|
Tomas Evensen is vice president and general manager of the Wind River Tools product division and also Chief Technology Officer.
Evensen joined Wind River as part of the acquisition of Integrated Systems, Inc. (ISI) in 2000. He was the Vice President of Engineering at Diab-SDS which was acquired by ISI in 1996. He has more than 24 years of experience in the OS and embedded tools area. Prior to joining Wind River he was the creator of the Diab Data C/C++ compilers.
Evensen received his MSEE at the Royal Institute of Technology in Stockholm, Sweden.
|
 Tomas Evensen, CTO, Wind River |
Developers of Linux applications seemingly face a two pronged dilemma when trying to design application to make good use of multi-core hardware: use threads and sacrifice the memory protections provided by processes use in favor of supposedly reduced overhead, or use process and supposedly sacrifice performance for memory protection.
This talk will explore some of the myths and facts regarding the use of threads and processes, present some real world measurements to the cost involved in using both taken using a modified subset of LMBench tool, discuss some very real but rarely mentioned reasons developers choose threads over processes and conclude by presenting CoProc (http://github.com/gby/coproc/), a proof of concept library which provides share-as-needed middle ground between threads and processes using Linux clone(2) and mmap(2) system calls. |
Gilad Ben-Yossef, Co-Author of "Building Embedded Linux Systems", 2nd edition, O'Reilly Press
|
|
Gas Chromatography-Mass Spectrometry (GCMS) היא שיטה המשלבת את התכונות של כרומטוגרפיה גזית ו ספקטרומטריית מסות לזיהוי חומרים שונים מתוך מדגם מורכב. יישומים של מערכת זו יכולים להיות: ניתוח אקולוגי, איתור סמים, חקר תרופות, וכן זיהוי של דגימות בלתי ידועות. ניתן להגדיל את הרגישות של המערכת, בסדר גודל, ע"י הוספת מרכיב סופר-סוני. המרכיב הסופר סוני דורש בקר אלקטרוני מורכב, רב תכליתי ומדויק. אנו נראה בהרצאה זו, כיצד ניתן לשלב את כל מרכיבי הבקר במערכת משובצת אחת המשלבת: מקורות מתח גבוה מתוכנתים, בקר זרם לחוט להט (filament), בקר לזרם פליטה, בקרי טמפרטורה, בקרים פנאומטיים, תקשורת למערכת GCMS, ותקשורת להתקני סביבה נוספים כמו משאבות טורבו ומדי ספיקה.
אלי פלקסר הוא פרופסור להנדסת אלקטרוניקה ומדעי המחשב באפקה – המכללה האקדמית להנדסה בת"א. הוא שימש כראש המחלקה לאלקטרוניקה עם הקמתה של המכללה. פרופסור פלקסר מייעץ במחקר ופיתוח לחברות בתעשיית ההייטק ולמכוני מחקר, בתחומים של פיסיקה שימושית, אלקטרוניקה ותוכנה (fes.co.il). בשנים הארונות הוא שימש כמדען ראשי במספר חברות. |
פרופסור אלי פלקסר, אפקה |
אסף גליל הוא מהנדס יישומים של חברת TenAsys (www.TenAsys.com) והחברה שבבעלותו מיצגת אותה בישראל.
לאסף למעלה מ 30 שנות נסיון במערכות משובצות מחשב עם דגש על מעבדים ותוכנות למשפחותX86
|
אסף גליל, Tenasys |
|
For many years, performance was the main optimization goal for many systems, including embedded systems. Recently, Power, Energy and Thermal related issues are becoming so important that need to be taken into consideration as first class citizens when designing and implementing systems.
The presentation presents a foundation that can be used to address non-trivial correlations between all these factors, and to take them into consideration when designing complex (hard) real time systems.
|
גיל קידר, הטכניון |
|
ירון שכטר, מנהל קבוצת יישומים ושיווק, Arrow OCS |
|
Shay Gal-On is EEMBC's Director of Software Engineering and leader of the EEMBC Technology Center. At EEMBC Shay created multiple standards the latest being EnergyBench and MultiBench. Prior to EEMBC Shay was Principal Performance Analyst in the Microprocessor Products Group at PMC Sierra
where he influenced the design of new processors, including instruction set design, and optimized both hardware and software products. Other roles in the microprocessor industry over the past decade include Improv Systems where Shay created unique hardware-software codesign tools, and Intel where Shay created production libraries for analysis and optimization of binaries used in simulators, compilers, and binary translators.
Summary:
The performance of a multicore processor implementation is
dependent on numerous factors, including memory and I/O
bandwidth, inter-core communications and cache coherency, OS
scheduling support, and the efficiency of synchronization.
A good benchmark must be able to stress each of these factors, help to isolate the bottlenecks, and provide results which help users
understand platform strengths and weaknesses.
Along with the inherent system-level characteristics, a good
benchmark suite must also address the different methodologies
used in software engineering to take advantage of concurrency.
This talk presents a suite of embedded benchmarks that allows processor and system designers to analyze, test, and improve multicore architectures and platforms as well as characterization
of the first standardized subset for this suite. This benchmark suite
uses standardized workloads, which can be individually
parameterized to vary the parallelism utilized by the algorithm,
and a test harness that provides compatibility with a wide variety
of multicore embedded processors and operating systems. By
applying incrementally challenging workloads, this benchmark
suite allows the testing of scalability and bottlenecks within the
system. Beyond helping designers to optimize programs for
specific processors and systems, the benchmarks also help users
assess the impact of memory bottlenecks, OS scheduling support,
efficiency of synchronization, and other related functions in
platforms using multicore processors. |
Shay Gal-On , EEMBC Director of Software Engineering |
|
ההרצאה תיתן סקירה של תהליך האתחול של לינוקס על מערכת משובצת מחשב (bootrom, boot loader, kernel ותוכנת מערכת), כיצד ניתן למדוד את העיכוב של כל שלב לתהליך העלייה ותציג מגוון טיפים וטריקים לקיצור זמן זמן זה. |
לירן בן חיים מחברת בינה וראש תחום embedded/rt במכללת הי-טק
|
|
Most embedded Linux systems not only contain a kernel, but also contain applications - running in user space. Apart from the application-specific binaries, most embedded systems also have a shell, a system logger and other generic apps. Over the years, many mechanisms have been designed to cross-compile and build these apps, including plain scripts, rpm, dpkg, buildroot, scratchbox, bitbake and others. This talk will address the community efforts, what the various commercial embedded Linux vendors have done and will compare different approaches, illustrate their advantages and drawbacks and try to show how to meet your deadlines in the least painful way. |
רון דוניו,
מנהל טכני, פרטק אמבדד
|