INFORMATION PROCESSING APPARATUS, INFORMATION PROCESSING METHOD, AND STORAGE MEDIUM

• US 20130145373A1
• Filed: 11/23/2012
• Published: 06/06/2013
• Est. Priority Date: 12/01/2011
• Status: Active Grant

• Current Assignee: Canon Ayutthaya Limited (Canon Inc.)
• Original Assignee: Canon Ayutthaya Limited (Canon Inc.)
• Inventors: Noro, Hideo
• Granted Patent: US 9,274,831 B2
• Time in Patent Office: Days
• US Class Current: 718/102
• CPC Class Codes: G06F 9/30087 Synchronisation or serialisation instructions
• CPC Class Codes: G06F 9/3889 controlled by multiple instructions, e.g. MIMD, decoupled access or execute
• CPC Class Codes: G06F 9/4881 Scheduling strategies for dispatcher, e.g. round robin, multi-level priority queues

Abstract

There is provided with an information processing apparatus for controlling execution of a plurality of threads which run on a plurality of calculation cores connected to a memory including a plurality of banks. A first selection unit is configured to select a thread as a continuing thread which receives data from other thread, out of threads which process a data group of interest, wherein the number of accesses for a bank associated with the selected thread is less than a predetermined count. A second selection unit is configured to select a thread as a transmitting thread which transmits data to the continuing thread, out of the threads which process the data group of interest.

Claims

1. An information processing apparatus for controlling execution of a plurality of threads which run on a plurality of calculation cores connected to a memory including a plurality of banks, said apparatus comprising:

a holding unit configured to hold the numbers of accesses by respective threads in a single calculation cycle for respective banks;

a first selection unit configured to select a thread as a continuing thread which receives data from other thread, out of threads which process a data group of interest, wherein the number of accesses for a bank associated with the selected thread is less than a predetermined count;

a second selection unit configured to select a thread as a transmitting thread which transmits data to the continuing thread, out of the threads which process the data group of interest;

a generation unit configured to generate an instruction which instructs the transmitting thread to write data in an area in a bank associated with the continuing thread, and an instruction which instructs the continuing thread to read data from the area in the bank associated with the continuing thread; and

an update unit configured to increment the number of accesses to the bank associated with the continuing thread when said generation unit generates an instruction.

2. The apparatus according to claim 1, wherein each instruction includes

to be synchronously executed by threads designated by the instruction,

a read cycle for reading initial data onto a register held by the calculation core on which the threads run,

a calculation cycle for making calculations using data read from areas in banks associated in advance with the threads and data in a register held by the calculation core, and storing a calculation result in the register held by the calculation core, and

a write cycle for writing data stored in the register held by the calculation core, in the memory.

3. The apparatus according to claim 2, wherein said first selection unit and said second selection unit are further configured to repetitively select a set of the continuing thread and the transmitting thread which exchange data in a pair of the write cycle and the subsequent calculation cycle out of the plurality of threads which process the data group of interest.

4. The apparatus according to claim 2, wherein said generation unit is further configured to generate an instruction which instructs the transmitting thread to write data in an area in a bank associated with the continuing thread in the write cycle, and an instruction which instructs the continuing thread to read data from the area in the bank associated with the continuing thread in the calculation cycle that follows the write cycle.

5. The apparatus according to claim 1, wherein said first selection unit is further configured to select a thread as the continuing thread, wherein the number of accesses for the bank associated with the selected thread is smaller.

6. The apparatus according to claim 1, wherein said second selection unit is further configured to select a thread as the transmitting thread, wherein the number of accesses for the bank associated with the selected thread is larger.

8. A non-transitory computer-readable storage medium storing a computer program for controlling a computer to function as respective units of an information processing apparatus of claim 1.

7. An information processing method for controlling execution of a plurality of threads which run on a plurality of calculation cores connected to a memory including a plurality of banks, in view of the numbers of accesses by respective threads in a single calculation cycle for respective banks, said method comprising the steps of:

selecting a thread as a continuing thread which receives data from other thread, out of threads which process a data group of interest, wherein the number of accesses for a bank associated with the selected thread is less than a predetermined count;

selecting a thread as a transmitting thread which transmits data to the continuing thread, out of the threads which process the data group of interest;

generating an instruction which instructs the transmitting thread to write data in an area in a bank associated with the continuing thread, and an instruction which instructs the continuing thread to read data from the area in the bank associated with the continuing thread; and

incrementing the number of accesses to the bank associated with the continuing thread when the instruction is generated in said generating step.