Zbigniew Mudza
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Zbigniew Mudza has a total of 14 co-authors. Here are the top co-authors by number of shared publications:
- Rafał Kiełbik
Department of Microelectronics and Computer Science, Lodz University of Technology, 93-005 Lodz, Poland
4 shared publications - Jaroslaw Jung
Department of Molecular Physics, Faculty of Chemistry, Lodz University of Technology, 116 Żeromskiego Str., 90-924 Lodz, Poland
2 shared publications Andrzej Napieralski
1 shared publicationDominik Stożek
1 shared publication- Grzegorz Jablonski
Department of Microelectronics and Computer Science, Lodz University of Technology, ul. Wólczańska 221, 93-005 Lodz, Poland
1 shared publication - Jacek Ulanski
Department of Molecular Physics, Lodz University of Technology, Zeromskiego 116, 90-924 Lodz, Poland
1 shared publication
Latest Publications
- Open Access
- 16 April 2022
- Published by MDPI AG in Electronics
- Vol. 11 (8) , 1261
Mapping Processing Elements of Custom Virtual CGRAs onto Reconfigurable Partitions
FPGAs can provide application-specific acceleration for computationally demanding tasks. However, they are rarely considered general-purpose platforms due to low productivity of software development and long reconfiguration time. These problems can be mitigated by implementing a coarser overlay atop the FPGA fabric. Combining this approach with partial reconfiguration allows for the modification of individual processing elements (PEs) of the virtual architecture without altering the rest of the system. Module relocation can be used to share implementation details between functionally equivalent PEs that use identical sets of resources, thus eliminating redundant placement and routing runs. Proper floorplanning is crucial for virtual Coarse-Grained Reconfigurable Architectures (CGRAs) with relocatable PEs considering their tendency to use nearest-neighbor connection patterns. It requires solving two problems—finding identical regions in the FPGA fabric and assigning individual partitions to certain locations. This article presents minor improvements of a state-of-the-art solution for the first and proposes a novel technique for solving the other. The proposed automated floorplanner uses modified breadth-first search with direction-based penalties to create initial floorplan consistent with geometry of logical array, then improves the result with 2-opt local optimization. Compared to simulated annealing solutions, the proposed approach allows for the reduction in the floorplanning time by two to three orders of magnitude without compromising the quality of the results.
- 16 April 2022
- Published by MDPI AG in Electronics
- Vol. 11 (8) , 1261
Open AccessMapping Processing Elements of Custom Virtual CGRAs onto Reconfigurable Partitions
FPGAs can provide application-specific acceleration for computationally demanding tasks. However, they are rarely considered general-purpose platforms due to low productivity of software development and long reconfiguration time. These problems can be mitigated by implementing a coarser overlay atop the FPGA fabric. Combining this approach with partial reconfiguration allows for the modification of individual processing elements (PEs) of the virtual architecture without altering the rest of the system. Module relocation can be used to share implementation details between functionally equivalent PEs that use identical sets of resources, thus eliminating redundant placement and routing runs. Proper floorplanning is crucial for virtual Coarse-Grained Reconfigurable Architectures (CGRAs) with relocatable PEs considering their tendency to use nearest-neighbor connection patterns. It requires solving two problems—finding identical regions in the FPGA fabric and assigning individual partitions to certain locations. This article presents minor improvements of a state-of-the-art solution for the first and proposes a novel technique for solving the other. The proposed automated floorplanner uses modified breadth-first search with direction-based penalties to create initial floorplan consistent with geometry of logical array, then improves the result with 2-opt local optimization. Compared to simulated annealing solutions, the proposed approach allows for the reduction in the floorplanning time by two to three orders of magnitude without compromising the quality of the results.
- 24 June 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 94-99
Exploiting Design Modularity and Relocation to Increase Productivity in FPGA-based Computing Systems
Long development cycles are a crucial disadvantage of FPGA-based systems. Reusing implementation results of a module for multiple instances and across different designs can mitigate this issue. Identical relative placement and routing for multiple instances of a module can be forced in Xilinx 7 Series devices with strict design constraining. A reference instance of a module can be implemented in a certain section of FPGA fabric. Fixed placement and routing constraints can be extracted from the obtained results and relocated to any identical section of FPGA fabric. The approach is especially useful when applied to reconfigurable partitions – it supports independent development and implementation of reconfigurable modules and static design. Also, in some cases it can be extended to relocating entire partial bitstreams.
- 24 June 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 94-99
Exploiting Design Modularity and Relocation to Increase Productivity in FPGA-based Computing Systems
Long development cycles are a crucial disadvantage of FPGA-based systems. Reusing implementation results of a module for multiple instances and across different designs can mitigate this issue. Identical relative placement and routing for multiple instances of a module can be forced in Xilinx 7 Series devices with strict design constraining. A reference instance of a module can be implemented in a certain section of FPGA fabric. Fixed placement and routing constraints can be extracted from the obtained results and relocated to any identical section of FPGA fabric. The approach is especially useful when applied to reconfigurable partitions – it supports independent development and implementation of reconfigurable modules and static design. Also, in some cases it can be extended to relocating entire partial bitstreams.
