Semiconductor integrated circuit design supporting apparatus, method, and program

A semiconductor integrated circuit design supporting apparatus, configured to cooperate with a high level synthesis means for executing high level synthesis of analyzing a behavioral description, which describes a circuit behavior of an integrated circuit by high level language, to generate register transfer level (RTL) description, the semiconductor integrated circuit design supporting apparatus being further configured to optimize an inserting position and a number of inserting stages of flip-flops (FFs) to be inserted in the RTL description through the high level synthesis by utilizing, in cooperation with the high level synthesis means, a scheduling result including latency of each output pin of each module of the integrated circuit and information on delay between signal lines at each clock boundary inside the each module, the scheduling result being output as an intermediate result when the high level synthesis means performs the high level synthesis, the semiconductor integrated circuit design supporting apparatus having a computer with a central processor unit configured to provide parts comprising:

• a module analyzing part for extracting modules to be subjected to latency adjustment from the behavioral description;
• a latency provisionally adjusting part for calculating, as a number of stages of pre-optimization inserting FFs, a number of FFs required to be inserted for latency adjustment between the modules to be subjected to latency adjustment;
• an FF inserting position candidate selecting part for selecting FF inserting position candidates inside a former-stage module to be subjected to latency adjustment and a latter-stage module to be subjected to latency adjustment in which the number of stages of pre-optimization inserting FFs, which is calculated by the latency provisionally adjusting part, is inserted, the FF inserting position candidates enabling achievement of the same timing and function as a case where the number of stages of pre-optimization inserting FFs is inserted between modules;
• an FF inserting position determining part for selecting, from the FF inserting position candidates selected by the FF inserting position candidate selecting part, an inserting position that enables one of reduction in scale of the entire integrated circuit and improvement in processing efficiency of the entire integrated circuit, and specifying the inserting position of an FF required to be inserted inside a module as a substitute for the pre-optimization inserting FFs;
• and
• a high level synthesis repeating part for adding, to the scheduling result, the FF that is specified to be inserted by the FF inserting position determining part so as to optimize the scheduling result, and then cooperating with the high level synthesis means again to generate the optimized RTL description based on the optimized scheduling result.

A semiconductor integrated circuit design supporting method, comprising:

• a high level synthesis step of executing high level synthesis processing of analyzing a behavioral description, which is stored in a storage part and describes a circuit behavior of an integrated circuit by high level language, to generate an RTL description,
• the semiconductor integrated circuit design supporting method optimizing an inserting position and a number of inserting stages of flip-flops (FFs) to be inserted in the RTL description through the high level synthesis processing by utilizing, in cooperation with the high level synthesis step, a scheduling result including latency of each output pin of each module of the integrated circuit and information on delay between signal lines at each clock boundary inside the each module, the scheduling result being output as an intermediate result when the high level synthesis processing is performed in the high level synthesis step, the semiconductor integrated circuit design supporting method further using a computer with a central processor unit to perform steps including:
• a module analyzing step of extracting modules to be subjected to latency adjustment from the behavioral description stored in the storage part;
• a latency provisionally adjusting step of calculating, as a number of stages of pre-optimization inserting FFs, a number of FFs required to be inserted for latency adjustment between the modules to be subjected to latency adjustment;
• an input latch FF determining step of extracting an input latch pin that is latched by an FF without interposing logic in a latter-stage module having, as an input, the FF that is required to be inserted, and calculating, as a number of stages of input latch FFs, a number of stages of input latch FFs that latch the input latch pin;
• a latency re-adjusting step of specifying, based on a relationship between the number of stages of pre-optimization inserting FFs, which is calculated in the latency provisionally adjusting step, and the number of stages of input latch FFs, which is calculated in the input latch FF determining step, the input latch FF and the FF that is required to be inserted, which cancel each other out, to thereby reduce a number of stages of FFs;
• and
• a high level synthesis repeating step of deleting, from the scheduling result, the FF specified to be reduced in the latency re-adjusting step so as to optimize the scheduling result, and then cooperating with the high level synthesis step again to generate the optimized RTL description based on the optimized scheduling result.

