The LOOPGEN IP collection provides fast hardware architectures for implementing nested loop structures. The collection comprises of three different architectures (variants), namely:
- HWLU, a mixed-level structural/RTL architecture,
- IXGENB, a behavioral-level and
- IXGENR, a high-performance, pure RTL description of a more generalized form of the architecture.
Hardware looping architectures have potential uses for data-intensive processing in embedded systems. The implemented architectures are able to execute perfect loop nests without any cycle overhead for updating the iteration vector. Actually, successive last iterations of nested loops are collapsed in a single cycle.
LOOPGEN can be used as a ROYALTY-FREE component for use in your projects.
Interesting features and characteristics of LOOPGEN include:
- three different architectural variants
- support for any number of loops and datapath bitwidth
- single-cycle iteration vector update
- 201-243 MHz achieved clock rates on Xilinx Virtex-6.
The LOOPGEN IPs comprise of the following deliverables:
- Documentation in ASCII text, PDF, HTML formats
- Vendor-independent VHDL code for all architectural variants
- Configurable testbench
- HDL code generators for the HWLU and IXGENR architectures
- Various helper scripts for simulation (GHDL, Modelsim) and synthesis.
-------------- All users that will register and download LOOPGEN within 2013 are eligible to the following:
- FREE updates for the entire lifetime of the product
- FREE email support.
Pricing information and sample downloads:
Best regards, Nikolaos Kavvadias Hardware and EDA tools developer, Research Scientist Lamia, Fthiotis, Greece