next time?

A usual PyPy sprints day goes through the following stages:

  1.  Planning Session: Tasks from previous days that have seen progress or are completed are noted in a shared document. Everyone adds new tasks and then assigns themselves to one or more tasks (usually in pairs). As soon as everybody is happy with their task and has a partner to work with, the planning session is concluded and the work can start.
  2. Discussions: A sprint is a good occasion to discuss difficult and important topics in person. We usually sit down in a separate area in the sprint room and discuss until a) nobody wants to discuss anymore or b) we found a solution to the problem. The good thing is that usally the outcome is b).
  3. Lunch: For lunch we prepare sandwiches and other finger food.
  4. Continue working until dinner, which we eat at a random restaurant in Leysin.
  5. Goto 1 the next day, if sprint has not ended.
Sprints are open to everybody and help newcomers to get started with PyPy (we usually pair you with a developer familiar with PyPy). They are perfect to discuss and find solutions to problems we currently face. If you are eager to join next year, please don't hesitate to register next year around January.
 

Sprint Summary   

Sprint goals included to work on the following topics:
  • Work towards releasing PyPy 3.5 (it will be released soon)
  • CPython Extension (CPyExt) modules on PyPy
  • Have fun in winter sports (a side goal)

Highlights

  • We have spent lots of time debugging and fixing memory issues on CPyExt. In particular, we fixed a serious memory leak where taking a memoryview would prevent numpy arrays from ever being freed. More work is still required to ensure that our GC always releases arrays in a timely manner.
  • Fruitful discussions and progress about how to flesh out some details about the unicode representation in PyPy. Our current goal is to use utf-8 as the unicode representation internally and have fast vectorized operations (indexing, check if valid, ...).
  • PyPy will participate in GSoC 2017 and we will try to allocate more resources to that than last year.
  • Profile and think about some details how to reduce the starting size of the interpreter. The starting point would be to look at the parser and reduce the amount of strings to keep alive.
  • Found a topic for a student's master thesis: correctly freeing cpyext reference cycles.
  • Run lots of Python3 code on top of PyPy3 and resolve issues we found along the way.
  • Initial work on making RPython thread-safe without a GIL.

List of attendees

- Stefan Beyer
- Antonio Cuni
- Maciej Fijalkowski
- Manuel Jacob
- Ronan Lamy
- Remi Meier
- Richard Plangger
- Armin Rigo
- Robert Zaremba
 
 


We would like to thank our donors for the continued support of the PyPy project and we looking forward to next years sprint in Leysin.

The PyPy Team





Wednesday, March 1, 2017

Async HTTP benchmarks on PyPy3

Hello everyone,

Since Mozilla announced funding, we've been working quite hard on delivering you a working Python 3.5.
 
We are almost ready to release an alpha version of PyPy 3.5. Our goal is to release it shortly after the sprint. Many modules have already been ported and  it can probably run many Python 3 programs already. We are happy to receive any feedback after the next release. 

To show that the heart (asyncio) of Python 3 is already working we have prepared some benchmarks. They are done by Paweł Piotr Przeradowski @squeaky_pl for a HTTP workload on serveral asynchronous IO libraries, namely the relatively new asyncio and curio libraries and the battle-tested tornado, gevent and Twisted libraries. To see the benchmarks check out https://github.com/squeaky-pl/zenchmarks and the instructions for reproducing can be found inside README.md in the repository. Raw results can be obtained from https://github.com/squeaky-pl/zenchmarks/blob/master/results.csv.

The purpose of the presented benchmarks is showing that the upcoming PyPy release is already working with unmodified code that runs on CPython 3.5. PyPy also manages to make them run significantly faster.

The benchmarks consist of HTTP servers implemented on the top of the mentioned libraries. All the servers are single-threaded relying on underlying event loops to provide concurrency. Access logging was disabled to exclude terminal I/O from the results. The view code consists of a lookup in a dictionary mapping ASCII letters to verses from the famous Zen of Python. If a verse is found the view returns it, otherwise a 404 Not Found response is served. The 400 Bad Request and 500 Internal Server Error cases are also handled.

The workload was generated with the wrk HTTP benchmarking tool. It is run with one thread opening up to 100 concurrent connections for 2 seconds and repeated 1010 times to get consecutive measures. There is a Lua script provided that instructs wrk to continuously send 24 different requests that hit different execution paths (200, 404, 400) in the view code. Also it is worth noting that wrk will only count 200 responses as successful so the actual request per second throughput is higher.

For your convenience all the used libraries versions are vendored into the benchmark repository. There is also a precompiled portable version of wrk provided that should run on any reasonably recent (10 year old or newer) Linux x86_64 distribution. The benchmark was performed on a public cloud scaleway x86_64 server launched in a Paris data center. The server was running Ubuntu 16.04.01 LTS and reported Intel(R) Xeon(R) CPU D-1531 @ 2.20GHz CPU. CPython 3.5.2 (shipped by default in Ubuntu) was benchmarked against a pypy-c-jit-90326-88ef793308eb-linux64 snapshot of the 3.5 compatibility branch of PyPy.

 
 
 
 
We want to thank Mozilla for supporting our work!

