Extending fused multiply-add instructions
Extending fused multiply-add instructions, the method comprising: receiving an extended fused multiply-add (FMA) instruction indicating one or more operands of a fused multiply-add (FMA) operation and one or more transformations to be applied to the one or more operands; and performing, based on the extended FMA instruction, the one or more transformations and the FMA operation.
1. A method of extending fused multiply-add instructions, the method comprising:
receiving an extended fused multiply-add (FMA) instruction comprising a first subset of bits indicating each operand of a fused multiply-add (FMA) operation, a second subset of bits indicating one or more transformations applicable to one or more operands of the FMA operation, and a parameter indicating whether or not the one or more transformations will be performed; and
performing, based on the extended FMA instruction and the parameter, the one or more transformations and the FMA operation.
2. The method of claim 1 , wherein performing the one or more transformations and the FMA operation comprises performing the one or more transformations based on one or more parameters of the extended FMA instruction indicating the one or more transformations.
3. The method of claim 1 , wherein performing the one or more transformations and the FMA operation comprises loading, from one or more registers, data describing the one or more transformations.
4. The method of claim 1 , wherein the extended FMA instruction is one of a plurality of extended FMA instructions each corresponding to a different combination of transformation-operand groupings.
5. The method of claim 1 , wherein the one or more transformations comprise a bit masking transformation.
6. The method of claim 1 , wherein the one or more transformations comprise a bit shifting operation.
7. The method of claim 1 , wherein the one or more transformations comprises a rounding operation.
8. The method of claim 1 , wherein the one or more transformations comprises a casting operation.
9. The method of claim 1 , wherein the one or more transformations comprises a conditional zeroing operation.
10. An apparatus for extending fused multiply-add instructions, the apparatus configured to perform steps comprising:
receiving an extended fused multiply-add (FMA) instruction comprising a first subset of bits indicating each operand of a fused multiply-add (FMA) operation, a second subset of bits indicating one or more transformations applicable to one or more operands of the FMA operation, and a parameter indicating whether or not the one or more transformations will be performed; and
performing, based on the extended FMA instruction and the parameter, the one or more transformations and the FMA operation.
11. The apparatus of claim 10 , wherein performing the one or more transformations and the FMA operation comprises performing the one or more transformations based on one or more parameters of the extended FMA instruction indicating the one or more transformations.
12. The apparatus of claim 10 , wherein performing the one or more transformations and the FMA operation comprises loading, from one or more registers, data describing the one or more transformations.
13. The apparatus of claim 10 , wherein the extended FMA instruction is one of a plurality of extended FMA instructions each corresponding to a different combination of transformation-operand groupings.
14. The apparatus of claim 10 , wherein the one or more transformations comprise a bit masking transformation.
15. The apparatus of claim 10 , wherein the one or more transformations comprise a bit shifting operation.
16. The apparatus of claim 10 , wherein the one or more transformations comprises a rounding operation.
17. The apparatus of claim 10 , wherein the one or more transformations comprises a casting operation.
18. The apparatus of claim 10 , wherein the one or more transformations comprises a conditional zeroing operation.
19. A computer program product disposed upon a non-transitory computer readable medium, the computer program product comprising computer program instructions for extending fused multiply-add instructions that, when executed, cause a computer system to perform steps comprising:
receiving an extended fused multiply-add (FMA) instruction comprising a first subset of bits indicating each operand of a fused multiply-add (FMA) operation, a second subset of bits indicating one or more transformations applicable to one or more operands of the FMA operation, and a parameter indicating whether or not the one or more transformations will be performed; and
performing, based on the extended FMA instruction and the parameter, the one or more transformations and the FMA operation.
20. The computer program product of claim 19 , wherein performing the one or more transformations and the FMA operation comprises performing the one or more transformations based on one or more parameters of the extended FMA instruction indicating the one or more transformations.