Note: This project is in the prototype phase, and is in development and tested with. Further optimization of this server for high performance use will be done at a later time.
A blazingly fast GPU accelerated pixelflut (video) server in Rust.
- Blazingly fast pixelflut rendering
- GPU accelerated
- Highly concurrent, to support many connections
- Linux, Windows and macOS
- Optional binary PX command for reduced bandwidth requirements (enabled by default).
For installation, Git and Rust cargo are required. Install the latest version of Rust with rustup.
Then, clone and install pixelpwnr-server
with:
# Clone the project
git clone https://github.com/timvisee/pixelpwnr-server.git
cd pixelpwnr-server
# Install pixelpwnr server
cargo install -f
# Start using pixelpwnr server
pixelpwnr-server --help
# or run it directly from Cargo
cargo run --release -- --help
Or just build it and invoke the binary directly (Linux/macOS):
# Clone the project
git clone https://github.com/timvisee/pixelpwnr-server.git
cd pixelpwnr-server
# Build the project (release version)
cargo build --release
# Start using pixelpwnr-server
./target/release/pixelpwnr-server --help
This implementation adds a new command to the protocol, which is laid out as follows:
PBxyrgba
where:
x
andy
are Little-Endian u16 values describing the X and Y coordinate of the pixel to set.r
,g
,b
anda
are single-byte values describing the R, G, B, and A components of the color to set the pixel to.- It is important to note that this command does not end in a newline. Appending a newline simply causes the server to interpret that newline as an empty command (which is fine).
If you wish to disable the binary pixel command, pass the --no-default-features
flag to cargo
- Rust (v1.31 or higher)
- Build essentials (Ubuntu package:
build-essential
) freetype2
development files (Ubuntu package:libfreetype6-dev
)
Here are some points that help with the pixelflut server performance, under heavy load:
- Use a
--release
build. - Use a CPU with as many cores as possible.
- Use a fast Ethernet connection, preferably 10Gb/s+.
- Use a dedicated graphics card.
- Use a Linux machine.
- Increase the file descriptor limit (on Linux).
- Quit as many other running programs.
This project is released under the GNU GPL-3.0 license. Check out the LICENSE file for more information.