Skip to content

BrokenSource/TurboPipe

Folders and files

NameName
Last commit message
Last commit date

Latest commit

Β 

History

54 Commits
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 

Repository files navigation

Important

Also check out ShaderFlow, where TurboPipe shines! πŸ˜‰

TurboPipe

Faster ModernGL Buffers inter-process data transfers for subprocesses


πŸ”₯ Description

TurboPipe speeds up sending raw bytes from moderngl.Buffer objects primarily to FFmpeg subprocess

The optimizations involved are:

  • Zero-copy: Avoid unnecessary memory copies or allocation (intermediate buffer.read())
  • C++: The core of TurboPipe is written in C++ for speed, efficiency and low-level control
  • Chunks: Write in chunks of 4096 bytes (RAM page size), so the hardware is happy (Unix)
  • Threaded:
    • Doesn't block Python code execution, allows to render next frame
    • Decouples the main thread from the I/O thread for performance

βœ… Don't worry, there's proper safety in place. TurboPipe will block Python if a memory address is already queued for writing, and guarantees order of writes per file-descriptor. Just call .sync() when done πŸ˜‰


πŸ“¦ Installation

It couldn't be easier! Just install the turbopipe package from PyPI:

# With pip (https://pip.pypa.io/)
pip install turbopipe

# With Poetry (https://python-poetry.org/)
poetry add turbopipe

# With PDM (https://pdm-project.org/en/latest/)
pdm add turbopipe

# With Rye (https://rye.astral.sh/)
rye add turbopipe

πŸš€ Usage

See also the Examples folder for comparisons, and ShaderFlow usage of it!

import subprocess

import moderngl
import turbopipe

# Create ModernGL objects
ctx = moderngl.create_standalone_context()
buffers = [ctx.buffer(reserve=1920*1080*3) for _ in range(2)]

# Make sure resolution, pixel format matches!
ffmpeg = subprocess.Popen(
    'ffmpeg -f rawvideo -pix_fmt rgb24 -r 60 -s 1920x1080 -i - -f null -'.split(),
    stdin=subprocess.PIPE
)

# Rendering loop of yours (eg. 1m footage)
for frame in range(60 * 60):
    buffer = buffers[frame % len(buffer)]
    turbopipe.sync(buffer)
    fbo.read_into(buffer)
    turbopipe.pipe(buffer, ffmpeg.stdin.fileno())

# Finalize writing, encoding
ffmpeg.stdin.close()
turbopipe.close()
ffmpeg.wait()

⭐️ Benchmarks

Note

The tests conditions are as follows:

  • The tests are the average of 3 runs to ensure consistency, with 5 GB of the same data being piped
  • These aren't tests of render speed; but rather the throughput speed of GPU -> CPU -> RAM -> IPC
  • All resolutions are wide-screen (16:9) and have 3 components (RGB) with 3 bytes per pixel (SDR)
  • The data is a random noise per-buffer between 128-135. So, multi-buffers runs are a noise video
  • Multi-buffer cycles through a list of buffer (eg. 1, 2, 3, 1, 2, 3... for 3-buffers)
  • All FFmpeg outputs are scrapped with -f null - to avoid any disk I/O bottlenecks
  • The gain column is the percentage increase over the standard method
  • When x264 is Null, no encoding took place (passthrough)
  • The test cases emoji signifies:
    • 🐒: Standard ffmpeg.stdin.write(buffer.read()) on just the main thread, pure Python
    • πŸš€: Threaded ffmpeg.stdin.write(buffer.read()) with a queue (similar to turbopipe)
    • πŸŒ€: The magic of turbopipe.pipe(buffer, ffmpeg.stdin.fileno())

Also see benchmark.py for the implementation

βœ… Check out benchmarks in a couple of systems below:

πŸ“¦ TurboPipe v1.0.4:

Desktop β€’ (AMD Ryzen 9 5900x) β€’ (NVIDIA RTX 3060 12 GB) β€’ (DDR4 2x32 GB 3200 MT/s) β€’ (Arch Linux)

Note: I have noted inconsistency across tests, specially at lower resolutions. Some 720p runs might peak at 2900 fps and stay there, while others are limited by 1750 fps. I'm not sure if it's the Linux EEVDF scheduler, or CPU Topology that causes this. Nevertheless, results are stable on Windows 11 on the same machine.

