1 / 10

CINT Retrospect

CINT Retrospect. ROOT Workshop Mar 2013 Agilent Technologies Masaharu Goto. Software for Scientific Research. Fundamental Needs -Users are not necessarily software experts Easy - Flexibility to deal with unpredictable phenomenon Interactive

tanuja
Download Presentation

CINT Retrospect

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. CINT Retrospect ROOT Workshop Mar 2013 Agilent Technologies Masaharu Goto

  2. Software for Scientific Research Fundamental Needs -Users are not necessarily software experts Easy -Flexibility to deal with unpredictable phenomenon Interactive -High performance computation for simulation Fast -Stable over long period of time Sustainable - Work with foreign libraryFlexible Interpreter for Standard Computer Language Software Computer Systems Measurement Unit Signal Source Device Under Test

  3. Before CINT (author’s experience) • In 1980s, • New computer language every year as computer evolves • Had to re-write library each time • Needed stable computer platform!!! Meanwhile, software guys were using C C/C++ language is - Industry standard good - Not easy bad - Compiler was very slow bad - No interpreter bad 1980 1990 Fortran NEC Basic Pascal HP Basic C++ C Decided to solve those issues by myself to obtain stable computation environment for >10 years

  4. CINT Original Concept Sustain scientific computation environment for many years • Support ANSI/ISO C/C++ Sustainable • As Implementation: • Multi Platform, Compiler independent • Exclude non-standard components such as graphics • As Scripting Language: • May not be fully compliant, but should be good enough for scripting • Easy to Use Quick & Easy • Fast Turn Around Time, Interactive • Language short-cut (Complexity of C++ remains as an issue, though) • Seamless interpreter-compiler integration Fast • Seamless access between interpreter and compiler code • Portable Component Flexible • Can be integrated with any foreign library or framework

  5. CINT Development • At beginning, it wasn’t a serious project 1990 • The author had absolutely no C/C++ knowledge • Intended to be local tool • It worked somehow… • Adding C++ features 1992 • Object Oriented Programming (Class, Inheritance, Virtual Function, etc…) • Complicated, but not so bad • Adding STL 1994 • Generic Programming - Great concept but extremely difficult to implement • Template itself is complicated enough, STL is a nightmare on top of it • ROOT collaboration 1995 • Add special feature • HEP became the largest user community

  6. Challenges • Complexity of C/C++ • Template and STL • Inheritance, Function overloading, Operator precedence, etc… • C/C++ as Interpreter language • Strongly Typed Language definition conflicts Ease of Use • Interpreter-Compiler Seamlessness • Interpreted class inheriting compiled class • (protected member, virtual function, Pointer to member function) • Dictionary code was an inevitable solution to realize this concept • Ad-hoc approach turned out to be right for CINT • Tried to think thoroughly but overwhelmed by challenges

  7. Alternatives were … • As time goes by…. • Before: Many MPUs, OSs, compilers • Very slow Compiler/Linker • Now: Only 2-3 MPU&OS choices - Intel / Win, Linux, Mac / VC++, gcc • Quick Compiler/Linker • Platform dependent features may not hurt so much… • Depending on Compiler/Processor specific undocumented scheme will simplify or eliminate dictionary code • Data stacking, Pointer to member, Name mangling, … • Just implement as another compiler • 1. Byte-code compiler + virtual machine • 2. Use compiler/linker in background (ACliCapproach) • Some features may be lost, but probably still be usable

  8. The Biggest Road-Block • Spam mail • At Iraq War (2003), Cyber attack began • Number of Spam mail increased >x10 • 100s of Spam mails per day • Slow internet connection • Spam filter were not ready • Took 2-3 hours just to download and filter e-mails • Stopped being an active developer but remain as a user • CINT already had enough capability for my original objective • -> “ROOT/CINT used in Test and Measurement Industry”

  9. Overview and Conclusion • CINT provided Easy-Interactive-Fast computation environment • Lasted more than 20 years helping our productivity 1980 1990 2000 2020 2010 Fortran NEC Basic Pascal HP Basic C++ C ROOT/CINT CINT Development Effort Thanks to CERN & Fermi-Lab collaboration, CINT was very successful

  10. Acknowledgement • Special thanks to • Rene Brun • Fons Rademakers • Philippe Canal • Axel Naumann • all ROOT/CINT developers and users

More Related