Instruction-set architecture for programmable Cyclic Redundancy Check (CRC) computations
View Patent ↗A method and apparatus to perform Cyclic Redundancy Check (CRC) operations on a data block using a plurality of different n-bit polynomials is provided. A flexible CRC instruction performs a CRC operation using a programmable n-bit polynomial. The n-bit polynomial is provided to the CRC instruction by storing the n-bit polynomial in one of two operands.
1. A processor comprising:
a fetch unit to fetch instructions;
a decode unit to decode the fetched instructions; and
an execution unit to execute a Cyclic Redundancy Check (CRC) instruction that identifies a first operand that stores a current n-bit residue and a programmable n-bit polynomial and a second operand that stores a 2n-bit data block, the CRC instruction to cause the execution unit to:
perform a sequence of micro-operations to expand the programmable n-bit polynomial to provide a pre-computed polynomial K, shuffle an n-bit portion from the first operand and an n-bit portion from the second operand based on an immediate data, and reduce an n-bit portion of the 2n-bit data block using the programmable n-bit polynomial, the current n-bit residue of the first operand and the pre-computed polynomial K, to provide an n-bit residue.
2. The processor of claim 1 , wherein the programmable n-bit polynomial has a maximal degree of n+1.
3. The processor of claim 1 , wherein the first operand has 2n-bits and the second operand has 2n-bits.
4. The processor of claim 3 , wherein n is 32.
5. The processor of claim 1 , wherein the first operand is both a source and a destination and the n-bit residue is to be stored in the first operand.
6. The processor of claim 1 , wherein the execution unit is to perform a shuffle word micro-instruction to shuffle bits within respective n-bit portions of the first and second operands.
7. The processor of claim 1 , wherein the sequence of micro-operations comprises:
a CRC reduction micro-instruction to reduce an n+8-bit data using the pre-computed polynomial K to provide an n+4-bit intermediate residue and to reduce the n+4 intermediate residue to the n-bit residue using the n-bit polynomial.
8. A method comprising:
performing a sequence of operations for a Cyclic Redundancy Check (CRC) instruction, the CRC instruction specifying a first operand and a second operand, the sequence of operations to perform a Cyclic Redundancy Check (CRC) operation on a 2n-bit data block stored in the second operand using a programmable n-bit polynomial stored in the first operand, comprising:
expanding the programmable polynomial to provide a pre-computed polynomial K; and
performing a sequence of micro-operations on the 2n-bit data block using the programmable n-bit polynomial and a current n-bit residue stored in the first operand and the pre-computed polynomial K to provide an n-bit residue for the 2n-bit data block, wherein performing the sequence of micro-operations comprises reducing an n+8-bit data using the pre-computed polynomial K to provide an n+4-bit intermediate residue and reducing the n+4 intermediate residue to the n-bit residue using the programmable n-bit polynomial.
9. The method of claim 8 , wherein the programmable n-bit polynomial has a maximal degree of n+1.
10. The method of claim 9 , wherein the first operand has 2n-bits and the second operand has 2n-bits.
11. The method of claim 10 , wherein n is 32.
12. The method of claim 9 , wherein the first operand is both a source and a destination and the n-bit residue is stored in the first operand.
13. The method of claim 9 , wherein performing the sequence of micro-operations comprises:
performing a shuffle by selecting an n-bit portion from the first operand and an n-bit portion from the second operand to store in a 2n-bit destination operand.
14. The method of claim 13 , further comprising performing the shuffle to shuffle bits within the respective n-bit portions.
15. A system comprising:
a processor including:
an execution unit to execute a Cyclic Redundancy Check (CRC) instruction that identifies a first operand that stores a current n-bit residue and a programmable n-bit polynomial and a second operand that stores a 2n-bit data block, the CRC instruction to cause the execution unit to:
perform a sequence of micro-operations to expand the programmable n-bit polynomial to provide a pre-computed polynomial K, shuffle an n-bit portion from the first operand and an n-bit portion from the second operand based on an immediate data, and reduce an n-bit portion of the 2n-bit data block using the programmable n-bit polynomial, the current n-bit residue of the first operand and the pre-computed polynomial K, to provide an n-bit residue; and
a dynamic random access memory coupled to the processor.
16. The system of claim 15 , wherein the processor comprises a fetch unit to fetch instructions.
17. The system of claim 16 , wherein the processor further comprises a decode unit to decode fetched instructions.
18. The system of claim 17 , wherein the processor further comprises a register file including a plurality of registers.
19. The system of claim 17 , wherein the processor further comprises a retirement unit to retire executed instructions.
20. The system of claim 19 , wherein the processor further comprises a cache coupled to the execution unit and the retirement unit.