A semiconductor integrated circuit design supporting method, comprising:

• a high level synthesis step of executing high level synthesis processing of analyzing a behavioral description, which is stored in a storage part and describes a circuit behavior of an integrated circuit by high level language, to generate an RTL description,
• the semiconductor integrated circuit design supporting method of optimizing an inserting position and a number of inserting stages of flip-flops (FFs) to be inserted in the RTL description through the high level synthesis processing by utilizing, in cooperation with the high level synthesis step, a scheduling result including latency of each output pin of each module of the integrated circuit and information on delay between signal lines at each clock boundary inside the each module, the scheduling result being output as an intermediate result when the high level synthesis processing is performed in the high level synthesis step, the semiconductor integrated circuit design supporting method further using a computer with a central processor unit to perform steps including:
• a module analyzing step of extracting modules to be subjected to latency adjustment from the behavioral description stored in the storage part;
• a latency provisionally adjusting step of calculating, as a number of stages of pre-optimization inserting FFs, a number of FFs required to be inserted for latency adjustment between the modules to be subjected to latency adjustment;
• an FF inserting position candidate selecting step of selecting FF inserting position candidates inside a former-stage module to be subjected to latency adjustment and a latter-stage module to be subjected to latency adjustment in which the number of stages of pre-optimization inserting FFs, which is calculated in the latency provisionally adjusting step, is inserted, the FF inserting position candidates enabling achievement of the same timing and function as a case where the number of stages of pre-optimization inserting FFs is inserted between modules;
• an FF inserting position determining step of selecting, from the FF inserting position candidates selected in the FF inserting position candidate selecting step, an inserting position that enables one of reduction in scale of the entire integrated circuit and improvement in processing efficiency of the entire integrated circuit, and specifying the inserting position of an FF that is required to be inserted inside a module as a substitute for the pre-optimization inserting FFs;
• and
• a high level synthesis repeating step of adding, to the scheduling result, the FF that is specified to be inserted in the FF inserting position determining step so as to optimize the scheduling result, and then cooperating with the high level synthesis step again to generate the optimized RTL description based on the optimized scheduling result.

A non-transitory computer readable medium containing a computer readable semiconductor integrated circuit design supporting program, the semiconductor integrated circuit design supporting program causing the computer with a central processor unit to perform steps including:

• a module analyzing step of extracting modules to be subjected to latency adjustment from a behavioral description;
• a latency provisionally adjusting step of calculating, as a number of stages of pre-optimization inserting flip flops (FFs), a number of FFs required to be inserted for latency adjustment between the modules to be subjected to latency adjustment;
• an input latch FF determining step of extracting an input latch pin that is latched by an FF without interposing logic in a latter-stage module having, as an input, the FF that is required to be inserted, and calculating, as a number of stages of input latch FFs, a number of stages of latch FFs that latch the input latch pin;
• a latency re-adjusting specifying, based on a relationship between the number of stages of pre-optimization inserting FFs, which is calculated by the latency provisionally adjusting step and the number of stages of input latch FFs, which is calculated by the input latch FF determining step the input latch FF and the FF that is required to be inserted, which cancel each other out, to thereby reduce a number of stages of FFs;
• and
• deleting, from a scheduling result, the FF specified to be reduced by the latency re-adjusting step—so as to optimize the scheduling result, and then cooperating with a high level synthesis means to generate an optimized register transfer level (RTL) description based on the optimized scheduling result.

A non-transitory computer readable medium containing a computer readable semiconductor integrated circuit design supporting program, the semiconductor integrated circuit design supporting program causing the computer with a central processor unit to perform steps including:

• a module analyzing step of extracting modules to be subjected to latency adjustment from a behavioral description;
• a latency provisionally adjusting step of calculating, as a number of stages of pre-optimization inserting flip flops (FFs), a number of FFs required to be inserted for latency adjustment between the modules to be subjected to latency adjustment;
• a FF inserting position candidate selecting step of selecting FF inserting position candidates inside a former-stage module to be subjected to latency adjustment and a latter-stage module to be subjected to latency adjustment in which the number of stages of pre-optimization inserting FFs, which is calculated by the latency provisionally adjusting step, is inserted, the FF inserting position candidates enabling achievement of the same timing and function as a case where the number of stages of pre-optimization inserting FFs subjected is inserted between modules;
• a FF inserting position determining step of selecting, from the FF inserting position candidates selected by the FF inserting position candidate selecting step, an inserting position that enables one of reduction in scale of the entire integrated circuit and improvement in processing efficiency of the entire integrated circuit, and specifying the inserting position of an FF that is required to be inserted inside a module as a substitute for the pre-optimization inserting FFs;
• and
• a high level synthesis repeating step of adding, to the scheduling result, the FF that is specified to be inserted by the FF inserting position determining step so as to optimize the scheduling result, and then cooperating with a high level synthesis means to generate an optimized register transfer level (RTL) description based on the optimized scheduling result.