Multithreaded programmable processor and system with partitioned operations

• US 7,987,344 B2
• Filed: 01/16/2004
• Issued: 07/26/2011
• Est. Priority Date: 08/16/1995
• Status: Expired due to Fees

• Current Assignee: Microunity Systems Engineering Incorporated
• Original Assignee: Microunity Systems Engineering Incorporated
• Inventors: Moussouris, John, Hansen, Craig
• Primary Examiner(s): Kindred; Alford W
• Assistant Examiner(s): Moll; Jesse R
• Application Number:
  US10/757,939
• Publication Number:
  US 20040210745A1
• Time in Patent Office: 2,748 Days
• Field of Search: 712/223, 712/221, 712/222, 712/9, 708/620, 708/523, 708/524
• US Class Current: 712/221
• CPC Class Codes: G06F 15/7832 on one IC chip (single chip microprocessors)
• CPC Class Codes: G06F 9/30014 with variable precision
• CPC Class Codes: G06F 9/30018 Bit or string instructions
• CPC Class Codes: G06F 9/30025 Format conversion instructions, e.g. Floating-Point to Integer, decimal conversion
• CPC Class Codes: G06F 9/30029 Logical and Boolean instructions, e.g. XOR, NOT
• CPC Class Codes: G06F 9/30032 Movement instructions, e.g. MOVE, SHIFT, ROTATE, SHUFFLE
• CPC Class Codes: G06F 9/30036 Instructions to perform operations on packed data, e.g. vector, tile or matrix operations
• CPC Class Codes: G06F 9/3004 to perform operations on memory
• CPC Class Codes: G06F 9/30043 LOAD or STORE instructions; Clear instruction
• CPC Class Codes: G06F 9/30054 Unconditional branch instructions
• CPC Class Codes: G06F 9/30087 Synchronisation or serialisation instructions
• CPC Class Codes: G06F 9/30101 Special purpose registers
• CPC Class Codes: G06F 9/30109 having multiple operands in a single register
• CPC Class Codes: G06F 9/30112 comprising data of variable length
• CPC Class Codes: G06F 9/3012 Organisation of register space, e.g. banked or distributed register file
• CPC Class Codes: G06F 9/30123 according to context, e.g. thread buffers
• CPC Class Codes: G06F 9/30145 Instruction analysis, e.g. decoding, instruction word fields
• CPC Class Codes: G06F 9/3016 Decoding the operand specifier, e.g. specifier format
• CPC Class Codes: G06F 9/30167 of immediate specifier, e.g. constants
• CPC Class Codes: G06F 9/3816 Instruction alignment, e.g. cache line crossing
• CPC Class Codes: G06F 9/3824 : Operand accessing
• CPC Class Codes: G06F 9/383 : Operand prefetching cache prefetching G06F12/0862
• CPC Class Codes: G06F 9/3851 : from multiple instruction streams, e.g. multistreaming
• CPC Class Codes: G06F 9/3861 : Recovery, e.g. branch miss-prediction, exception handling error detection or correction G06F11/00
• CPC Class Codes: G06F 9/3873 : Variable length pipelines, e.g. elastic pipeline
• CPC Class Codes: G06F 9/3885 : using a plurality of independent parallel functional units

Abstract

A programmable processor and method for improving the performance of processors by incorporating an execution unit configurable to execute a plurality of instruction streams from the plurality of threads, wherein each instruction stream includes a group instruction that operates on a plurality of data elements in partitioned fields of at least one of the registers to produce a catenated result.

Claims

1. A programmable processor comprising:

a data path capable of transmitting data;

an external interface operable to receive data from an external source and communicate the received data over the data path;

a register file containing a plurality of registers each having a register width, the register file coupled to the data path and configured to support processing of a plurality of threads and to store a plurality of multiple-bit data elements in partitioned fields, each of the multiple-bit data elements having an elemental width smaller than the register width;

an execution unit coupled to the data path, the execution unit configured to execute a plurality of instruction streams from the plurality of threads in a multistage pipeline such that the multistage pipeline is capable of including instructions from different ones of the instruction streams in different stages of the multistage pipeline, each instruction stream including a single arithmetic instruction that specifies an arithmetic operation to cause multiple instances of the arithmetic operation to be performed, each instance of the arithmetic operation to be performed using a different one of the plurality of multiple-bit data elements in partitioned fields of at least one of the registers to produce a catenated result, the single arithmetic instruction causing a plurality of multiple-bit data elements in partitioned fields to be read in parallel from a register included in the register file, and causing the catenated result to be written in parallel to one of the registers included in the register file; and