720p x264 Buffers Framerate Bandwidth Gain
🐒 Null 1 882 fps 2.44 GB/s
πŸš€ Null 1 793 fps 2.19 GB/s -10.04%
πŸŒ€ Null 1 1911 fps 5.28 GB/s 116.70%
🐒 Null 4 857 fps 2.37 GB/s
πŸš€ Null 4 891 fps 2.47 GB/s 4.05%
πŸŒ€ Null 4 2309 fps 6.38 GB/s 169.45%
🐒 ultrafast 4 714 fps 1.98 GB/s
πŸš€ ultrafast 4 670 fps 1.85 GB/s -6.10%
πŸŒ€ ultrafast 4 1093 fps 3.02 GB/s 53.13%
🐒 slow 4 206 fps 0.57 GB/s
πŸš€ slow 4 208 fps 0.58 GB/s 1.37%
πŸŒ€ slow 4 214 fps 0.59 GB/s 3.93%
1080p x264 Buffers Framerate Bandwidth Gain
🐒 Null 1 410 fps 2.55 GB/s
πŸš€ Null 1 399 fps 2.48 GB/s -2.60%
πŸŒ€ Null 1 794 fps 4.94 GB/s 93.80%
🐒 Null 4 390 fps 2.43 GB/s
πŸš€ Null 4 391 fps 2.43 GB/s 0.26%
πŸŒ€ Null 4 756 fps 4.71 GB/s 94.01%
🐒 ultrafast 4 269 fps 1.68 GB/s
πŸš€ ultrafast 4 272 fps 1.70 GB/s 1.48%
πŸŒ€ ultrafast 4 409 fps 2.55 GB/s 52.29%
🐒 slow 4 115 fps 0.72 GB/s
πŸš€ slow 4 118 fps 0.74 GB/s 3.40%
πŸŒ€ slow 4 119 fps 0.75 GB/s 4.34%
1440p x264 Buffers Framerate Bandwidth Gain
🐒 Null 1 210 fps 2.33 GB/s
πŸš€ Null 1 239 fps 2.64 GB/s 13.84%
πŸŒ€ Null 1 534 fps 5.91 GB/s 154.32%
🐒 Null 4 219 fps 2.43 GB/s
πŸš€ Null 4 231 fps 2.56 GB/s 5.64%
πŸŒ€ Null 4 503 fps 5.56 GB/s 129.75%
🐒 ultrafast 4 141 fps 1.56 GB/s
πŸš€ ultrafast 4 150 fps 1.67 GB/s 6.92%
πŸŒ€ ultrafast 4 226 fps 2.50 GB/s 60.37%
🐒 slow 4 72 fps 0.80 GB/s
πŸš€ slow 4 71 fps 0.79 GB/s -0.70%
πŸŒ€ slow 4 75 fps 0.83 GB/s 4.60%
2160p x264 Buffers Framerate Bandwidth Gain
🐒 Null 1 81 fps 2.03 GB/s
πŸš€ Null 1 107 fps 2.67 GB/s 32.26%
πŸŒ€ Null 1 213 fps 5.31 GB/s 163.47%
🐒 Null 4 87 fps 2.18 GB/s
πŸš€ Null 4 109 fps 2.72 GB/s 25.43%
πŸŒ€ Null 4 212 fps 5.28 GB/s 143.72%
🐒 ultrafast 4 59 fps 1.48 GB/s
πŸš€ ultrafast 4 67 fps 1.68 GB/s 14.46%
πŸŒ€ ultrafast 4 95 fps 2.39 GB/s 62.66%
🐒 slow 4 37 fps 0.94 GB/s
πŸš€ slow 4 43 fps 1.07 GB/s 16.22%
πŸŒ€ slow 4 44 fps 1.11 GB/s 20.65%
Desktop β€’ (AMD Ryzen 9 5900x) β€’ (NVIDIA RTX 3060 12 GB) β€’ (DDR4 2x32 GB 3200 MT/s) β€’ (Windows 11)
