Vector move instruction in a vector data processing system and method therefor
First Claim
1. A method for executing an instruction in a data processor, comprising the steps of:
- storing a first vector value in a vector register;
storing a digital data value in a processing element;
storing an enable value in a first storage circuit, the enable value corresponding to the processing element, the enable value selectively enabling the processing element to participate in execution of the instruction;
storing a history value in a second storage circuit, the history value corresponding to the processing element, the history value being used to selectively modify the enable value;
receiving the instruction at an input means;
decoding the first instruction to provide a first plurality of control values using an instruction decode means, the instruction decode means being coupled to the input means for receiving the instruction;
negating at least one of a plurality of bits of the first vector value stored in the vector register to provide a negated vector value in response to a first portion of the plurality of control values; and
storing a preselected one of a plurality of bits of the negated vector value in the second storage circuit as the history value in response to a second portion of the plurality of control values.
7 Assignments
0 Petitions
Accused Products
Abstract
A "vnmvh" instruction reduces a substantial number of instructions and the temporary use of a register in a software code which executes nested conditional constructs in a vector data processor (10). When the vnmvh instruction is executed, all processing elements in the vector data processor participate in the function regardless of a setting of a status bit (Vt bit) (FIG. 6). During execution of the vnmvh instruction, the least significant bits of vector register specified in an operand are negated and moved into a plurality of history bits (Vh bits) (FIG. 6). The functionality provided by execution of vnmvh instruction allows a user to execute a nested conditional construct efficiently and effectively.
37 Citations
18 Claims
-
1. A method for executing an instruction in a data processor, comprising the steps of:
-
storing a first vector value in a vector register; storing a digital data value in a processing element; storing an enable value in a first storage circuit, the enable value corresponding to the processing element, the enable value selectively enabling the processing element to participate in execution of the instruction; storing a history value in a second storage circuit, the history value corresponding to the processing element, the history value being used to selectively modify the enable value; receiving the instruction at an input means; decoding the first instruction to provide a first plurality of control values using an instruction decode means, the instruction decode means being coupled to the input means for receiving the instruction; negating at least one of a plurality of bits of the first vector value stored in the vector register to provide a negated vector value in response to a first portion of the plurality of control values; and storing a preselected one of a plurality of bits of the negated vector value in the second storage circuit as the history value in response to a second portion of the plurality of control values. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9)
-
-
10. A method for executing an instruction in a data processor, comprising the steps of:
-
storing a first vector value in a vector register; storing a plurality of digital data values in a plurality of processing elements, each of the plurality of digital data values corresponding to one of the plurality of processing elements; storing a plurality of enable values in a first storage circuit, each of the plurality of enable values corresponding to one of the plurality of processing elements and each of the plurality of enable values selectively enables a corresponding one of the plurality of processing elements to participate in execution of the instruction; storing a plurality of history value in a second storage circuit, each of the plurality of history values corresponding to one of the plurality of processing elements and each of the plurality of history values being used to selectively modify a corresponding one of the plurality of enable values; receiving the instruction at an input means; decoding the first instruction to provide a first plurality of control values using an instruction decode means, the instruction decode means being coupled to the input means for receiving the instruction; negating a portion of a plurality of bits of the first vector value stored in the vector register to provide a negated vector value in response to a first portion of the plurality of control values; and storing the negated vector value in the second storage circuit as the plurality of history values in response to a second portion of the plurality of control values. - View Dependent Claims (11, 12, 13, 14, 15, 16, 17, 18)
-
Specification