Cool_Tools/bemaniutils
1
0
Fork 0
mirror of synced 2025-02-17 19:19:19 +01:00
Code Releases Activity

Comments for cythonized code, and dropping python implementation of blend from compiling.

Browse Source
This commit is contained in:
Jennifer Taylor 2021-05-31 18:14:22 +00:00
parent a4746f2934
commit 38f0483cb5
1 changed files with 22 additions and 6 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

28
setup.py
View File

@ -14,18 +14,24 @@ def extensions():
return [
*cythonize(
[
# Hot code for anything constructing or parsing a remote game packet.
Extension(
"bemani.protocol.binary",
[
"bemani/protocol/binary.py",
]
),
# Even though we have a C++ implementation of this, some of the code
# is still used as a wrapper to the C++ implementation and it is very
# hot code (almost every packet touches this).
Extension(
"bemani.protocol.lz77",
[
"bemani/protocol/lz77.py",
]
),
# Alternative, orders of magnitude faster, memory-unsafe version of
# LZ77 which drastically speeds up packet processing time.
Extension(
"bemani.protocol.lz77cpp",
[
@ -35,36 +41,42 @@ def extensions():
extra_compile_args=["-std=c++14"],
extra_link_args=["-std=c++14"],
),
# Every single backend service uses this class for construction and
# parsing, so compiling this makes sense.
Extension(
"bemani.protocol.node",
[
"bemani/protocol/node.py",
]
),
# This is the top-level protocol marshall which gets touched at least
# once per packet, so its worth it to squeeze more speed out of this.
Extension(
"bemani.protocol.protocol",
[
"bemani/protocol/protocol.py",
]
),
# This is used to implement a convenient way of parsing/creating binary
# data and it is memory-safe accessses of bytes so it is necessarily
# a bottleneck.
Extension(
"bemani.protocol.stream",
[
"bemani/protocol/stream.py",
]
),
# This gets used less frequently (only on the oldest games) but it is
# still worth it to get a bit of a speed boost by compiling.
Extension(
"bemani.protocol.xml",
[
"bemani/protocol/xml.py",
]
),
Extension(
"bemani.format.afp.blend.blend",
[
"bemani/format/afp/blend/blend.py",
]
),
# This is a memory-unsafe, orders of magnitude faster threaded implementation
# of the pure python blend code which takes rendering rough animations down
# from over an hour to around a minute.
Extension(
"bemani.format.afp.blend.blendcpp",
[
@ -75,12 +87,16 @@ def extensions():
extra_compile_args=["-std=c++14"],
extra_link_args=["-std=c++14"],
),
# These types include operations such as matrix math and color conversion so
# it is worth it to speed this up when rendering animations.
Extension(
"bemani.format.afp.types.generic",
[
"bemani/format/afp/types/generic.py",
]
),
# DXT is slow enough that it might be worth it to write a C++ implementation of
# this at some point, but for now we squeeze a bit of speed out of this by compiling.
Extension(
"bemani.format.dxt",
[