720p x264 Buffers Framerate Bandwidth Gain
🐒 Null 1 981 fps 2.71 GB/s
πŸš€ Null 1 1145 fps 3.17 GB/s 16.74%
πŸŒ€ Null 1 1504 fps 4.16 GB/s 53.38%
🐒 Null 4 997 fps 2.76 GB/s
πŸš€ Null 4 1117 fps 3.09 GB/s 12.08%
πŸŒ€ Null 4 1467 fps 4.06 GB/s 47.14%
🐒 ultrafast 4 601 fps 1.66 GB/s
πŸš€ ultrafast 4 616 fps 1.70 GB/s 2.57%
πŸŒ€ ultrafast 4 721 fps 1.99 GB/s 20.04%
🐒 slow 4 206 fps 0.57 GB/s
πŸš€ slow 4 206 fps 0.57 GB/s 0.39%
πŸŒ€ slow 4 206 fps 0.57 GB/s 0.13%
1080p x264 Buffers Framerate Bandwidth Gain
🐒 Null 1 451 fps 2.81 GB/s
πŸš€ Null 1 542 fps 3.38 GB/s 20.31%
πŸŒ€ Null 1 711 fps 4.43 GB/s 57.86%
🐒 Null 4 449 fps 2.79 GB/s
πŸš€ Null 4 518 fps 3.23 GB/s 15.48%
πŸŒ€ Null 4 614 fps 3.82 GB/s 36.83%
🐒 ultrafast 4 262 fps 1.64 GB/s
πŸš€ ultrafast 4 266 fps 1.66 GB/s 1.57%
πŸŒ€ ultrafast 4 319 fps 1.99 GB/s 21.88%
🐒 slow 4 119 fps 0.74 GB/s
πŸš€ slow 4 121 fps 0.76 GB/s 2.46%
πŸŒ€ slow 4 121 fps 0.75 GB/s 1.90%
1440p x264 Buffers Framerate Bandwidth Gain
🐒 Null 1 266 fps 2.95 GB/s
πŸš€ Null 1 308 fps 3.41 GB/s 15.87%
πŸŒ€ Null 1 402 fps 4.45 GB/s 51.22%
🐒 Null 4 276 fps 3.06 GB/s
πŸš€ Null 4 307 fps 3.40 GB/s 11.32%
πŸŒ€ Null 4 427 fps 4.73 GB/s 54.86%
🐒 ultrafast 4 152 fps 1.68 GB/s
πŸš€ ultrafast 4 156 fps 1.73 GB/s 3.02%
πŸŒ€ ultrafast 4 181 fps 2.01 GB/s 19.36%
🐒 slow 4 77 fps 0.86 GB/s
πŸš€ slow 4 79 fps 0.88 GB/s 3.27%
πŸŒ€ slow 4 80 fps 0.89 GB/s 4.86%
2160p x264 Buffers Framerate Bandwidth Gain
🐒 Null 1 134 fps 3.35 GB/s
πŸš€ Null 1 152 fps 3.81 GB/s 14.15%
πŸŒ€ Null 1 221 fps 5.52 GB/s 65.44%
🐒 Null 4 135 fps 3.36 GB/s
πŸš€ Null 4 151 fps 3.76 GB/s 11.89%
πŸŒ€ Null 4 220 fps 5.49 GB/s 63.34%
🐒 ultrafast 4 66 fps 1.65 GB/s
πŸš€ ultrafast 4 70 fps 1.75 GB/s 6.44%
πŸŒ€ ultrafast 4 82 fps 2.04 GB/s 24.31%
🐒 slow 4 40 fps 1.01 GB/s
πŸš€ slow 4 43 fps 1.09 GB/s 9.54%
πŸŒ€ slow 4 44 fps 1.10 GB/s 10.15%
Laptop β€’ (Intel Core i7 11800H) β€’ (NVIDIA RTX 3070) β€’ (DDR4 2x16 GB 3200 MT/s) β€’ (Windows 11)