Cheers,
fijal, squeaky_pl and the PyPy Team

Tuesday, January 24, 2017

Leysin Winter Sprint: 25/26th Feb. - 4th March 2017

The next PyPy sprint will be in Leysin, Switzerland, for the twelveth time. This is a fully public sprint: newcomers and topics other than those proposed below are welcome.

Goals and topics of the sprint

The list of topics is very open.

  • The main topic is Python 3.5 support in PyPy, as most py3.5 contributors should be present. It is also a good topic if you have no or limited experience with PyPy contribution: we can easily find something semi-independent that is not done in py3.5 so far, and do pair-programming with you.
  • Any other topic is fine too: JIT compiler optimizations, CFFI, the RevDB reverse debugger, improving to speed of your program on PyPy, etc.
  • And as usual, the main side goal is to have fun in winter sports :-) We can take a day off (for ski or anything else).

Exact times

Work days: starting 26th Feb (~noon), ending March 4th (~noon).

I have pre-booked the week from Saturday Feb 25th to Saturday March 4th. If it is possible for you to arrive Sunday before mid-afternoon, then you should get a booking from Sunday only. The break day should be around Wednesday.

It is fine to stay a few more days on either side, or conversely to book for a part of that time only.

Location & Accomodation

Leysin, Switzerland, "same place as before".

Let me refresh your memory: both the sprint venue and the lodging will be in a pair of chalets built specifically for bed & breakfast: http://www.ermina.ch/. The place has a good ADSL Internet connection with wireless installed. You can also arrange your own lodging elsewhere (as long as you are in Leysin, you cannot be more than a 15 minutes walk away from the sprint venue).

Please confirm that you are coming so that we can adjust the reservations as appropriate.

The options of rooms are a bit more limited than on previous years because the place for bed-and-breakfast is shrinking; but we should still have enough room for us. The price is around 60 CHF, breakfast included, in shared rooms (3 or 4 people). If there are people that would prefer a double or single room, please contact me and we'll see what choices you have. There are also a choice of hotels in Leysin.

Please register by Mercurial:

https://bitbucket.org/pypy/extradoc/
https://bitbucket.org/pypy/extradoc/raw/extradoc/sprintinfo/leysin-winter-2017/

or on the pypy-dev mailing list if you do not yet have check-in rights:

http://mail.python.org/mailman/listinfo/pypy-dev

You need a Swiss-to-(insert country here) power adapter. There will be some Swiss-to-EU adapters around, and at least one EU-format power strip.

Saturday, November 12, 2016

PyPy2.7 v5.6 released - stdlib 2.7.12 support, C-API improvements, and more

We have released PyPy2.7 v5.6 [0], about two months after PyPy2.7 v5.4. This new PyPy2.7 release includes the upstream stdlib version 2.7.12.

We continue to make incremental improvements to our C-API compatibility layer (cpyext). We pass all but 12 of the over-6000 tests in the upstream NumPy test suite, and have begun examining what it would take to support Pandas and PyQt.

Work proceeds at a good pace on the PyPy3.5 version due to a grant from the Mozilla Foundation, and some of those changes have been backported to PyPy2.7 where relevant.

The PowerPC and s390x backend have been enhanced with the capability to use SIMD instructions for micronumpy loops.

We changed timeit to now report average +/- standard deviation, which is better than the misleading minimum value reported in CPython.

We now support building PyPy with OpenSSL 1.1 in our built-in _ssl module, as well as maintaining support for previous versions.

CFFI has been updated to 1.9, improving an already great package for interfacing with C.

As always, this release fixed many issues and bugs raised by the growing community of PyPy users. We strongly recommend updating. You can download the PyPy2.7 v5.6 release here:
Downstream packagers have been hard at work. The Debian package is already available, and the portable PyPy versions are also ready, for those who wish to run PyPy on other Linux distributions like RHEL/Centos 5.

We would like to thank our donors for the continued support of the PyPy project.

We would also like to thank our contributors and encourage new people to join the project. PyPy has many layers and we need help with all of them: PyPy and RPython documentation improvements, tweaking popular modules to run on pypy, or general help with making RPython’s JIT even better.

What is PyPy?

PyPy is a very compliant Python interpreter, almost a drop-in replacement for CPython 2.7. It’s fast (PyPy and CPython 2.7.x performance comparison) due to its integrated tracing JIT compiler.
We also welcome developers of other dynamic languages to see what RPython can do for them.
This release supports:
  • x86 machines on most common operating systems (Linux 32/64 bits, Mac OS X 64 bits, Windows 32 bits, OpenBSD, FreeBSD)
  • newer ARM hardware (ARMv6 or ARMv7, with VFPv3) running Linux,
  • big- and little-endian variants of PPC64 running Linux,
  • s390x running Linux

What else is new?

(since the release of PyPy 5.4 in August, 2016)
There are many incremental improvements to RPython and PyPy, the complete listing is here.
 
Please update, and continue to help us make PyPy better.

Cheers, The PyPy team

[0] We skipped 5.5 since we share a code base with PyPy3, and PyPy3.3-v.5.5-alpha was released last month