- Open Access
- 10 September 2020
- Published by MDPI AG in Electronics
- Vol. 9 (9) , 1482
Methodology of Firmware Development for ARUZ—An FPGA-Based HPC System
ARUZ is a large scale, massively parallel, FPGA-based reconfigurable computational system dedicated primarily to molecular analysis. This paper presents a methodology for ARUZ firmware development that simplifies the process, offers low-level optimization, and facilitates verification. According to this methodology, firstly an expanded, generic, all-in-one VHDL description of variable Processing Elements (PEs) is developed manually. GCC preprocessing is then used to extract only the desired target functionality. A dedicated software instantiates and connects PEs in form of a scalable network, divides it into subsets for chips and generates its HDL description. As a result, individual HDL-coded specification, optimized for certain analysis, is provided for the synthesis tool. Code reuse and automated generation of up to 81% of the code economizes the workload. Using well-optimized VHDL for core description rather than High Level Synthesis eliminates unnecessary overhead. The PE network can be scaled inversely proportional to PEs complexity, in order to efficiently utilize available resources. Moreover, downscaling the problem makes verification during HDL simulations and testing the prototype systems easier.
- 10 September 2020
- Published by MDPI AG in Electronics
- Vol. 9 (9) , 1482
Open AccessMethodology of Firmware Development for ARUZ—An FPGA-Based HPC System
ARUZ is a large scale, massively parallel, FPGA-based reconfigurable computational system dedicated primarily to molecular analysis. This paper presents a methodology for ARUZ firmware development that simplifies the process, offers low-level optimization, and facilitates verification. According to this methodology, firstly an expanded, generic, all-in-one VHDL description of variable Processing Elements (PEs) is developed manually. GCC preprocessing is then used to extract only the desired target functionality. A dedicated software instantiates and connects PEs in form of a scalable network, divides it into subsets for chips and generates its HDL description. As a result, individual HDL-coded specification, optimized for certain analysis, is provided for the synthesis tool. Code reuse and automated generation of up to 81% of the code economizes the workload. Using well-optimized VHDL for core description rather than High Level Synthesis eliminates unnecessary overhead. The PE network can be scaled inversely proportional to PEs complexity, in order to efficiently utilize available resources. Moreover, downscaling the problem makes verification during HDL simulations and testing the prototype systems easier.
- 1 June 2020
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 117-123
Relocatable Partial Bitstreams For Virtual Overlay Architectures atop Field-Programmable Gate Arrays
Intermediate virtual architecture overlays atop physical FPGA chips provide convenient abstraction level, which can increase productivity in FPGA-targeted application development. Individual reconfigurable modules of the overlay can be reprogrammed independently using partial reconfiguration. Homogeneous reconfigurable modules can be programmed using common configuration data, on condition that appropriate implementation constraints and proper floorplanning of the virtual architecture are provided. This paper presents methodology that can be used to generate relocatable bitstreams for Xilinx 7 series FPGA devices. The methodology is based on using constraints to force Xilinx Vivado Design Suite tools to implement multiple reconfigurable partition in the same way. Partial Reconfiguration Flow is used to implement multiple variants of individually reconfigurable partitions and Isolation Design Flow is used for feed-through prevention.
- 1 June 2020
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 117-123
Relocatable Partial Bitstreams For Virtual Overlay Architectures atop Field-Programmable Gate Arrays
Intermediate virtual architecture overlays atop physical FPGA chips provide convenient abstraction level, which can increase productivity in FPGA-targeted application development. Individual reconfigurable modules of the overlay can be reprogrammed independently using partial reconfiguration. Homogeneous reconfigurable modules can be programmed using common configuration data, on condition that appropriate implementation constraints and proper floorplanning of the virtual architecture are provided. This paper presents methodology that can be used to generate relocatable bitstreams for Xilinx 7 series FPGA devices. The methodology is based on using constraints to force Xilinx Vivado Design Suite tools to implement multiple reconfigurable partition in the same way. Partial Reconfiguration Flow is used to implement multiple variants of individually reconfigurable partitions and Isolation Design Flow is used for feed-through prevention.
- 1 November 2018
- Published by Elsevier in Computer Physics Communications
- Vol. 232, 22-34
ARUZ — Large-scale, massively parallel FPGA-based analyzer of real complex systems
- 1 November 2018
- Published by Elsevier in Computer Physics Communications
- Vol. 232, 22-34
ARUZ — Large-scale, massively parallel FPGA-based analyzer of real complex systems