Providing code sections for matrix of arithmetic logic units in a processor
The present invention relates to a processor having a trace cache and a plurality of ALUs arranged in a matrix, comprising an analyser unit located between the trace cache and the ALUs, wherein the analyser unit analyses the code in the trace cache, detects loops, transforms the code, and issues to the ALUs sections of the code combined to blocks for joint execution for a plurality of clock cycles.
1. A processor having at least one core, the least one core comprising:
a plurality of ALUs, at least some of the plurality of ALUs arranged in a multi-dimensional array;
an instruction fetcher configured to fetch instructions of a program;
an instruction decoder configured to decode the fetched instructions of the program, the decoded instructions comprising at least data load instructions for loading data and data processing instructions for processing at least some of the data loaded according to the data load instructions;
an instruction scheduler configured to issue the decoded instructions in order of the program; and
at least one load unit configured to receive at least some of the data load instructions issued by the instruction scheduler and to provide, according to the received data load instructions, loaded data to the multi-dimensional array of the at least some ALUs as operand data to be processed according to the data processing instructions,
the at least some of the ALUs arranged in the multi-dimensional array configured to receive at least some of the data processing instructions issued from the instruction scheduler and to process, according to the received data processing instructions, at least some of the loaded data provided by the at least one load unit,
wherein each of the at least some ALUs arranged in the multi-dimensional array comprises at least two inputs for operands, each execution unit being adapted to, in response to receiving the instruction:
is adapted to wait for operands identifiable as its operands to be processed according to its received data processing instructions, to arrive,
is adapted to process its arrived identified operands according to its received data processing instructions, and
is adapted to then at least one of:
store a result of the processing as an identifiable result in a register, or
forward the result of the processing as the identifiable result to another ALU.
2. The processor of claim 1 , wherein the plurality of ALUs execute the instructions in order of operand availability.
3. The processor of claim 2 , wherein the instruction scheduler issues the decoded instructions in program order.
4. The processor of claim 3 , wherein the instruction scheduler is configured to select a particular ALU of the plurality of ALUs for instruction issue according to states received by the instruction scheduler from the plurality of ALUs.
5. The processor of claim 3 , wherein the at least one core is implemented on a chip.
6. The processor of claim 1 , wherein the at least one load unit is at least one load and store unit.