wherein each of the multiple-bit data elements has an elemental width, and the data path has a data path width multiple times greater than the elemental width, to allow multiple-bit data elements used for the multiple instances of the arithmetic operation to be transmitted in parallel from the register file to the execution unit, and wherein the execution unit is operable to receive, in parallel, multiple-bit data elements for the multiple instances of the arithmetic operation and execute the multiple instances of the single arithmetic instruction to produce the catenated result.

2. The processor of claim 1 wherein the execution unit comprises a pipeline having a plurality of stages and wherein the pipeline interleaves execution of instructions from the plurality of instruction streams.

3. The processor of claim 2 wherein the pipeline is operable to simultaneously contain states of execution of at least two instructions from different instruction streams.

4. The processor of claim 2 wherein execution of the instructions is interleaved in a round-robin manner.

5. The processor of claim 1 wherein the processor ensures only one thread from the plurality of threads can have an exception handled at any given time.

6. The processor of claim 1 further comprising a virtual memory addressing unit and a cache operable to store data communicated between the external interface and the data path.

7. The processor of claim 1 wherein the execution unit is further operable to, in response to decoding another single instruction specifying a first and a second register each containing a plurality of floating-point operands, multiply the plurality of floating-point operands in the first register by the plurality of floating-point operands in the second register to produce a plurality of products and provide the plurality of products to partitioned fields of a result register as a second catenated result.

8. The processor of claim 1 wherein the arithmetic operation comprises an integer operation.

9. The processor of claim 1 wherein the arithmetic operation comprises a floating-point operation.

10. A programmable processor comprising:

a data path capable of transmitting data;

an external interface operable to receive data from an external source and communicate the received data over the data path;

first and second register files containing a plurality of registers each having a register width, the first and second register files coupled to the data path and configured to support processing of first and second threads, respectively, and to store a plurality of multiple-bit data elements in partitioned fields, each of the multiple-bit data elements having an elemental width smaller than the register width;

an execution unit coupled to the data path, the execution unit configured to execute first and second instruction streams from the first and second threads, respectively, in a multistage pipeline such that the multistage pipeline is capable of including instructions from different ones of the instruction streams in different stages of the multistage pipeline, the first and second instruction streams each including a single arithmetic instruction that specifies an arithmetic operation to cause multiple instances of the arithmetic operation to be performed, each instance of the arithmetic operation to be performed using a different one of multiple-bit data elements in partitioned fields of at least one of the registers to produce a catenated result, the single arithmetic instruction causing a plurality of multiple-bit data elements in partitioned fields to be read in parallel from a register included in the register file, and causing the catenated result to be written in parallel to one of the registers included in the register file; and

wherein each of the multiple-bit data elements has an elemental width, and the data path has a data path width multiple times greater than the elemental width, to allow multiple-bit data elements used for the multiple instances of the arithmetic operation to be transmitted in parallel from the first register file and from the second register file to the execution unit, and wherein the execution unit is operable to receive, in parallel, multiple-bit data elements for the multiple instances of the arithmetic operation and execute the multiple instances of the single arithmetic instruction to produce the catenated result.

11. The processor of claim 10 wherein the execution unit comprises a pipeline having a plurality of stages and wherein the pipeline interleaves execution of instructions from the first instruction stream with instructions from the second instruction stream.

12. The processor of claim 11 wherein the pipeline is operable to simultaneously contain states of execution of an instruction from the first instruction stream and an instruction from the second instruction stream.

13. The processor of claim 11 wherein execution of the instructions is interleaved in a round-robin manner.

14. The processor of claim 11 wherein the execution unit is further operable to, in response to decoding another single instruction specifying a first and a second register each containing a plurality of floating-point operands, multiply the plurality of floating-point operands in the first register by the plurality of floating-point operands in the second register to produce a plurality of products and provide the plurality of products to partitioned fields of a result register as a second catenated result.

