SMASH: High-Level Synthesis of Memory-Intensive Hardware Systems

A thesis by Pravil Gupta from the University of Southern California addressing high-level synthesis for memory-intensive application-specific systems. It describes the SMASH program, which combines datapath synthesis with hierarchical storage architecture design under cost and performance constraints. Experimental validation includes a case study on rapid prototyping of a JPEG image compression system.

Use Cases

• Evaluating area-delay trade-offs for hardware components based on the described synthesis techniques.
• Designing hierarchical storage architectures for memory-intensive systems based on the described methodology.
• Integrating synthesis tools into a system-level design flow, as illustrated by the JPEG compression experiment.

Strengths

• Thesis is based on a specific synthesis program (SMASH) with defined tasks like datapath synthesis and storage hierarchy design.
• Experimental results validate the techniques, including a concrete application to JPEG image compression system prototyping.

Limitations

• Description metadata is limited; actual data quality requires manual inspection after download.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count and file formats are unknown, which may limit suitability assessment.

Provenance

Source

University of Southern California thesis by Pravil Gupta.

Collection Method

Research and development of the SMASH synthesis program.