Development build. This page describes main, not a released package. The latest published Lodestar.Gpu is 0.1.0 — read its documentation.

Home › Gpu › GPU kernels

Uploads a host block to the accelerator.

public static DeviceDenseBlock Upload(GpuContext context, ReadOnlySpan<double> values, int rowCount, int columnCount)

Parameters — context is the accelerator to upload to. values holds rowCount × columnCount numbers, row-major. rowCount and columnCount are the block's shape, and the product of the two has to be exactly values.Length.

Returns — DeviceDenseBlock, owning device memory the caller disposes.

Exceptions — ArgumentNullException when context is null; ArgumentOutOfRangeException when a dimension is below 1; ArgumentException when values is not exactly the block.

Example — the shape a caller writes.

using Lodestar.Gpu.Compute;

using var context = GpuContext.Create(preferCpu: true);
using var block = DeviceDenseBlock.Upload(context, [1.0, 2.0], rowCount: 1, columnCount: 2);

int wide = block.ColumnCount;  // => 2

Remarks — this is the first crossing of the bus in a chain, and the only one a caller writes by hand. Everything after it comes from a kernel that leaves its result resident, which is the overload on TiledSparseDenseProduct.Multiply that takes a block and returns one.

Applies to — net10.0, netstandard2.1.

See also — DeviceDenseBlock.Download.