Note: Must select NVIDIA GPU on their Control Panel instead of Intel iGPU

720p x264 Buffers Framerate Bandwidth Gain
🐒 Null 1 786 fps 2.17 GB/s
πŸš€ Null 1 903 fps 2.50 GB/s 14.91%
πŸŒ€ Null 1 1366 fps 3.78 GB/s 73.90%
🐒 Null 4 739 fps 2.04 GB/s
πŸš€ Null 4 855 fps 2.37 GB/s 15.78%
πŸŒ€ Null 4 1240 fps 3.43 GB/s 67.91%
🐒 ultrafast 4 484 fps 1.34 GB/s
πŸš€ ultrafast 4 503 fps 1.39 GB/s 4.10%
πŸŒ€ ultrafast 4 577 fps 1.60 GB/s 19.37%
🐒 slow 4 143 fps 0.40 GB/s
πŸš€ slow 4 145 fps 0.40 GB/s 1.78%
πŸŒ€ slow 4 151 fps 0.42 GB/s 5.76%
1080p x264 Buffers Framerate Bandwidth Gain
🐒 Null 1 358 fps 2.23 GB/s
πŸš€ Null 1 427 fps 2.66 GB/s 19.45%
πŸŒ€ Null 1 566 fps 3.53 GB/s 58.31%
🐒 Null 4 343 fps 2.14 GB/s
πŸš€ Null 4 404 fps 2.51 GB/s 17.86%
πŸŒ€ Null 4 465 fps 2.89 GB/s 35.62%
🐒 ultrafast 4 191 fps 1.19 GB/s
πŸš€ ultrafast 4 207 fps 1.29 GB/s 8.89%
πŸŒ€ ultrafast 4 234 fps 1.46 GB/s 22.77%
🐒 slow 4 62 fps 0.39 GB/s
πŸš€ slow 4 67 fps 0.42 GB/s 8.40%
πŸŒ€ slow 4 74 fps 0.47 GB/s 20.89%
1440p x264 Buffers Framerate Bandwidth Gain
🐒 Null 1 180 fps 1.99 GB/s
πŸš€ Null 1 216 fps 2.40 GB/s 20.34%
πŸŒ€ Null 1 264 fps 2.92 GB/s 46.74%
🐒 Null 4 178 fps 1.97 GB/s
πŸš€ Null 4 211 fps 2.34 GB/s 19.07%
πŸŒ€ Null 4 250 fps 2.77 GB/s 40.48%
🐒 ultrafast 4 98 fps 1.09 GB/s
πŸš€ ultrafast 4 110 fps 1.23 GB/s 13.18%
πŸŒ€ ultrafast 4 121 fps 1.35 GB/s 24.15%
🐒 slow 4 40 fps 0.45 GB/s
πŸš€ slow 4 41 fps 0.46 GB/s 4.90%
πŸŒ€ slow 4 43 fps 0.48 GB/s 7.89%
2160p x264 Buffers Framerate Bandwidth Gain
🐒 Null 1 79 fps 1.98 GB/s
πŸš€ Null 1 95 fps 2.37 GB/s 20.52%
πŸŒ€ Null 1 104 fps 2.60 GB/s 32.15%
🐒 Null 4 80 fps 2.00 GB/s
πŸš€ Null 4 94 fps 2.35 GB/s 17.82%
πŸŒ€ Null 4 108 fps 2.70 GB/s 35.40%
🐒 ultrafast 4 41 fps 1.04 GB/s
πŸš€ ultrafast 4 48 fps 1.20 GB/s 17.67%
πŸŒ€ ultrafast 4 52 fps 1.30 GB/s 27.49%
🐒 slow 4 17 fps 0.43 GB/s
πŸš€ slow 4 19 fps 0.48 GB/s 13.16%
πŸŒ€ slow 4 19 fps 0.48 GB/s 13.78%

πŸŒ€ Conclusion

TurboPipe significantly increases the feeding speed of FFmpeg with data, especially at higher resolutions. However, if there's few CPU compute available, or the video is too hard to encode (/slow preset), the gains are insignificant over the other methods (bottleneck). Multi-buffering didn't prove to have an advantage, debugging shows that TurboPipe C++ is often starved of data to write (as the file stream is buffered on the OS most likely), and the context switching, or cache misses, might be the cause of the slowdown.

The theory supports the threaded method being faster, as writing to a file descriptor is a blocking operation for python, but a syscall under the hood, that doesn't necessarily lock the GIL, just the thread. TurboPipe speeds that even further by avoiding an unecessary copy of the buffer data, and writing directly to the file descriptor on a C++ thread. Linux shows a better performance than Windows in the same system after the optimizations, but Windows wins on the standard method.

Interestingly, due either Linux's scheduler on AMD Ryzen CPUs, or their operating philosophy, it was experimentally seen that Ryzen's frenetic thread switching degrades a bit the single thread performance, which can be "fixed" with prepending the command with taskset --cpu 0,2 (not recommended at all), comparatively speaking to Windows performance on the same system (Linux πŸš€ = Windows 🐒). This can also be due the topology of tested CPUs having more than one Core Complex Die (CCD). Intel CPUs seem to stick to the same thread for longer, which makes the Python threaded method often slightly faster.

Personal experience

On realistically loads, like ShaderFlow's default lightweight shader export, TurboPipe increases rendering speed from 1080p260 to 1080p360 on my system, with mid 80% CPU usage than low 60%s. For DepthFlow's default depth video export, no gains are seen, as the CPU is almost saturated encoding at 1080p130.


πŸ“š Future work

  • Disable/investigate performance degradation on Windows iGPUs
  • Improve the thread synchronization and/or use a ThreadPool
  • Maybe use mmap instead of chunks writing on Linux
  • Split the code into a libturbopipe? Not sure where it would be useful πŸ˜