15. A data processing system comprising:

(a) a bus coupling components in the data processing system;

(b) an external memory coupled to the bus;

(c) a programmable microprocessor coupled to the bus and capable of operation independent of another host processor, the microprocessor comprising;

a data path capable of transmitting data;

an external interface operable to receive data from an external source and communicate the received data over the data path;

a register file containing a plurality of registers each having a register width, the register file coupled to the data path and configured to support processing of a plurality of threads and to store a plurality of multiple-bit data elements in partitioned fields, each of the multiple-bit data elements having an elemental width smaller than the register width;

an execution unit coupled to the data path, the execution unit configured to execute a plurality of instruction streams from the plurality of threads in a multistage pipeline such that the multistage pipeline is capable of including instructions from different ones of the instruction streams in different stages of the multistage pipeline, each instruction stream including a single arithmetic instruction that specifies an arithmetic operation to cause multiple instances of the arithmetic operation to be performed, each instance of the arithmetic operation to be performed using a different one of the plurality of data elements in partitioned fields of at least one of the registers to produce a catenated result, the single arithmetic instruction causing a plurality of multiple-bit data elements in partitioned fields to be read in parallel from a register included in the register file, and causing the catenated result to be written in parallel to one of the registers included in the register file; and

wherein each of the multiple-bit data elements has an elemental width, and the data path has a data path width multiple times greater than the elemental width, to allow multiple-bit data elements used for the multiple instances of the arithmetic operation to be transmitted in parallel from the register file to the execution unit, and wherein the execution unit is operable to receive, in parallel, multiple-bit data elements for the multiple instances of the arithmetic operation and execute the multiple instances of the single arithmetic instruction to produce the catenated result.

16. The system of claim 15 wherein the execution unit comprises a pipeline having a plurality of stages and wherein the pipeline interleaves execution of instructions from the plurality of instruction streams.

17. The system of claim 16 wherein the pipeline is operable to simultaneously contain states of execution of at least two instructions from different instruction streams.

18. The system of claim 16 wherein execution of the instructions is interleaved in a round-robin manner.

19. The system of claim 15 wherein the processor ensures only one thread from the plurality of threads can have an exception handled at any given time.

20. The system of claim 15 further comprising a virtual memory addressing unit and a cache operable to store data communicated between the external interface and the data path.

21. The system of claim 15 wherein the execution unit is further operable to, in response to decoding another single instruction specifying a first and a second register each containing a plurality of floating-point operands, multiply the plurality of floating-point operands in the first register by the plurality of floating-point operands in the second register to produce a plurality of products and provide the plurality of products to partitioned fields of a result register as a second catenated result.

22. A data processing system comprising:

(a) a bus coupling components in the data processing system;

(b) an external memory coupled to the bus;

(c) a programmable microprocessor coupled to the bus and capable of operation independent of another host processor, the microprocessor comprising;

a data path capable of transmitting data

an external interface operable to receive data from an external source and communicate the received data over the data path;

first and second register files containing a plurality of registers each having a register width, the first and second register files coupled to the data path and configured to support processing of first and second threads, respectively, and to store a plurality of multiple-bit data elements in partitioned fields, each of the multiple-bit data elements having an elemental width smaller than the register width;

an execution unit coupled to the data path, the execution unit configured to execute first and second instruction streams from the first and second threads, respectively, in a multistage pipeline such that the multistage pipeline is capable of including instructions from different ones of the instruction streams in different stages of the multistage pipeline, the first and second instruction streams each including a single arithmetic instruction that specifies an arithmetic operation to cause multiple instances of the arithmetic operation to be performed, each instance of the arithmetic operation to be performed using a different one of the plurality of multiple-bit data elements in partitioned fields of at least one of the registers to produce a catenated result, the single arithmetic instruction causing a plurality of multiple-bit data elements in partitioned fields to be read in parallel from a register included in the register file, and causing the catenated result to be written in parallel to one of the registers included in the register file; and

wherein each of the multiple-bit data elements has an elemental width, and the data path has a data path width multiple times greater than the elemental width, to allow multiple-bit data elements used for the multiple instances of the arithmetic operation to be transmitted in parallel from the first register file and from the second register file to the execution unit, and wherein the execution unit is operable to receive, in parallel, multiple-bit data elements for the multiple instances of the arithmetic operation and execute the multiple instances of the single arithmetic instruction to produce the catenated result.

