In-lane vector shuffle instructions
In-lane vector shuffle operations are described. In one embodiment a shuffle instruction specifies a field of per-lane control bits, a source operand and a destination operand, these operands having corresponding lanes, each lane divided into corresponding portions of multiple data elements. Sets of data elements are selected from corresponding portions of every lane of the source operand according to per-lane control bits. Elements of these sets are copied to specified fields in corresponding portions of every lane of the destination operand. Another embodiment of the shuffle instruction also specifies a second source operand, all operands having corresponding lanes divided into multiple data elements. A set selected according to per-lane control bits contains data elements from every lane portion of a first source operand and data elements from every corresponding lane portion of the second source operand. Set elements are copied to specified fields in every lane of the destination operand.
1. A processor comprising:
a plurality of registers; and
an execution unit, including circuitry, coupled with the plurality of registers, the execution unit in response to a decoded single instruction to generate and store a result of the single instruction in a destination operand, the single instruction to specify a source operand, the destination operand, and an immediate operand, wherein the source operand and the destination operand are each to have a first lane and a second lane, wherein the first lane of the source operand is to store a first plurality of data elements, the second lane of the source operand is to store a second plurality of data elements, and wherein the immediate operand is to specify a first plurality of control bits, a second plurality of control bits, a third plurality of control bits, and a fourth plurality of control bits, the execution unit to use the first, second, third, and fourth pluralities of control bits for each of the first and second lanes, wherein the execution unit is to:
copy one of the first plurality of data elements specified by the first plurality of control bits to a first data element position of the first lane of the destination operand, copy one of the first plurality of data elements specified by the second plurality of control bits to a second data element position of the first lane of the destination operand, copy one of the first plurality of data elements specified by the third plurality of control bits to a third data element position of the first lane of the destination operand, and copy one of the first plurality of data elements specified by the fourth plurality of control bits to a fourth data element position of the first lane of the destination operand; and
copy one of the second plurality of data elements specified by the first plurality of control bits to a first data element position of the second lane of the destination operand, copy one of the second plurality of data elements specified by the second plurality of control bits to a second data element position of the second lane of the destination operand, copy one of the second plurality of data elements specified by the third plurality of control bits to a third data element position of the second lane of the destination operand, and copy one of the second plurality of data elements specified by the fourth plurality of control bits to a fourth data element position of the second lane of the destination operand.
2. The processor of claim 1 , wherein the source operand is a 256-bit operand, and wherein each of the first lane of the source operand and the second lane of the source operand is a 128-bit lane.
3. The processor of claim 2 , wherein each of the first plurality of data elements is a 32-bit data element and each of the second plurality of data elements is a 32-bit data element.
4. The processor of claim 1 , wherein the immediate operand is an 8-bit operand.
5. The processor of claim 4 , wherein the first plurality of control bits, the second plurality of control bits, the third plurality of control bits, and the fourth plurality of control bits each consist of 2 bits.
6. The processor of claim 1 , wherein the destination operand is a 256-bit operand, and wherein each of the first lane of the destination operand and the second lane of the destination operand is a 128-bit lane.
7. The processor of claim 1 , wherein the first lane of the source operand occupies one half of the source operand and the second lane of the source operand occupies another half of the source operand.
8. A system comprising:
a processor;
a memory; and
a bus to communicatively couple the processor to a plurality of other system components, wherein the processor includes:
an execution unit including circuitry, the execution unit in response to a decoded single instruction to generate and store a result of the single instruction in a destination operand, the single instruction to specify a source operand, the destination operand, and an immediate operand, wherein the source operand and the destination operand are each to have a first lane and a second lane, the first lane of the source operand to store a first plurality of data elements, the second lane of the source operand to store a second plurality of data elements, and the immediate operand to specify a first plurality of control bits, a second plurality of control bits, a third plurality of control bits, and a fourth plurality of control bits, the execution unit to use the first, second, third, and fourth pluralities of control bits for each of the first and second lanes, the execution unit to:
copy one of the first plurality of data elements specified by the first plurality of control bits to a first data element position of the first lane of the destination operand, copy one of the first plurality of data elements specified by the second plurality of control bits to a second data element position of the first lane of the destination operand, copy one of the first plurality of data elements specified by the third plurality of control bits to a third data element position of the first lane of the destination operand, and copy one of the first plurality of data elements specified by the fourth plurality of control bits to a fourth data element position of the first lane of the destination operand; and
copy one of the second plurality of data elements specified by the first plurality of control bits to a first data element position of the second lane of the destination operand, copy one of the second plurality of data elements specified by the second plurality of control bits to a second data element position of the second lane of the destination operand, copy one of the second plurality of data elements specified by the third plurality of control bits to a third data element position of the second lane of the destination operand, and copy one of the second plurality of data elements specified by the fourth plurality of control bits to a fourth data element position of the second lane of the destination operand.
9. The system of claim 8 , wherein the source operand is a 256-bit operand, and wherein each of the first lane of the source operand and the second lane of the source operand is a 128-bit lane.
10. The system of claim 9 , wherein each of the first plurality of data elements is a 32-bit data element and each of the second plurality of data elements is a 32-bit data element.
