The Cosmic Cube is a parallel computer with 64 nodes connected in a binary 6-cube network, designed to simulate future VLSI systems. It supports high concurrency and is scalable, with potential for future machines with thousands of nodes. The system uses message-passing for communication between nodes, which differs from shared-storage multiprocessors. The Cosmic Cube's architecture separates processor-storage communication and interprocess communication, allowing efficient handling of concurrent processes. The system's operating system kernel manages processes, messages, and system calls, and is implemented in software to simulate future single-chip nodes. The Cosmic Cube's message-passing model is simpler and more economical than shared-storage systems, especially with many processors. It is effective for scientific and engineering computations, with the 64-node version being significantly faster than a VAX11/780. The system's software includes a kernel that handles message routing and process management, and supports concurrent programming through processes and virtual communication channels. The Cosmic Cube's hardware is based on Intel 8086 processors, with 128K bytes of storage and a 700-watt power consumption. The system has been used for various applications, including simulations of physical systems and circuit analysis. The Cosmic Cube's design has influenced future systems, with potential for larger, more efficient machines. The system's software and hardware have been developed through extensive research and experimentation, with ongoing work supported by DARPA. The Cosmic Cube's architecture and software have shown promise for high-concurrency computing, with applications in science and engineering. The system's performance and efficiency have been demonstrated through various benchmarks and applications, showing its potential for future computing.The Cosmic Cube is a parallel computer with 64 nodes connected in a binary 6-cube network, designed to simulate future VLSI systems. It supports high concurrency and is scalable, with potential for future machines with thousands of nodes. The system uses message-passing for communication between nodes, which differs from shared-storage multiprocessors. The Cosmic Cube's architecture separates processor-storage communication and interprocess communication, allowing efficient handling of concurrent processes. The system's operating system kernel manages processes, messages, and system calls, and is implemented in software to simulate future single-chip nodes. The Cosmic Cube's message-passing model is simpler and more economical than shared-storage systems, especially with many processors. It is effective for scientific and engineering computations, with the 64-node version being significantly faster than a VAX11/780. The system's software includes a kernel that handles message routing and process management, and supports concurrent programming through processes and virtual communication channels. The Cosmic Cube's hardware is based on Intel 8086 processors, with 128K bytes of storage and a 700-watt power consumption. The system has been used for various applications, including simulations of physical systems and circuit analysis. The Cosmic Cube's design has influenced future systems, with potential for larger, more efficient machines. The system's software and hardware have been developed through extensive research and experimentation, with ongoing work supported by DARPA. The Cosmic Cube's architecture and software have shown promise for high-concurrency computing, with applications in science and engineering. The system's performance and efficiency have been demonstrated through various benchmarks and applications, showing its potential for future computing.