IP Library Granted Patent US 7,346,760
Granted Patent B2
US 7,346,760 · App. 09/855,594 · Granted Mar 18, 2008

Data processing apparatus of high speed process using memory of low speed and low power consumption

View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 7,346,760
App. No.
09/855,594
Filed
May 16, 2001
Granted
Mar 18, 2008
Kind
B2
Art Unit
2183
USPC
712/200
Abstract

When fetching an instruction from a plurality of memory banks, a first pipeline cycle corresponding to selection of a memory bank and a second pipeline cycle corresponding to instruction readout are generated to carry out a pipeline process. Only the selected memory bank can be precharged to allow reduction of power consumption. Since the first and second pipeline cycles are effected in parallel, the throughput of the instruction memory can be improved.

Claims (27)

1. A data processing apparatus comprising:

an instruction memory in which an instruction is stored;

a data memory in which data is stored;

an instruction decoder decoding a fetched instruction;

a memory operation unit coupled to said instruction memory, said data memory and said instruction decoder, fetching an instruction stored in said instruction memory, and accessing said data memory according to a decode result of said instruction decoder; and

an integer operation unit carrying out an integer operation according to a decode result of said instruction decoder,

said instruction memory including a plurality of instruction memory banks,

said memory operation unit generating a pipeline cycle corresponding to selection of an instruction memory bank to be accessed in a following pipeline cycle and a pipeline cycle corresponding to an access to an instruction memory bank without any accesses to other instruction memory banks to carry out low power consumption pipeline processing when a plurality of instructions are fetched from the plurality of instruction memory banks, wherein

said instruction memory further includes a high speed instruction memory, and

said memory operation unit generates a pipeline cycle corresponding to instruction readout to carry out a pipeline process when fetching an instruction from said high speed instruction memory, and performs a pipeline control having a same throughput and a different latency compared to the access to said instruction memory when accessing to said high speed instruction memory,

said data memory includes a plurality of data memory banks,

said memory operation unit generates a pipeline cycle corresponding to selection of a data memory bank and a pipeline cycle corresponding to data access to carry out a pipeline process when accessing said plurality of data memory banks,

said data memory further includes a second bank select circuit decoding an address including a high order address to generate chip select signals of said plurality of data memory banks in order to divide said plurality of data memory banks into two different regions, and

said second bank select circuit decodes an address including a low order address to generate chip select signals of said plurality of data memory banks so that a different data memory bank in said plurality of data memory banks is accessed when data at continuous addresses in said two different regions are accessed.

2. A data processing apparatus comprising:

an instruction memory in which an instruction is stored;

a data memory in which data is stored;

an instruction decoder decoding a fetched instruction;

a register file;

a memory operation unit coupled to said instruction memory, said data memory and said instruction decoder, fetching an instruction stored in said instruction memory, and accessing said data memory according to a decoded result of said instruction decoder; and

an integer operation unit carrying out an integer operation according to a decoded result of said instruction decoder,

said memory operation unit retaining an instruction in a loop of instructions corresponding to a repeat instruction in a dedicated register in said register file when said repeat instruction is executed, and executing the loop of instructions while fetching the instruction retained in said dedicated register;

wherein said memory operation unit generates a pipeline cycle corresponding to selection of an instruction memory bank to be accessed in a following pipeline cycle and a pipeline cycle corresponding to an access to an instruction memory bank without any accesses to other instruction memory banks to carry out a low power consumption pipeline process when fetching an instruction from said plurality of instruction memory banks, wherein said register file comprises a processor status word,

wherein said memory operation unit:

refers to a first flag indicating a first execution cycle of said loop in said processor status word at the first execution cycle, and retains, in the register of said register file, the instruction in the loop of instructions fetched from said instruction memory when said repeat instruction is executed, and

refers to a second flag indicating a second execution cycle to a last execution cycle of said loop in said processor status word at the second execution cycle to the last execution cycle, and executes said loop while fetching the instruction retained in said dedicated register.

3. The data processing apparatus according to claim 2 , wherein said memory operation unit saves a plurality of registers including said dedicated register and switches a task in a task switch operation.

Assignments (1)
CHANGE OF ADDRESS Recorded Nov 29, 2017
From: RENESAS ELECTRONICS CORPORATION
To: RENESAS ELECTRONICS CORPORATION
Reel/Frame 044928/0001 →