Browse Prior Art Database

Visualization and Optimization of Hot Spot Distribution

IP.com Disclosure Number: IPCOM000023759D
Original Publication Date: 2004-Apr-25
Included in the Prior Art Database: 2004-Apr-25
Document File: 1 page(s) / 19K

Publishing Venue

Siemens

Related People

Juergen Carstens: CONTACT

Abstract

Some chipcard NVM (Non-Volatile Memory) technologies have restrictions or preferences on where to place data (sector, page) which are often modified regarding the endurance ('hot spots') of NVM pages. To optimize the endurance an effective allocation of hot spots is necessary. Up to now the user has to identify and allocate these hot spots manually by analyzing the application and allocate one hot spot after the other. This is not only a lot of work but it is also susceptible to mistakes. It is therefore proposed to modify the software simulator as well as the linker to allow them to do this work automatically. An existing simulator can be extended with a detailed NVM model which counts every write access while running the application and showing the physical memory structure and the contents of the counters. The user can set a certain threshold above which a page is considered a hot spot and displayed in a different color. With this support the user is able to immediately identify these hot spots.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 74% of the total text.

Page 1 of 1

S

Visualization and Optimization of Hot Spot Distribution

Idea: Juergen Duve, FR-Sophia Antipolis

Some chipcard NVM (Non-Volatile Memory) technologies have restrictions or preferences on where to place data (sector, page) which are often modified regarding the endurance ('hot spots') of NVM pages. To optimize the endurance an effective allocation of hot spots is necessary. Up to now the user has to identify and allocate these hot spots manually by analyzing the application and allocate one hot spot after the other. This is not only a lot of work but it is also susceptible to mistakes.

It is therefore proposed to modify the software simulator as well as the linker to allow them to do this work automatically. An existing simulator can be extended with a detailed NVM model which counts every write access while running the application and showing the physical memory structure and the contents of the counters. The user can set a certain threshold above which a page is considered a hot spot and displayed in a different color. With this support the user is able to immediately identify these hot spots.

For optimization of allocation there is an extended linker that now has data about the architecture and special requirements of the NVM for allocating the hot spots in an efficient way. Therefore a new keyword or identifier has to be introduced to mark the hot spots for the linker. This has to be done on source code level. These variables can then be handled differently by...