gnprice

$ python -m timeit -v -s 'import bench_fnclass as b' 'b.as_fn()'
10 loops -> 0.000801 secs
100 loops -> 0.00851 secs
1000 loops -> 0.113 secs
10000 loops -> 7.75 secs
raw times: 7.2 7.22 7.24
10000 loops, best of 3: 720 usec per loop

$ python -V
Python 2.6.6

gnprice

# Times are PyPy / CPython.
#   Python 2.7.0 (e837df3968a1, Mar 15 2011, 03:00:18)
#   [PyPy 1.5.0-alpha0 with GCC 4.4.3] on linux2
#   Python 2.6.6 (r266:84292, Sep 15 2010, 16:22:56) 
#   [GCC 4.4.5] on linux2
# The CPython is 64-bit, the PyPy 32-bit, which I hope is not
# confounding anything.  Both running on the same machine.

# Times measured with
#  $ $PYTHON -m timeit -n10000 -r3 -s 'import bench_fnclass as b, gc; gc.enable()' "b.$func()"

gnprice

From 98fc9cd0b28921dc02ee4c366c96e4caeb7d8475 Mon Sep 17 00:00:00 2001
From: Greg Price <[email protected]>
Date: Wed, 31 Oct 2012 15:53:59 -0700
Subject: [PATCH] Suppress SIGPROF

We don't handle it well -- the exact reason is unclear, but threading
seems to trigger it -- so just squelch it.

Building with -pg is necessary for getting call-graph information with
e.g. the Linux 'perf' profiler.  But -pg also causes the libc startup

gnprice

... actually, see https://paper.dropbox.com/doc/Mypy-strict-optional-UQPcKDn5VDdSXQme37o1G for much more easily-editable notes.

gnprice

<?xml version="1.0"?>
<!DOCTYPE fontconfig SYSTEM "fonts.dtd">
<!--

    Problem: After installing the Noto fonts (on my Debian `stretch`
      system, with `apt install fonts-noto`), Japanese text in my
      browser ends up displayed in the "Medium" style of the font
      rather than the "Regular" style.  The "Medium" style is too
      heavy for normal text.

gnprice

# This is a snapshot of my ~/.gitconfig file, minus a few bits
# that are obsolete or non-reusable.
#
# For explanation of each setting, see `git help config`
# or https://git-scm.com/docs/git-config .
#
[user]
	name = Greg Price
	email = [email protected]
[alias]

gnprice

pragma solidity ^0.5.0;

contract LowestUnique {
  uint256 ante = 0.01 ether;

  enum Stage { INIT, IN_PROGRESS, END }
  Stage private stage;
  uint private commitDeadline;
  uint private revealDeadline;

gnprice

{
  "\UF729"  = moveToBeginningOfParagraph:; // home
  "\UF72B"  = moveToEndOfParagraph:; // end
  "$\UF729" = moveToBeginningOfParagraphAndModifySelection:; // shift-home
  "$\UF72B" = moveToEndOfParagraphAndModifySelection:; // shift-end
  "^\UF729" = moveToBeginningOfDocument:; // ctrl-home
  "^\UF72B" = moveToEndOfDocument:; // ctrl-end
  "^$\UF729" = moveToBeginningOfDocumentAndModifySelection:; // ctrl-shift-home
  "^$\UF72B" = moveToEndOfDocumentAndModifySelection:; // ctrl-shift-end

gnprice

These are benchmark results from the "pyperformance" benchmark suite:
  https://pyperformance.readthedocs.io/
on CPython built with various options for optimization.

For context, including the specific optimizations involved, see:
  https://github.com/NixOS/nixpkgs/pull/84072

gnprice

{ attr
}:

# Scaffold to build just one package with modified stdenv scripts.
#
# Dependencies come unmodified from <nixpkgs>, including the stdenv's
# toolchain.  This is useful for (a) iterating quickly on changes to
# stdenv scripts (like setup.sh), and (b) more easily comparing the
# build log against a baseline, by keeping inputs' paths the same so
# that any changes caused by changes in the scripts' behavior stand out.