System and method of floating point multiply operation processing

1. A processor comprising:

an integer multiplier configured to execute an integer multiply instruction to multiply significand bits of at least one floating point operand of a floating point multiply operation to generate an intermediate result; and

a floating point multiplier configured to execute a special purpose floating point multiply accumulate instruction with respect to the intermediate result and the at least one floating point operand to generate a final floating point multiplication result.

2. The processor of claim 1 , wherein the integer multiply instruction includes at least one of:

a first instruction to perform integer multiplication of low-order significand bits of the at least one floating point operand with low-order significand bits of a second operand to generate a first value;

a second instruction to perform integer multiplication of the low-order significand bits of the at least one floating point operand and high-order significand bits of the second operand to generate a second value; or

a third instruction to perform integer multiplication of the low-order significand bits of the second operand and high-order significand bits of the at least one floating point operand to generate a third value.

3. The processor of claim 2 , further comprising a register configured to store the intermediate result, and wherein the intermediate result corresponds to a sum of the third value, the second value, and a reduced-length version of the first value.

4. The processor of claim 3 , wherein the intermediate value includes a dedicated bit that indicates whether the reduced-length version of the first value is inexact.

5. The processor of claim 1 , wherein the floating point multiplier is configured to execute the special purpose floating point multiply accumulate instruction by:

performing floating-point multiplication that includes multiplying high-order significand bits of the at least one floating point operand with high-order significand bits of a second operand to generate high-order significand bits of a floating point value; and

accumulating the high-order significand bits with the intermediate result to generate the final floating point multiplication result.

6. The processor of claim 1 , further comprising a decoder configured to replace a received floating point multiply instruction with a set of instructions that includes a plurality of integer multiply instructions and the special purpose floating point multiply accumulate instruction.

7. The processor of claim 6 , wherein the set of instructions further includes an instruction that is executable to:

determine whether the at least one floating point operand of the floating point multiply operation is denormal; and

in response to determining that the at least one floating point operand is denormal and that an exponent of a second operand of the floating point multiply operation is larger than a threshold, modify a value of at least one of the at least one floating point operand or the second operand.

8. A method of performing a floating point multiply operation at a processor, the method comprising:

executing, by an integer multiplier, a plurality of integer multiply instructions to multiply significand bits of at least one floating point operand of the floating point multiply operation to generate an intermediate result; and

executing, by a floating point multiplier, a special purpose floating point multiply accumulate instruction with respect to the intermediate result and the at least one floating point operand of the floating point multiply operation to generate a floating point multiplication result.

9. The method of claim 8 , wherein the plurality of integer multiply instructions includes:

a first instruction to perform integer multiplication of low-order significand bits of the at least one floating point operand with low-order significand bits of a second operand to generate a first value;

a second instruction to perform integer multiplication of the low-order significand bits of the at least one floating point operand and high-order significand bits of the second operand to generate a second value; and

a third instruction to perform integer multiplication of the low-order significand bits of the second operand and high-order significand bits of the at least one floating point operand to generate a third value.

10. The method of claim 9 , wherein the intermediate result corresponds to a sum of the third value, the second value, and a reduced-length version of the first value.

11. The method of claim 10 , wherein the intermediate value includes a dedicated bit that indicates whether the reduced-length version of the first value is inexact.

12. The method of claim 8 , wherein executing the special purpose floating point multiply accumulate instruction causes the processor to:

perform floating-point multiplication that includes multiplying high-order significand bits of the at least one floating point operand with high-order significand bits of a second operand to generate high-order significand bits of a floating point value; and

accumulate the high-order significand bits with the intermediate result to generate the floating point multiplication result.

13. The method of claim 8 , further comprising, prior to executing the plurality of integer multiply instructions:

receiving a floating point multiply instruction; and

replacing the floating point multiply instruction with the plurality of integer multiply instructions and the special purpose floating point multiply accumulate instruction.

14. The method of claim 13 , further comprising, in response to receiving the floating point multiply instruction:

determining whether the at least one floating point operand of the floating point multiply operation is denormal; and

in response to determining that the at least one floating point operand is denormal and that an exponent of a second operand of the floating point multiply operation is larger than a threshold, modifying a value of at least one of the at least one floating point operand or the second operand.

15. An apparatus comprising:

means for executing a plurality of integer multiply instructions to multiply significand bits of at least one floating point operand of a floating point multiply operation to generate an intermediate result; and

means for executing a special purpose floating point multiply accumulate instruction with respect to the intermediate result and the at least one floating point operand to generate a final floating point multiplication result.

16. The apparatus of claim 15 , further comprising means for replacing a floating point multiply instruction with the plurality of integer multiply instructions and the special purpose floating point multiply accumulate instruction.

17. The apparatus of claim 15 , wherein the plurality of integer multiply instructions includes:

a first instruction to perform integer multiplication of low-order significand bits of the at least one floating point operand with low-order significand bits of a second operand to generate a first value;

a second instruction to perform integer multiplication of the low-order significand bits of the at least one floating point operand and high-order significand bits of the second operand to generate a second value; and

a third instruction to perform integer multiplication of the low-order significand bits of the second operand and high-order significand bits of the at least one floating point operand to generate a third value.

18. The apparatus of claim 17 , further comprising means for storing the intermediate result, and wherein the intermediate result corresponds to a sum of the third value, the second value, and a reduced-length version of the first value.

19. A method comprising:

identifying a first operand of a multi-operand floating-point multiply operation as denormal;

in response to determining that an exponent of a second operand of the multi-operand floating-point operation is larger than a threshold, modifying a value of at least one of the first operand or the second operand;

executing, at an integer multiplier, at least one integer multiply instruction using the modified value to generate an intermediate result of the multi-operand floating-point multiply operation, and

generating, at a floating point multiplier, a final result of the multi-operand floating-point multiply operation at least partially based on the intermediate result.

20. The method of claim 19 , wherein generating the final result of the multi-operand floating-point multiply operation includes

executing, at the floating point multiplier, a special purpose floating point multiply accumulate instruction with respect to the intermediate result and at least one floating point operand.