US 12,118,644 B2 · Data loading method and apparatus for convolution operation

Abstract

Disclosed are a data loading method and apparatus for a convolution operation, a computer device, a storage medium and a computer program product. The method includes: splitting a feature image to be loaded into the cache unit into a plurality of sub-feature images; determining a target cache line corresponding to each of the sub-feature images in each of the cache lines according to a positional relationship of each of the sub-feature images in the feature image; wherein target cache lines corresponding to at least two sub-feature images with the same positional relationship are located in the same cache set, and target cache lines corresponding to at least two sub-feature images with an adjacent positional relationship are located in different cache sets; loading a data content of each of the sub-feature images into the target cache line corresponding to each of the sub-feature images.

Claims14

Dependent claims

A data loading method for a convolution operation, applied to a cache unit which is formed by a plurality of cache banks arranged in an array, the same row of cache banks forming a cache line, a plurality of cache lines forming a cache set, the data loading method comprising:

splitting a feature image to be loaded into the cache unit into a plurality of sub-feature images;
determining a target cache line corresponding to each of the sub-feature images in each of the cache lines according to a positional relationship of each of the sub-feature images in the feature image;
wherein target cache lines corresponding to at least two sub-feature images with the same positional relationship are located in the same cache set, and target cache lines corresponding to at least two sub-feature images with an adjacent positional relationship are located in different cache sets;
loading a data content of each of the sub-feature images into the target cache line corresponding to each of the sub-feature images.

1 more element

A data loading apparatus for a convolution operation, applied to a cache unit which is formed by a plurality of cache banks arranged in an array, the cache unit comprising at least one cache set, each cache set comprising at least one cache line, the apparatus comprising:

a split module, configured to split a feature image to be loaded into the cache unit into a plurality of sub-feature images;
a determination module, configured to determine a target cache line corresponding to each of the sub-feature images in the at least one cache line according to a positional relationship of each of the sub-feature images in the feature image, wherein target cache lines corresponding to at least two sub-feature images with the same positional relationship are located in the same cache set, and target cache lines corresponding to at least two sub-feature images with an adjacent positional relationship are located in different cache sets;
a loading module, configured to load data contents of the sub-feature images into target cache lines corresponding to the sub-feature images.

Dependencies are inferred from the claim text.

Data loading method and apparatus for convolution operation

Life of this patent

Data loading method and apparatus for convolution operation

Life of this patent

Abstract

Claims14