Masking for coarse grained reconfigurable architecture

A system comprising:

• multiple memory-compute nodes coupled to provide a reconfigurable compute fabric, wherein each of the nodes comprises a hybrid threading processor and hybrid threading fabric, and wherein the hybrid threading fabric of each node comprises multiple memory-compute tiles and each of the memory-compute tiles includes respective processing and storage elements;
• a first node comprising N memory-compute tiles, wherein the N memory-compute tiles are coupled using a first portion of a synchronous compute fabric, and wherein operations performed by the respective processing and storage elements of the N memory-compute tiles are configured to be enabled or disabled based on information in a mask field of data propagated through the first portion of the synchronous compute fabric;
• and
• a first tile of the N memory-compute tiles, the first tile comprising:
• a first processing element;
• and
• an operand multiplexer including a mask bit input and one or more data bit inputs, wherein an output of the operand multiplexer is coupled to an input of the first processing element;
• wherein the first processing element and the operand multiplexer are configured to convert one bit of information from the mask field to a corresponding mask byte of information for further propagation through the first portion of the synchronous compute fabric.

A method comprising:

• receiving, via an asynchronous compute fabric, first data for a synchronous compute thread at a memory interface of a first node of a memory-compute system, wherein the memory interface is coupled to the asynchronous compute fabric and a synchronous compute fabric of the system, and wherein the first data includes a mask field and a payload field;
• and
• at a first memory-compute tile:
• receiving the first data from the memory interface;
• based on information in the mask field of the first data, selectively enabling a processing element of the first memory-compute tile and operating on information in the payload field of the first data to execute a portion of the compute thread;
• generating second data with a mask field and a payload field, wherein the second data and the first data comprise different information in one or both of the mask field and the payload field;
• and
• communicating the second data to a second memory-compute tile using the synchronous compute fabric.