Abstract: An indirect VLIW (iVLIW) architecture is described which contains a minimum of two instruction memories. The first instruction memory (SIM) contains short-instruction-words (SIWs) of a fixed length. The second instruction memory (VIM), contains very-long-instruction-words (VLIWs) which allow execution of multiple instructions in parallel. Each SIW may be fetched and executed as an independent instruction by one of the available execution units. A special class of SIW is used to reference the VIM indirectly to either execute or load a specified VLIW instruction (called an “XV” instruction for “eXecute VLIW”, or LV for “Load VLIW”). In these cases, the SIW instruction specifies how the location of the VLIW is to be accessed. Other aspects of this approach relate to the application of data memory addressing techniques for execution or loading of VLIWs that parallel the addressing modes used for data memory accesses.

Abstract: An improved manifold array (ManArray) architecture addresses the problem of configurable application-specific instruction set optimization and instruction memory reduction using an instruction abbreviation process thereby further optimizing the general ManArray architecture for application to high-volume and portable battery-powered type of products. In the ManArray abbreviation process a standard 32-bit ManArray instruction is reduced to a smaller length instruction format, such as 14-bits. An application is first programmed using the full ManArray instruction set using the native 32-bit instructions. After the application program is completed and verified, an instruction-abbreviation tool analyzes the 32-bit application program and generates the abbreviated program using the abbreviated instructions. This instruction abbreviation process allows different program-reduction optimizations tailored for each application program. This process develops an optimized instruction set for the intended application.

Abstract: An array processor includes processing elements arranged in clusters which are, in turn, combined in a rectangular array. Each cluster is formed of processing elements which preferably communicate with the processing elements of at least two other clusters. Additionally each inter-cluster communication path is mutually exclusive, that is, each path carries either north and west, south and east, north and east, or south and west communications. Due to the mutual exclusivity of the data paths, communications between the processing elements of each cluster may be combined in a single inter-cluster path. That is, communications from a cluster which communicates to the north and east with another cluster may be combined in one path, thus eliminating half the wiring required for the path. Additionally, the length of the longest communication path is not directly determined by the overall dimension of the array, as it is in conventional torus arrays.

Abstract: A very long instruction word (VLIW) processor typically requires a large number of register file ports due to the parallel execution of the sub-instructions comprising the VLIW. By splitting a general purpose register file into separate address and compute register files, the number of compute register file ports is significantly reduced. This reduction is particularly evident when multiple load and store execution units with indexed addressing modes are supported. The implication is that a faster register file and dedicated address registers are achieved in the programming model. The savings comes at the cost of providing support for data movement between the compute register file and the address register file. In addition, address arithmetic, table look-up, and store to table functions are desirable functions that cannot be obviously obtained when the address registers are separated from the compute registers.

Abstract: Processing element to processing element switch connection control is described using a receive model that precludes communication hazards from occurring in a synchronous MIMD mode of operation. Such control allows different communication topologies and various processing effects such as an array transpose, hypercomplement or the like to be efficiently achieved utilizing architectures, such as the manifold array processing architecture. An encoded instruction method reduces the amount of state information and setup burden on the programmer taking advantage of the recognition that the majority of algorithms will use only a small fraction of all possible mux settings available. Thus, by means of transforming the PE identification based upon a communication path specified by a PE communication instruction an efficient switch control mechanism can be used.

Abstract: General purpose flags (ACFs) are defined and encoded utilizing a hierarchical one-, two- or three-bit encoding. Each added bit provides a superset of the previous functionality. With condition combination, a sequential series of conditional branches based on complex conditions may be avoided and complex conditions can then be used for conditional execution. ACF generation and use can be specified by the programmer. By varying the number of flags affected, conditional operation parallelism can be widely varied, for example, from mono-processing to octal-processing in VLIW execution, and across an array of processing elements (PE)s. Multiple PEs can generate condition information at the same time with the programmer being able to specify a conditional execution in one processor based upon a condition generated in a different processor using the communications interface between the processing elements to transfer the conditions.

Abstract: A reconfigurable register file integrated in an instruction set architecture capable of extended precision operations, and also capable of parallel operation on lower precision data is described. A register file is composed of two separate files with each half containing half as many registers as the original. The halves are designated even or odd by virtue of the register addresses which they contain. Single width and double width operands are optimally supported without increasing the register file size and without increasing the number of register file ports. Separate extended registers are also employed to provide extended precision for operations such as multiply-accumulate operations.

Abstract: An array processor includes processing elements arranged in clusters which are, in turn, combined in a rectangular array. Each cluster is formed of processing elements which preferably communicate with the processing elements of at least two other clusters. Additionally each inter-cluster communication path is mutually exclusive, that is, each path carries either north and west, south and east, north and east, or south and west communications. Due to the mutual exclusivity of the data paths, communications between the processing elements of each cluster may be combined in a single inter-cluster path. That is, communications from a cluster which communicates to the north and east with another cluster may be combined in one path, thus eliminating half the wiring required for the path. Additionally, the length of the longest communication path is not directly determined by the overall dimension of the array, as it is in conventional torus arrays.

Abstract: A hierarchical instruction set architecture (ISA) provides pluggable instruction set capability and support of array processors. The term pluggable is from the programmer's viewpoint and relates to groups of instructions that can easily be added to a processor architecture for code density and performance enhancements. One specific aspect addressed herein is the unique compacted instruction set which allows the programmer the ability to dynamically create a set of compacted instructions on a task by task basis for the primary purpose of improving control and parallel code density. These compacted instructions are parallelizable in that they are not specifically restricted to control code application but can be executed in the processing elements (PEs) in an array processor. The ManArray family of processors is designed for this dynamic compacted instruction set capability and also supports a scalable array of from one to N PEs.

Abstract: A variety of advantageous mechanisms for improved data transfer control within a data processing system are described. A DMA controller is described which is implemented as a multiprocessing transfer engine supporting multiple transfer controllers which may work independently or in cooperation to carry out data transfers, with each transfer controller acting as an autonomous processor, fetching and dispatching DMA instructions to multiple execution units. In particular, mechanisms for initiating and controlling the sequence of data transfers are provided, as are processes for autonomously fetching DMA instructions which are decoded sequentially but executed in parallel.

Abstract: Techniques are described for providing mechanisms of data distribution to and collection of data from multiple memories in a data processing system. The system may suitably be a manifold array (ManArray) processing system employing an array of processing elements. Virtual to physical processing element (PE) identifier translation is employed in conjunction with a ManArray PE interconnection topology to support a variety of communication models, such as hypercube and such. Also, PE addressing nodes are based upon logically nested parameterized loops. Mechanisms for updating loop parameters, as well as exemplary instruction formats are also described.