| ||||||||||||||||||||
|
|
|
|
||||||||||||||||||
|
|
|
|
||||||||||||||||||
|
|
||||||||||||||||||||
|
|
||||||||||||||||||||
![]() ![]() Audio Designline Automotive Designline CommsDesign Digital Home Designline DSP Designline EDA Designline EEDesign EETimes Embedded Green Supply Line EET Supply Network Industrial Control Designline Mobile Handset Designline Teardown.com TechOnline PL Designline Planet Analog Power Managment DL RF Designline Semiconductor Insights Silicon Strategies VI Designline Webinar Wireless Designline Analog Europe Automotive Europe Power Management Europe Automotive Europe
|
|
Micro Digital Inc. recently announced that the GoFast floating point library has been ported to the Thumb-2 instruction set for Cortex processors. GoFast for Thumb-2 was designed for high-performance operation in embedded applications and ease of use including "link and go" compatibility with the IAR compiler. GoFast provides ROMable, reentrant, IEEE, and ANSI compatible Thumb-2 single- and double-precision floating-point support.
GoFast is a family of high-performance, reentrant, floating-point libraries that are ANSI C compatible, IEEE 754 compliant. It's designed to directly replace a C compiler's runtime floating-point library. GoFast boosts the performance of an application's math calculations and, in many cases, eliminates the need for hardware floating-point coprocessors. This reduces product manufacturing cost.
Basic floating point functions in the EWARM library, such as add and sub, are coded efficiently, so GoFast offers no speed improvement for them. However, for more complex functions such as square root, exponent, log, trigonometric, hyperbolic, and others, the performance gain from GoFast is typically 3:1 and as high as 10:1. This can result in significant cost savings for GPS, surveying, guidance, and other applications doing intensive trigonometric calculations by reducing hardware cost for a required performance level. Many other types of calculations involve square roots, exponents, logs, and hyperbolic functions, which will also benefit greatly from GoFast.
The accuracy of each GoFast Floating Point Library is within one (least significant) bit for arithmetic functions and two bits for transcendental functions, in most cases. The IEEE 754 Floating Point Format defines special representations for underflow, overflow, and invalid operation. The GoFast routines use these formats and adhere to the IEEE 754 error handling procedures in all applicable cases. Quality assurance and testing procedures guarantee proper product operation. In addition, each delivery includes target specific test programs assuring confidence of product operation.
It's available now for $3500 for royalty-free, one-product license, with full source code, and 90-day support. For more information and performance tables, visit www.smxrtos.com/ussw/gofast.htm.
| ||||||||||||||||||
|
||||||||||||||||||||
|
|
||||||||||||||||||||
|
All materials on this site Copyright © 2010 TechInsights, a Division of United Business Media LLC. All rights reserved. Privacy Statement | Terms of Service. |
||||||||||||||||||||