23. The system of claim 22 wherein the execution unit comprises a pipeline having a plurality of stages and wherein the pipeline interleaves execution of instructions from the first instruction stream with instructions from the second instruction stream.

24. The system of claim 23 wherein the pipeline is operable to simultaneously contain states of execution of an instruction from the first instruction stream and an instruction from the second instruction stream.

25. The system of claim 23 wherein execution of the instructions is interleaved in a round-robin manner.

26. The system of claim 23 wherein the execution unit is further operable to, in response to decoding another single instruction specifying a first and a second register each containing a plurality of floating-point operands, multiply the plurality of floating-point operands in the first register by the plurality of floating-point operands in the second register to produce a plurality of products and provide the plurality of products to partitioned fields of a result register as a second catenated result.

27. A programmable processor comprising:

a data path capable of transmitting data;

an external interface operable to received data from an external source and communicate the received data over the data path;

a register file containing a plurality of registers each having a register width, the register file coupled to the data path and configured to support processing of a plurality of threads and to store a plurality of multiple-bit data elements in partitioned fields, each of the multiple-bit data elements having an elemental width smaller than the register width;

an execution unit coupled to the data path, the execution unit configured to execute a plurality of instruction streams from the plurality of threads in a multistage pipeline such that the multistage pipeline is capable of including instructions from different ones of the instruction streams in different stages of the multistage pipeline, each instruction stream including a single floating-point arithmetic instruction that specifies a floating-point arithmetic operation to cause multiple instances of the floating-point arithmetic operation to be performed, each instance of the floating point arithmetic operation to be performed using a different one of the plurality of multiple-bit data elements in partitioned fields of at least one of the registers to produce a catenated result, the single floating-point arithmetic instruction causing a plurality of multiple-bit data elements in partitioned fields to be read in parallel from a register included in the register file, and causing the catenated result to be written in parallel to one of the registers included in the register file; and

wherein each of the multiple-bit data elements has an elemental width, and the data path has a data path width multiple times greater than the elemental width, to allow multiple-bit data elements used for the multiple instances of the floating-point arithmetic operation to be transmitted in parallel from the register file to the execution unit, and wherein the execution unit is operable to receive, in parallel, multiple-bit data elements for the multiple instances of the floating-point arithmetic operation and execute the multiple instances of the single floating-point arithmetic instruction to produce the catenated result.

28. A data processing system comprising:

(a) a bus coupling components in the data processing system;

(b) an external memory coupled to the bus;

(c) a programmable microprocessor coupled to the bus and capable of operation independent of another host processor, the microprocessor comprising;

a data path capable of transmitting data;

an external interface operable to receive data from an external source and communicate the received data over the data path;

a register file containing a plurality of registers each having a register width, the register file coupled to the data path and configured to support processing of a plurality of threads and to store a plurality of multiple-bit data elements in partitioned fields, each of the multiple-bit data elements having an elemental width smaller than the register width;

an execution unit coupled to the data path, the execution unit configured to execute a plurality of instruction streams from the plurality of threads in a multistage pipeline such that the multistage pipeline is capable of including instructions from different ones of the instruction streams in different stages of the multistage pipeline, each instruction stream including a single floating-point arithmetic instruction that specifies a floating-point arithmetic operation to cause multiple instances of the floating-point arithmetic operation to be performed, each instance of the floating-point arithmetic operation to be performed using a different one of the plurality of multiple-bit data elements in partitioned fields of at least one of the registers to produce a catenated result, the single floating-point arithmetic instruction causing a plurality of multiple-bit data elements in partitioned fields to be read in parallel from a register included in the register file, and causing the catenated result to be written in parallel to one of the registers included in the register file; and

wherein each of the multiple-bit data elements has an elemental width, and the data path has a data path width multiple times greater than the elemental width, to allow multiple-bit data elements used for the multiple instances of the floating-point arithmetic operation to be transmitted in parallel from the register file to the execution unit, and wherein the execution unit is operable to receive, in parallel, multiple-bit data elements for the multiple instances of the floating-point arithmetic operation and execute the multiple instances of the single floating-point arithmetic instruction to produce the catenated result.