US2006253836A1PendingUtilityA1

Method for creating parallel processing system

Assignee: IPFLEX INCPriority: Jun 27, 2003Filed: Jun 25, 2004Published: Nov 9, 2006
Est. expiryJun 27, 2023(expired)· nominal 20-yr term from priority
Inventors:Shimura Hiroshi
G06F 30/30
45
PatentIndex Score
0
Cited by
0
References
0
Claims

Abstract

A definition file included in the present invention includes a plurality of parallel descriptions that respectively define a plurality of parallel processes performed independently. The plurality of parallel descriptions include a first parallel description showing a first parallel process with a plurality of data inputs including at least one data input into which output data of another parallel process is inputted, with data with the same latency from input in a parallel processing system are inputted into the plurality of data inputs. A forming method includes a first step of generating, based on a hardware library in which information on a plurality of types of elements is stored, hardware configuration information including circuit configurations for executing the parallel processes that include at least one of the plurality of types of elements; and a second step of adding a delay element to the hardware configuration information so that data with a same latency from input into the parallel processing system are inputted into the plurality of data inputs of the circuit configuration for executing the first parallel process, and therefore the hardware configuration information is generated by this method.

Claims

exact text as granted — not AI-modified
1 . A method for forming, in accordance with a definition file, a parallel processing system that includes a plurality of types of elements that operate in parallel, 
 the definition file including a plurality of parallel descriptions that respectively define a plurality of parallel processes performed independently, the plurality of parallel descriptions including a first parallel description showing a first parallel process with a plurality of data inputs including at least one data input into which output data of another parallel process is inputted,    and the method comprising: 
 a first step of generating, based on a hardware library in which information on the plurality of types of elements is stored, hardware configuration information including circuit configurations for executing the parallel processes defined by the parallel descriptions of the definition file, the circuit configurations including at least one of the plurality of types of elements; and  
 a second step of adding a delay element to the hardware configuration information so that data with a same latency from input into the parallel processing system are inputted into a plurality of data inputs of a circuit configuration for executing the first parallel process.  
   
     
     
         2 . The method for forming according to  claim 1 , wherein the parallel processing system is reconfigurable to different hardware configurations by changing connections between the plurality of types of elements and the hardware configuration information includes information showing a plurality of different hardware configurations.  
     
     
         3 . The method for forming according to  claim 1 , wherein the plurality of types of elements include a plurality of types of operation units of a scale whereby one operation unit is capable of processing parallel process defined by one parallel description of the definition file.  
     
     
         4 . The method for forming according to  claim 1 , wherein the plurality of types of elements include a plurality of types of operation units for executing different operations in byte or word units.  
     
     
         5 . The method for forming according to  claim 1 , wherein information on a number of cycles consumed by respective types of elements is stored in the hardware library, and 
 in the second step, the delay element corresponding to a number of cycles consumed by at least one of the plurality of types of elements is added.    
     
     
         6 . The method for forming according to  claim 1 , wherein the plurality of parallel descriptions include a second parallel description that defines a second parallel process including shared processing that is common to at least part of a third parallel process defined by a third parallel description, 
 in the first step, a shared circuit configuration including at least one of the plurality of types of elements is generated for the shared processing, and    in the second step, the delay element is added to a circuit configuration for executing a difference between the second parallel process and the shared processing as the circuit configuration for executing the first parallel process.    
     
     
         7 . An apparatus for forming, in accordance with a definition file, a parallel processing system that includes a plurality of types of elements that operate in parallel, 
 the definition file including a plurality of parallel descriptions that respectively define a plurality of parallel processes performed independently, the plurality of parallel descriptions including a first parallel description showing a first parallel process with a plurality of data inputs including at least one data input into which output data of another parallel process is inputted,    the apparatus comprising: 
 a first means for generating, based on a hardware library in which information on the plurality of types of elements is stored, hardware configuration information including circuit configurations for executing the parallel processes defined by the parallel descriptions of the definition file, the circuit configurations including at least one of the plurality of types of elements; and  
 a second means for adding a delay element to the hardware configuration information so that data with a same latency from input into the parallel processing system are inputted into a plurality of data inputs of a circuit configuration for executing the first parallel process.  
   
