Issuing instructions to multiple execution units
A single chip sequential processor comprising at least one ALU-Block, where said sequential processor is capable of maintaining its op-codes while processing data such as to overcome the necessity of requiring a new instruction in every clock cycle.
1. A microprocessor comprising:
an arrangement of a plurality of data processing units;
a plurality of data load units operative to load data from data memory;
an instruction fetch unit operative to fetch instructions from instruction memory;
an instruction decode unit operative to decode the fetched instructions for execution on the microprocessor; and
an instruction issue unit operative to issue the decoded instructions, in instruction order of source code, to the data processing units,
one or more of the data processing units operative to process the issued instructions upon arrival, at the one or more data processing units, of operand data of the data via a bus from a data load unit of the plurality of data load units or from a different data processing unit of the plurality of data processing units.
2. The microprocessor of claim 1 wherein the instruction issue unit is operative to issue instructions by linearly advancing through a two-dimensional array of the plurality of data processing units in a row-wise and column-wise order in which one or more opcodes are issued to a first row of the two-dimensional array of the plurality of data processing units in successive columns and then one or more successive opcodes are issued to a next row of the two-dimensional array of the plurality of data processing units in the successive columns.
3. The microprocessor of claim 2 wherein the one or more opcodes are issued to the first row of the two-dimensional array of the plurality of data processing units until the first row is completed or a termination is detected, and then the one or more successive opcodes are issued to the next row of the two-dimensional array of the plurality of data processing units.
4. The microprocessor of claim 1 wherein the microprocessor is implemented on a single silicon die.
5. The microprocessor of claim 1 wherein the microprocessor is implemented on a stack of silicon dies.
6. The microprocessor of claim 1 wherein the arrangement of the plurality of data processing units is an array having more than one dimension.
7. A method for processing instructions by a microprocessor, the method comprising:
loading data from data memory using a plurality of data load units;
fetching instructions from instruction memory using an instruction fetch unit;
decoding the fetched instructions for execution on the microprocessor using an instruction decode unit;
issuing the decoded instructions, in instruction order of source code, to a plurality of data processing units using an instruction issue unit; and
processing, by one or more of the plurality of data processing units, the issued instructions upon arrival, at the one or more data processing units, of operand data of the data via a bus from a data load unit of the plurality of data load units or from a different data processing unit of the plurality of data processing units.
8. The method of claim 7 wherein issuing the decoded instructions includes linearly advancing through a two-dimensional array of the plurality of data processing units in a row-wise and column-wise order in which one or more opcodes are issued to a first row of the two-dimensional array of the plurality of data processing units in successive columns and then one or more successive opcodes are issued to a next row of the two-dimensional array of the plurality of data processing units in the successive columns.
9. The method of claim 8 wherein the one or more opcodes are issued to the first row of the two-dimensional array of the plurality of data processing units until the first row is completed or a termination is detected, and then the one or more successive opcodes are issued to the next row of the two-dimensional array of the plurality of data processing units.
10. The method of claim 7 wherein the microprocessor is implemented on a single silicon die.
11. The method of claim 7 wherein the microprocessor is implemented on a stack of silicon dies.
12. The method of claim 7 wherein the plurality of data processing units are arranged in an array having more than one dimension.