摘要:
In a data processing system, a system memory includes first memory modules having a data path of a first bit width and second memory modules having a data path of a second bit width with the first bit width being less than the second bit width. A central subsystem includes a cache memory unit and processing units for initiating requests for data transfers of the second bit width between the system memory and the subsystem processing units. An interface coupling the system memory and the central subsystem for bidirectional data transfers generates, in response to a memory request of a second bit width wherein the requested data is stored in a first memory module, additional memory requests until sufficient data has been retrieved from the system memory to satisfy the central subsystem request. The interface also monitors data transfers between the system processing units and the system memory and transfers the data transfers to the central subsystem in order to update and to retain the integrity of the cache memory in the central subsystem.
摘要:
A cache memory for use in a data processing system wherein data words identified by even address numbers are stored separately from data words associated with odd address numbers to enable the simultaneous transfer of two successively addressed data words to or from the cache memory by the transferring of a data word associated with an odd address number and a data word associated with an even address number.
摘要:
A data processing system includes a memory subsystem for storing operands and instructions and a central processing unit (CPU) for manipulating the operands by executing the instructions. The CPU includes a control store for generating signals for controlling the CPU operation. Shifters made up of multiplexers shift operands between an outer bus and a write bus in response to control store signals. The multiplexers shift the operands left or right 1, 2 or 4-bit positions including open shifts and circular shifts and also perform byte position shifting and twinning.
摘要:
A technique and apparatus for storing data and addressing stored data in a memory from which multiple words of data are to be retrieved in parallel is disclosed. The memory is addressed by providing the address of the first of N consecutive words to be retrieved in parallel. The data is stored in memory in physical data words which contain N logical data words such that the addressing of one physical data word will result in N logical data words being read in parallel from the memory. Each physical data word contains the contents of the logical data word having the same address as that of the physical data word in its leftmost position followed in the next right position by the contents of the logical data word having the next higher address, and so on until the rightmost position of the physical data word contains the contents of the logical data word with an address equal to the physical data word address plus N-1. This results in the contents of each logical data word being stored N times in the memory, but eliminates the need for data alignment as the N logical data words are read in parallel from the memory.
摘要:
A data processing system for providing word, byte or bit addressing. A word location in a memory device may be addressed based upon the contents of a base address register. Indirect addressing may be provided to another word location based upon a word index value in an index register. Effective byte or bit addressing of the addressed word is provided in response to byte and bit index values which are produced by means of the index register. An instruction word indicates the type of addressing and directs the use of different control words included in a control storage device in order to implement the desired operation.
摘要:
A logic control system is disclosed for accommodating the flow of both procedural information and CPU (central processing unit) instructions from a central memory system to a CPU without compromising memory bandwidths or CPU execution speeds because of transfer delays or timing variances. Instruction modifications and plural task assignments are accommodated during instruction execution.
摘要:
In a data processing system, a central processor unit requests procedural data words or non-procedural data words stored in the system memory. A control store device executes firmware instructions including a local bus field for controlling the transfer of the requested procedural data words and non-procedural data words to the central processor unit. The requested procedural data words and non-procedural data words are transferred to the central processing unit by an interfacing device including a data bus latch for receiving the procedural data words and non-procedural data words transferred from the memory, a prefetch buffer for storing up to four words, a first set of OR gate circuits for selectively transferring the procedural data words stored in the prefetch buffer to a procedural data multiplexer for assembling either a procedural data word or a procedure address, and a second set of OR gate circuits for selectively transferring either a procedural data word or non-procedural data word to the source bus or a procedural data address or non-procedural data address to the source bus for transfer to the central processor unit.
摘要:
A clock system for providing rectangular wave forms or wave trains, with each wave train having a selectable predetermined clock cycle period. A rectangular wave train is generated by a generator comprising a delay line coupled to an INVERTER. By using a second delay line to delay the rectangular wave by a selectable predetermined delay period, a control signal is formed which when fed into the generator produces a second rectangular wave train with a clock cycle period equal to that of the rectangular wave clock cycle period plus the period of the second predetermined delay. The addition of a synchronization circuit permits the clock cycle period to be dynamically selected during a clock cycle. This provides a rectangular train with the period of each clock cycle being any of the predetermined clock cycle periods independent of the clock cycle period of preceding or succeeding clock cycles.
摘要:
A final effective address of an operand is generated in a microprogrammed data processing system by use of a base address register which may include an unindexed address, an index register which may include an index address value, an instruction register which may include an instruction word, which instruction word provides control over the addressing of a control store dependent upon the state of a selected one of a plurality of test conditions. The addressed control store word provides signals for controlling the operation of the system, including the branching between such major operations as instruction fetching, addressing, reading, writing, and execution as well as branching between minor operations which are included in the major operations.
摘要:
A data processing system includes a microprogram controlled central processing unit that executes instructions. The instruction words include a data type field for identifying the type of operand processed during the execution of the instruction. The data type field signals and a number of control signals are applied to the address terminals of a read only memory. The read only memory output signals are tested by microwords of a microprogram to branch to firmware routines to process the operand type.