     
     
         8 . The apparatus according to  claim 7 , wherein the parallel processing system is reconfigurable to different hardware configurations by changing connections between the plurality of types of elements and the hardware configuration information includes information showing a plurality of different hardware configurations.  
     
     
         9 . A program product for causing a computer to execute a process for designing, in accordance with a definition file, a system that includes a plurality of types of elements that operate in parallel, 
 the definition file including a plurality of parallel descriptions that respectively define a plurality of parallel processes performed independently, the plurality of parallel descriptions including a first parallel description showing a first parallel process with a plurality of data inputs including at least one data input into which output data of another parallel process is inputted,    the process for designing comprising: 
 a first step of generating, based on a hardware library in which information on the plurality of types of elements is stored, hardware configuration information including circuit configurations for executing the parallel processes defined by the parallel descriptions of the definition file, the circuit configurations including at least one of the plurality of types of elements; and  
 a second step of adding a delay element to the hardware configuration information so that data with a same latency from input into the parallel processing system are inputted into a plurality of data inputs of a circuit configuration for executing the first parallel process.  
   
     
     
         10 . A computer-readable recording medium on which a definition file is stored, the definition file including a plurality of parallel descriptions that respectively define a plurality of parallel processes performed independently and in synchronization by a system that includes a plurality of elements that operate in parallel, the plurality of parallel descriptions including a first parallel description showing a first parallel process with a plurality of data inputs including at least one data input into which output data of another parallel process is inputted and showing that data with a same latency from input into the system are inputted into the plurality of data inputs.  
     
     
         11 . The recording medium according to  claim 10 , wherein the plurality of parallel descriptions respectively define a plurality of parallel processes to be executed in synchronization with a clock for operations by the plurality of elements.  
     
     
         12 . A method of simulating, based on a definition file, a system that includes a plurality of types of elements that operate in parallel, 
 the definition file including a plurality of parallel descriptions that respectively define a plurality of parallel processes preformed independently, the plurality of parallel descriptions including a first parallel description showing a first parallel process with a plurality of data inputs including at least one data input into which output data of another parallel process is inputted,    the method comprising a step of executing a plurality of parallel processes defined by the definition file in synchronization, wherein in the step of executing, data with a same latency from input in the system are inputted into the plurality of data inputs of the first parallel process.    
     
     
         13 . A simulator for simulating, based on a definition file, a system that includes a plurality of types of elements that operate in parallel, 
 the definition file including a plurality of parallel descriptions that respectively define a plurality of parallel processes performed independently, the plurality of parallel descriptions including a first parallel description showing a first parallel process with a plurality of data inputs including at least one data input into which output data of another parallel process is inputted,    the simulator comprising a means of executing a plurality of parallel processes defined by the definition file in synchronization, wherein in the means of executing, data with a same latency from input in the system are inputted into the plurality of data inputs of the first parallel process.    
     
     
         14 . A program product for causing a computer to simulate, based on a definition file, a system that includes a plurality of types of elements that operate in parallel, 
 the definition file including a plurality of parallel descriptions that respectively define a plurality of parallel processes preformed independently, the plurality of parallel descriptions including a first parallel description showing a first parallel process with a plurality of data inputs including at least one data input into which output data of another parallel process is inputted,    wherein when the computer simulates a step of executing a plurality of parallel processes defined by the definition file in synchronization, data with a same latency from input in the system are inputted into the plurality of data inputs of the first parallel process.

Join the waitlist — get patent alerts

Track US2006253836A1 — get alerts on status changes and closely related new filings.

We store only your email — no account needed. See our privacy policy.