This library helps to interact with NVIDIA GPUs from zig. Provides high level interface to communicate with GPU. It can detect cuda installation and link to a project's binary on Linux/MacOS. Check Customization to give cuda manual path.
- Memory Allocation in GPU with defined size.
- Copying data from host to gpu and viceversa.
- Compiling (.cu) and loading kernels (.ptx) both from file and text.
- Running kernels with grid/blocks/threads configuration.
- Generate random numbers
Check test.zig file for code samples.
Scroll below to go through an example of incrementing each value in an array parallely using GPU.
Download and save the library path in build.zig.zon
file by running
zig fetch --save https://github.com/akhildevelops/cudaz/tarball/master
Add cudaz module in your project's build.zig
file that will link to your project's binary.
//build.zig
const std = @import("std");
pub fn build(b: *std.Build) !void {
// exe points to main.zig that uses cudaz
const exe = b.addExecutable(.{ .name = "main", .root_source_file = .{ .path = "src/main.zig" }, .target = b.host });
// Point to cudaz dependency
const cudaz_dep = b.dependency("cudaz", .{});
// Fetch and add the module from cudaz dependency
const cudaz_module = cudaz_dep.module("cudaz");
exe.root_module.addImport("cudaz", cudaz_module);
// Dynamically link to libc, cuda, nvrtc
exe.linkLibC();
exe.linkSystemLibrary("cuda");
exe.linkSystemLibrary("nvrtc");
// Run binary
const run = b.step("run", "Run the binary");
const run_step = b.addRunArtifact(exe);
run.dependOn(&run_step.step);
}
// src/main.zig
const std = @import("std");
const Cuda = @import("cudaz");
const CuDevice = Cuda.Device;
const CuCompile = Cuda.Compile;
const CuLaunchConfig = Cuda.LaunchConfig;
// Cuda Kernel
const increment_kernel =
\\extern "C" __global__ void increment(float *out)
\\{
\\ int i = blockIdx.x * blockDim.x + threadIdx.x;
\\ out[i] = out[i] + 1;
\\}
;
pub fn main() !void {
// Initialize allocator
var GP = std.heap.GeneralPurposeAllocator(.{}){};
defer _ = GP.deinit();
const allocator = GP.allocator();
// Initialize GPU
const device = try CuDevice.default();
defer device.free();
// Copy data from host to GPU
const data = [_]f32{ 1.2, 2.8, 0.123 };
const cu_slice = try device.htodCopy(f32, &data);
defer cu_slice.free();
// Compile and load the Kernel
const ptx = try CuCompile.cudaText(increment_kernel, .{}, allocator);
defer allocator.free(ptx);
const module = try CuDevice.loadPtxText(ptx);
const function = try module.getFunc("increment");
// Run the kernel on the data
try function.run(.{&cu_slice.device_ptr}, CuLaunchConfig{ .block_dim = .{ 3, 1, 1 }, .grid_dim = .{ 1, 1, 1 }, .shared_mem_bytes = 0 });
// Retrieve incremented data back to the system
const incremented_arr = try CuDevice.syncReclaim(f32, allocator, cu_slice);
defer incremented_arr.deinit();
}
For running above code system refer to the example project: increment
- It is intelligent to identify and link to installed cuda libraries. If needed, provide cuda installation path manually by mentioning build parameter
zig build -DCUDA_PATH=<cuda_folder>
.
Inspired from Rust Cuda library: https://github.com/coreylowman/cudarc/tree/main