Packed data rearrangement control indexes generation processors, methods, systems and instructions

Patent metadata

Field	Value
Publication number	US-9904547-B2
Application number	US-201113977217-A
Country	US
Kind code	B2
Filing date	Dec 22, 2011
Priority date	Dec 22, 2011
Publication date	Feb 27, 2018
Grant date	Feb 27, 2018

Abstract

A method of an aspect includes receiving a packed data rearrangement control indexes generation instruction. The packed data rearrangement control indexes generation instruction indicates a destination storage location. A result is stored in the destination storage location in response to the packed data rearrangement control indexes generation instruction. The result includes a sequence of at least four non-negative integers representing packed data rearrangement control indexes. In an aspect, values of the at least four non-negative integers are not calculated using a result of a preceding instruction. Other methods, apparatus, systems, and instructions are disclosed.

First claim

What is claimed is:

1. A method comprising:
receiving a packed data rearrangement control indexes generation instruction, the packed data rearrangement control indexes generation instruction indicating one or more numerical pattern defining parameters, and indicating a destination storage location;
using each of the one or more numerical pattern defining parameters to evaluate a numerical pattern defining relation that is implicit to the packed data rearrangement control indexes generation instruction; and
storing a result in the destination storage location in response to the packed data rearrangement control indexes generation instruction, the result including a sequence of at least four non-negative integers representing packed data rearrangement control indexes, each of the one or more numerical pattern defining parameters affecting a value of each of the at least four non-negative integers, at least one of the one or more numerical pattern defining parameters being used to generate a value of at least a plurality of the at least four non-negative integers,
wherein values of the at least four non-negative integers are not calculated using a result of a preceding instruction, and
wherein it is implicit to an opcode of the instruction that the result includes the sequence of integers.

2. The method of claim 1 , wherein receiving comprises receiving the packed data rearrangement control indexes generation instruction that does not indicate a source packed data operand having a plurality of packed data elements in an architecturally-visible storage location.

3. The method of claim 1 , wherein storing the result comprises storing the sequence of at least eight non-negative integers having a numerical pattern, and wherein the numerical pattern is based predominantly on the opcode of the packed data rearrangement control indexes generation instruction and the one or more numerical pattern defining parameters.

4. The method of claim 1 , wherein receiving comprises receiving the packed data rearrangement control indexes generation instruction indicating an integer offset, and wherein storing comprises storing the sequence of the at least four non-negative integers with a smallest of the at least four non-negative integers differing from zero by the integer offset.

5. The method of claim 1 , wherein receiving comprises receiving the packed data rearrangement control indexes generation instruction indicating a constant integer stride, and wherein storing comprises storing the sequence of the at least four non-negative integers with all consecutive integers differing by the constant integer stride.

6. The method of claim 1 , wherein storing comprises storing the sequence of the at least four non-negative integers with all consecutive integers differing by a constant stride, and wherein receiving comprises receiving the instruction having the opcode that fixes that all the consecutive integers differ by the constant stride.

7. The method of claim 1 , wherein receiving comprises receiving the instruction indicating a plurality of numerical pattern defining parameters.

8. The method of claim 1 , wherein storing comprises storing a result including the sequence of at least thirty-two non-negative integers representing packed data rearrangement control indexes.

9. The method of claim 1 , wherein the one or more numerical pattern defining parameters comprise a stride.

10. The method of claim 1 , further comprising:
receiving a packed data rearrangement instruction selected from a permute instruction and a shuffle instruction, the packed data rearrangement instruction indicating the packed data rearrangement control indexes, indicating at least one source packed data, and indicating a second destination storage location; and
storing a packed data result in the second destination storage location in response to the packed data rearrangement instruction, the packed data result including data elements from the at least one source packed data rearranged according to the packed data rearrangement control indexes.

11. An apparatus comprising:
a destination storage location;
a decode unit to decode a packed data rearrangement control indexes generation instruction that is to indicate one or more numerical pattern defining parameters, and is to indicate the destination storage location; and
an execution unit coupled with the destination storage location, the execution unit, in response to the packed data rearrangement control indexes generation instruction being decoded, to store a result in the destination storage location, the result to include a sequence of at least four non-negative integers that are to represent packed data rearrangement control indexes, wherein at least a plurality of the at least four non-negative integers of the sequence are to have been generated from one of the one or more numerical pattern defining parameters, and wherein the result is to be stored without calculating values of the least four non-negative integers from a result of a preceding instruction,
wherein it is implicit to an opcode of the instruction that the result is to include the sequence of integers, and
wherein said one of the one or more numerical pattern defining parameters is not to be provided by an immediate.

12. The apparatus of claim 11 , wherein the execution unit is to store the result responsive to the packed data rearrangement control indexes generation instruction that is not to indicate a source packed data operand having a plurality of packed data elements in an architecturally-visible storage location.

13. The apparatus of claim 11 , wherein the execution unit, in response to the packed data rearrangement control indexes generation instruction, is to store the sequence of at least eight non-negative integers that are to have a numerical pattern, and wherein the numerical pattern is to be based predominantly on the opcode of the packed data rearrangement control indexes generation instruction and the one or more numerical pattern defining parameters.

14. The apparatus of claim 11 , wherein each of the one or more numerical pattern defining parameters is to affect a value of each of the at least four non-negative integers.

15. The apparatus of claim 11 , wherein the execution unit, in response to the packed data rearrangement control indexes generation instruction, is to store the sequence of at least eight non-negative integers that are to have a numerical pattern that is based entirely on the opcode of the packed data rearrangement control indexes generation instruction and the one or more numerical pattern defining parameters.

16. The apparatus of claim 11 , wherein the packed data rearrangement control indexes generation instruction is to indicate an integer offset, and wherein the execution unit, in response to the packed data rearrangement control indexes generation instruction, is to store the sequence of the at least four non-negative integers with a smallest of the at least four non-negative integers to differ from zero by the integer offset.

17. The apparatus of claim 11 , wherein the execution unit, in response to the opcode of the packed data rearrangement control indexes generation instruction, is to store all consecutive integers in the sequence of the at least four non-negative integers to differ by a constant stride.

18. The apparatus of claim 11 , wherein the one or more numerical pattern defining parameters comprise a plurality of numerical pattern defining parameters.

19. The apparatus of claim 11 , wherein the one or more numerical patt

Assignees

• Ould Ahmed Vall Elmoustapha
• Abraham Seth
• Valentine Robert
• Sperber Zeev
• Gradstein Amit
• Intel Corp

Inventors

• Ould Ahmed Vall Elmoustapha
• Abraham Seth
• Valentine Robert
• Sperber Zeev
• Gradstein Amit

Classifications

• G06F9/30032

  Movement instructions, e.g. MOVE, SHIFT, ROTATE, SHUFFLE · CPC title

• G06F9/30036

  Instructions to perform operations on packed data, e.g. vector, tile or matrix operations · CPC title

• G06F9/30043 Primary

  LOAD or STORE instructions; Clear instruction · CPC title

• G06F9/30163

  with implied specifier, e.g. top of stack · CPC title

• G06F9/30167

  of immediate specifier, e.g. constants · CPC title