US 5,669,011 A · Partially decoded instruction cache

Abstract

A microprocessor partially decodes instructions retrieved from main memory before placing them into the microprocessor's integrated instruction cache. Each storage location in the instruction cache includes two slots for decoded instructions. One slot controls one of the microprocessor's integer pipelines and a port to the microprocessor's data cache. A second slot controls the second integer pipeline or one of the microprocessor's floating point units. The instructions retrieved from main memory are decoded by a loader unit which decodes the instructions from the compact form as stored in main memory and places them into the two slots of the instruction cache entry according to their functions. In addition, auxiliary information is placed in the cache entry along with the instruction to control parallel execution as well as emulation of complex instructions. A bit in each instruction cache entry indicates whether the instructions in the two slots are independent, so that they can be executed in parallel, or dependent, so that they must be executed sequentially. Using a single bit for this purpose allows two dependent instructions to be stored in the slots of the single cache entry.

Claims16

Dependent claims

A microprocessor that includes at least first and second parallel-operating execution units such that each of the first and second execution units responds to periodic cycles of a system clock signal by executing instructions in parallel with the execution of instructions by the other of the first and second execution units such that the microprocessor executes instructions at a rate of greater than one instruction per system clock cycle, encoded instructions being retrieved from a main memory that is external to the microprocessor and that is connectable to the microprocessor via a system bus, the microprocessor comprising:

an instruction cache that includes a plurality of individual instruction cache entries, each individual instruction cache entry including a first storage slot for storing a first partially-decoded instruction and a second storage slot for storing a second partially-decoded instruction;
a bus interface unit connected to the system bus for sequentially retrieving first and second encoded instructions from the external main memory, the first encoded instruction having a length different than the length of the second encoded instruction, and
an instruction loader that is connected to the bus interface unit for sequentially translating the first encoded instruction to a first partially-decoded instruction and the second encoded instruction to a second partially-decoded instruction, the instruction loader being connected to the instruction cache for sequentially storing the first and second partially-decoded instructions in the first and second storage slots, respectively, of a selected instruction cache entry.

A microprocessor-implemented method of executing instructions, the microprocessor including at least first and second parallel-operating execution units such that each of the first and second execution units responds to periodic cycles of a system clock signal by executing instructions in parallel with the execution of instructions by the other of the first and second execution units such that the microprocessor executes instructions at a rate greater than one instruction per system clock cycle, encoded instructions being retrieved from a main memory that is external to the microprocessor and that is connectable to the microprocessor via a system bus, the microprocessor further including a plurality of instruction cache entries, each instruction cache entry including a first storage slot for storing a first partially-decoded instruction and a second storage slot for storing a second partially-decoded instruction, the method comprising:

retrieving first and second encoded instructions from the external main memory, the first encoded instruction having a length different than the length of the second encoded instruction;
sequentially decoding the first and second encoded instructions to provide first and second partially-decoded instructions;
sequentially storing the first and second partially-decoded instructions in the first and second storage slots, respectively, of a selected instruction cache entry.

A microprocessor-implemented method of executing instructions, the microprocessor including at least first and second parallel-operating execution units such that each of the first and second execution units responds to periodic cycles of a system clock signal by executing instructions in parallel with the execution of instructions by the other of the first and second execution units such that the microprocessor executes instructions at a rate greater than one instruction per system clock cycle, the microprocessor further including an instruction cache having a plurality of instruction cache entries, each instruction cache entry including a first storage slot for storing a first-partially decoded instruction and a second storage slot for storing a second partially-decoded instruction, the method comprising:

retrieving first and second encoded instructions from a main memory that is external to the microprocessor, the first encoded instruction having a length different than the length of the second encoded instruction;
determining whether the first encoded instruction will be executed by the first execution unit or by the second execution unit;
decoding the first and second encoded instructions to provide first and second partially-decoded instructions, respectively;
in the event that the determining step determines that the first encoded instruction will be executed by the first execution unit, storing the first partially-decoded instruction in the first storage slot of a selected instruction cache entry and the second partially-decoded instruction in the second storage slot of the selected instruction cache entry and, in the event that the determining step determines that the first encoded instruction will be executed by the second execution unit, storing the first partially-decoded instruction in the second storage slot of the selected instruction cache entry and the second-partially decoded instruction in the first storage slot of the selected instruction cache entry;
and
providing the partially-decoded instruction stored in the first storage slot to the first execution unit in parallel with providing the partially-decoded instruction stored in the second storage slot to the second execution unit.

3 more elements

A microprocessor-implemented method of executing instructions, the microprocessor including at least first and second parallel-operating execution units such that each of the first and second execution units responds to periodic cycles of a system clock signal by executing instructions in parallel with the execution of instructions by the other of the first and second execution units such that the microprocessor executes instructions at a rate greater than one instruction per system clock cycle, encoded instructions being retrieved from a main memory that is external to the microprocessor and that is connectable to the microprocessor via a system bus, the encoded instructions including encoded core instructions and encoded non-core instructions, the microprocessor further including a plurality of instruction cache entries, each instruction cache entry including a first storage slot for storing a first partially-decoded instruction and a second storage slot for storing a second partially-decoded instruction, the method comprising:

(a) retrieving first and second encoded instructions from the external main memory on sequential cycles of the system clock signal, the first encoded instruction having a length different than the length of the second encoded instruction;
(b) determining whether the first and second encoded instructions are encoded core instructions or encoded non-core instructions;
(c) in the event that both the first encoded instruction and the second encoded instruction are core instructions;
(i) sequentially decoding the first and second encoded core instructions to provide first and second partially-decoded core instructions;
(ii) sequentially storing the first and second partially-decoded core instructions in the first and second storage slots, respectively, of a selected instruction cache entry;
(iii) determining whether the first and second partially-decoded core instructions are independent such that the first and second partially-decoded core instructions may be executed in parallel, or are dependent such that the first and second partially-decoded core instructions must be executed sequentially;
and
(iv) in the event that the first and second partially-decoded core instructions are determined to be independent, simultaneously retrieving the first and second partially-decoded core instructions from the selected instruction cache entry for parallel execution by the first and second execution units, respectively;
and
(d) in the event that either the first encoded instruction or the second encoded instruction is an encoded non-core instruction, generating a corresponding sequence of partially-decoded core instructions for sequential execution by a selected one of the first and second execution units.

7 more elements

Dependencies are inferred from the claim text.

Partially decoded instruction cache

Life of this patent

Partially decoded instruction cache

Life of this patent

Abstract

Claims16