Method for embedding rows prefetching in recommendation models

A method comprising:

• identifying, by a prefetch engine based on a user query, a plurality of data items targeted by the user query, wherein two or more of the plurality of data items are stored in a non-contiguous manner in one or more data storage areas;
• retrieving, from a prefetch data storage by the prefetch engine, a first group of data items comprising one or more data items of the plurality of data items;
• and
• sending, by the prefetch engine, the first group of data items to a data processing stage, prior to a second group of data items being fetched from the one or more data storage areas, wherein the second group of data items comprises one or more data items of the plurality of data items different from the first group of data items.

A computing system comprising:

• a memory configured to store instructions of one or more tasks and source data to be processed by the one or more tasks;
• an integrated circuit configured to execute the instructions using the source data, wherein the integrated circuit comprises:
• a prefetch engine configured to:
• identify, based on a user query, a plurality of data items targeted by the user query, wherein two or more of the plurality of data items are stored in a non-contiguous manner in one or more data storage areas;
• retrieve, from a prefetch data storage, a first group of data items comprising one or more data items of the plurality of data items;
• and
• send the first group of data items to a data processing stage, prior to a second group of data items being fetched from the one or more data storage areas, wherein the second group of data items comprises one or more data items of the plurality of data items different from the first group of data items.