Managing shared memory used by compute nodes

A method for managing shared memory used by a plurality of compute nodes, comprising:

• allocating memory in a shared globally addressable memory using a memory write module;
• writing read-only data to allocated memory of the shared globally addressable memory;
• publishing the read-only data as being readable by read compute nodes via the shared globally addressable memory;
• mapping shared globally addressable memory for reading by the read compute nodes;
• and
• allowing access to read-only data in the shared globally addressable memory by the read compute nodes.

A system for managing shared memory used by a plurality of compute nodes, comprising:

• a shared globally addressable memory to enable access to shared data by the plurality of compute nodes;
• a memory interface to process requests to the shared globally addressable memory from the plurality of compute nodes;
• a memory write module of the memory interface to allocate shared globally addressable memory in the shared globally addressable memory and write read-only data to the globally addressable memory from a writing compute node;
• a read module of the memory interface to map read-only data in the globally addressable shared memory as read-cacheable memory and read-only to reading compute nodes;
• and
• a re-write access module to allow the write compute node to restrict access to the read-only data in the shared globally addressable memory while the read-only data is updated.