Dissertation > Excellent graduate degree dissertation topics show

Optimized Design of the High Performance Floating-Point Multiply-Add-Fused Unit

Author: ZhangJingBo
Tutor: XingZuoCheng
School: National University of Defense Science and Technology
Course: Electronic Science and Technology
Keywords: Floating-point multiply-add units Semi-custom design optimization Full-custom design optimization Optimize the design process
CLC: TP332.2
Type: Master's thesis
Year: 2007
Downloads: 74
Quote: 1
Read: Download Dissertation


The floating-point multiply-add components (MAF) is one of the components of high-performance microprocessor core computing , its speed and power consumption has a great impact on the entire microprocessor performance . Study of semi-custom / full-custom the mixed design optimization to achieve high-performance floating-point multiply-add components with a wide range of application value and practical significance . X microprocessor floating point multiply-add components to study and optimize the design of low latency to support double-precision floating-point multiply-add structure based on using a combination of methods of semi-custom design optimization and full-custom design optimization , in design and optimization algorithm , coding , logic implementation structure , circuit layout levels , significantly enhance the performance of the floating-point multiply-add components . 0.13μm CMOS process , logic synthesis and layout simulation results show that the optimized floating-point multiply-add units operating frequency up to 1GHz, to the optimized design of the target . The results of this research include the following points : 1 . Analysis to study low-latency floating-point multiply-add components based on proposed full floating-point multiply-add components optimized design and optimize the design process summarized a floating-point multiply-add components . Improved addend on order shift calculation and processing order shifter implementation structure , development order shift amount with maximum parallelism of the order shifter so addend on order shift logic operation delay is reduced by 15% . 3 of the ten stops pipelined floating - point multiply-add components , so that the proportion of moderate to optimize the design of full-custom design , to ensure the feasibility of the design optimization objectives . Logically divided in a logical progression , and fine-tune stations logic delay to balancing stations , floating-point multiply-add components to achieve optimal performance . Floating - point multiply-add components optimize the key sub- module of the full-custom design . Full-custom design to achieve high-performance the 4-2 compressor standard cell , 52 or doors and AND gate 108 . ,4- 2 compression maximum delay in the worst conditions of 317ps and maximum delay of 52 -bit or doors 147ps , 193ps maximum delay of 108 bit and doors . Optimize the design of full-custom design module used in the floating-point multiply-add components to achieve a better optimization results .

Related Dissertations

  1. Design of the Lower Power Arithmetic and Logical Unit,TP332
  2. Optimized Design and Physical Implementation of the High Performance Trap Logic Unit,TP332
  3. Design of Wearable Computer’ Multiprocessor Based on SOPC,TP332
  4. Research of Processor Performance Optimization Based on Function Dynamic Reuse,TP332
  5. Design and Implementation of Processor TI DSP VC33 Virtual Kernel,TP332
  6. The Research on A Scheduling Algorithm for Real-time Tasks in Multi-core Systems,TP332
  7. CPU overhead virtual computing environment measurement system,TP332
  8. Dynamic analysis based on multi-core acceleration method,TP332
  9. Multiprocessor rate monotonic scheduling algorithm,TP332
  10. Optical vector matrix multiplier theory and method of,TP332.22
  11. Design and Implementation of Fault Injection Tools Targeted at Itanium 2 Processors,TP332
  12. Research and Implementation of MPI on CELL Broad Engine,TP332
  13. Design and Verification of an 8-bit MCU Based on Wishbone Bus,TP332
  14. FPGA-based Design and Implement of 16-bit Microprocessor for Teaching,TP332
  15. Implementation of Soft-Core Processor and DDFS Based on FPGA,TP332.1
  16. The Design of FPU Based on PowerPC,TP332
  17. 6502 -compatible microprocessor IP Development and Design,TP332
  18. Research on Abstraction for Model Checking Cache Coherence Protocol,TP332
  19. Accelerating Function and Timing Verification Convergence on L2-Cache of Multi-core Processor,TP332
  20. Program optimization based on multi-core systems technology research,TP332
  21. The Design of 8-bit RISC MPU,TP332

CLC: > Industrial Technology > Automation technology,computer technology > Computing technology,computer technology > Electronic digital computer (not a continuous role in computer ) > Arithmetic unit and the controller (CPU) > Arithmetic unit
© 2012 www.DissertationTopic.Net  Mobile