11. The system of claim 8 , wherein the immediate operand is an 8-bit operand.
12. The system of claim 11 , wherein the first plurality of control bits, the second plurality of control bits, the third plurality of control bits, and the fourth plurality of control bits each consist of 2 bits.
13. The system of claim 8 , wherein the destination operand is a 256-bit operand, and wherein each of the first lane of the destination operand and the second lane of the destination operand is a 128-bit lane.
14. The system of claim 8 , wherein the first lane of the source operand occupies one half of the source operand and the second lane of the source operand occupies another half of the source operand.
15. A processor comprising:
a plurality of registers; and
an execution unit coupled with the plurality of registers, the execution unit including circuitry, the execution unit in response to a decoded single instruction to generate and store a result in a destination operand, the single instruction to specify a first source operand, a second source operand, the destination operand, and an immediate operand, the first source operand, the second source operand, and the destination operand each to have a first lane and a second lane, the first lane of the first source operand to store data elements X 4 , X 3 , X 2 , and X 1 (X 4 -X 1 ), the second lane of the first source operand to store data elements X 8 , X 7 , X 6 , and X 5 (X 8 -X 5 ), the first lane of the second source operand to store data elements Y 4 , Y 3 , Y 2 , and Y 1 (Y 4 -Y 1 ), the second lane of the second source operand to store data elements Y 8 , Y 7 , Y 6 , and Y 5 (Y 8 -Y 5 ), and the immediate operand to specify a first plurality of control bits, a second plurality of control bits, a third plurality of control bits, and a fourth plurality of control bits, the execution unit to use the first, second, third, and fourth pluralities of control bits for each of the first and second lanes of the destination operand, the execution unit to:
copy one of the data elements X 4 -X 1 specified by the first plurality of control bits to a first data element position of the first lane of the destination operand, copy one of the data elements X 4 -X 1 specified by the second plurality of control bits to a second data element position of the first lane of the destination operand, copy one of the data elements Y 4 -Y 1 specified by the third plurality of control bits to a third data element position of the first lane of the destination operand, and copy one of the data elements Y 4 -Y 1 specified by the fourth plurality of control bits to a fourth data element position of the first lane of the destination operand; and
copy one of the data elements X 8 -X 5 specified by the first plurality of control bits to a first data element position of the second lane of the destination operand, copy one of the data elements X 8 -X 5 specified by the second plurality of control bits to a second data element position of the second lane of the destination operand, copy one of the data elements Y 8 -Y 5 specified by the third plurality of control bits to a third data element position of the second lane of the destination operand, and copy one of the data elements Y 8 -Y 5 specified by the fourth plurality of control bits to a fourth data element position of the second lane of the destination operand.
16. The processor of claim 15 , wherein the first source operand is a 256-bit operand and each of the first lane of the first source operand and the second lane of the first source operand is a 128-bit lane, wherein each of the data elements X 4 -X 1 is a 32-bit data element and each of the data elements X 8 -X 5 is a 32-bit data element.
17. A processor comprising:
a plurality of registers; and
an execution unit coupled with the plurality of registers, the execution unit including circuitry, the execution unit in response to a decoded single instruction to generate and store a result in a destination operand, the single instruction to specify a source operand, the destination operand, and an immediate operand, the source operand and the destination operand each to have a first lane and a second lane, an upper portion of the first lane of the source operand to store data elements D, C, B, and A (D-A), an upper portion of the second lane of the source operand to store data elements H, G, F, and E (H-E), and the immediate operand to specify a first plurality of control bits, a second plurality of control bits, a third plurality of control bits, and a fourth plurality of control bits, the execution unit to use the first, second, third, and fourth pluralities of control bits for each of the upper portion of the first lane of the source operand and the upper portion of the second lane of the source operand, the execution unit to:
copy one of the data elements D-A specified by the first plurality of control bits to a first data element position of an upper portion of the first lane of the destination operand, copy one of the data elements D-A specified by the second plurality of control bits to a second data element position of the upper portion of the first lane of the destination operand, copy one of the data elements D-A specified by the third plurality of control bits to a third data element position of the upper portion of the first lane of the destination operand, and copy one of the data elements D-A specified by the fourth plurality of control bits to a fourth data element position of the upper portion of the first lane of the destination operand; and
copy one of the data elements H-E specified by the first plurality of control bits to a first data element position of an upper portion of the second lane of the destination operand, copy one of the data elements H-E specified by the second plurality of control bits to a second data element position of the upper portion of the second lane of the destination operand, copy one of the data elements H-E specified by the third plurality of control bits to a third data element position of the upper portion of the second lane of the destination operand, and copy one of the data elements H-E specified by the fourth plurality of control bits to a fourth data element position of the upper portion of the second lane of the destination operand.
18. The processor of claim 17 , wherein the source operand is a 256-bit operand and each of the first lane of the source operand and the second lane of the source operand is a 128-bit lane, wherein the upper portion of the first lane of the source operand occupies bits 127 - 64 of the source operand and the upper portion of the second lane of the source operand occupies bits 255 - 192 of the source operand.