工作学习总结、个人生活感悟!

Variable Precision Floating Point Modules

上一篇 / 下一篇  2007-06-20 13:11:45

  • 文件大小: 2 MB
  • 文件版本: V2.0
  • 开发商: 本站原创
  • 文件来源: 本地
  • 界面语言: english
  • 授权方式: 破解
  • 运行平台: Win2000/WinXP

My FP library is not a free IP core,but it's very cheap for sale。

To Get Password,Please Contact:jianglw@ihep.ac.cn


Introduction

Note: we have posted a new zip file with fixes to float2fix and fix2float modules. December 9, 2005.

We have developed a library of fully parameterized hardware modules for variable precision floating-point arithmetic. The library includes modules for format control, arithmetic operators and conversion to and from any fixed-point format. All the modules have been implemented in VHDL and designed to run on different FPGA devices. They have been tested on reconfigurable computing engines fromAnnapolis Microsystems.

An important feature of these modules is that denormalize, and normalize/round are decoupled from the arithmetic operations. This gives the designer full control over where and when to normalize floating point calculations, and results in savings of area in their hardware implementation. The currently implemented arithmetic operators are floating point add/sub and floating point multiply. Floating Point Divide, Square Root, and Multiply Accumulate (MAC) are planned in the near future.


Floating Point Modules

Fixed-point to floating-point conversion (unsigned)

Fixed-point to floating-point conversion (signed)

Floating-point to fixed-point conversion (unsigned)

Floating-point to fixed-point conversion (signed)

Floating-point adder/subtracter

Floating-point multiplier

Floating-point round and normalizer


Download Area

Modules updated are: 

Updated Modules

parameterized_mux

parameterized_adder

parameterized_comparator

parameterized_variable_shifter_row

shift_adjust

 

Removed Modules

one_bit_comparator

mux2

shift_control

 

Added Modulesparameterized_shifter

To use the modules, first unzip the zip file to some directory. In this directory, there are two subdirectories, one is namedlibrary_modules, which contains the modules for format control, operators and conversion to and from any fixed-point format. The other is namedfloatlib; it contains all the necessary sub-modules. The dependencies of these modules can be found by click the hyperlink on the left side. To simulate, first compile all the vhd files in directoryfloatlib, then compile the module you want simulate in thelibrary_modules,configure library modules according to the dependencies.


My FP library is not a free IP core,but it's very cheap for sale。

To Get Password,Please Contact:jianglw@ihep.ac.cn

FPGA/CPLD器件价格查询

TAG: Floating Point Modules library IP

 

评分:0

我来说两句

显示全部

:loveliness: :handshake :victory: :funk: :time: :kiss: :call: :hug: :lol :'( :Q :L ;P :$ :P :o :@ :D :( :)

数据统计

  • 访问量: 1300
  • 日志数: 2
  • 文件数: 1
  • 建立时间: 2007-06-10
  • 更新时间: 2007-07-02

RSS订阅

Open Toolbar