IPSI Overview 2011: Research + Teaching

Post on 19-Jan-2016

38 views 0 download

Tags:

description

IPSI Overview 2011: Research + Teaching. An Overview of IPSI Belgrade Projects for High-Tech Computer Industry in the USA and EU. IPSI Belgrade. IPSI Belgrade - Jointly founded by German/USA/Serbian capital - PowerPoint PPT Presentation

Transcript of IPSI Overview 2011: Research + Teaching

1/77

IPSI Overview 2011:IPSI Overview 2011: Research + Teaching Research + Teaching

An Overview of IPSI Belgrade Projects

for High-Tech Computer Industry

in the USA and EU

2/77

IPSI BelgradeIPSI Belgrade• IPSI BelgradeIPSI Belgrade - Jointly founded by German/USA/Serbian capital

• Some of the the major past and present partners: - University of Ulm, Germany - Fraunhofer IPSI, Darmstadt, Germany - University of Birmingham, UK - Maxeler, FinSoft, GTech, IFACT, UK - NYU, School of Continuous Professional Studies, USA - HP, MS, DowJones, WallStreetJournal, USA

3/77

Employees and AssociatesEmployees and Associates

• CEOCEOProf. Dr. Veljko Milutinovic, Fellow of the IEEE

• Senior ConsultantsSenior ConsultantsProf. Dr. Erich Neuhold, Fellow of the IEEE Prof. Dr. Hiro Fujii, Fellow of the IEEE Dr. Hovard Moskowitz, Fellow of SigmaXi

• Chief OfficersChief OfficersZoran Babovic, Goran Rakocevic, Aleksandar Prijic, Nenad Korolija, etc.

4/77

5/77

The IPSI Belgrade CoverageThe IPSI Belgrade Coverage- Workspaces of the Future- Environments for Cooperative Working and Learning- Virtual Information and Knowledge Environments- Mobile Interactive Media- Open Adaptive Information Management Systems- Publication Engineering and Technology- Networks and WWW- Infrastructure for E-Business on the Internet- Hardware Design and Operating Systems- Novel Computer Architectures

6/77

General Project StructureGeneral Project Structure

Industrial Research:

Phase #1: Survey, and Generation of Embryonic Ideas

Phase #2: Analytic Analysis and Comparison (1+K)

Phase #3: Simulation Analysis and Comparison

Phase #4: Implementation Analysis and Comparison

7/77

General Project StructureGeneral Project Structure

Industrial Development:

Phase #1: Product Requirements and Testing

Phase #2: Intra Module Coding and Testing

Phase #3: Inter Module Integration and Testing

Phase #4: Exhaustive Verification and Testing

8/77

Some Recent Educ Projects Frankfurt/M, Frankfurt/O, Magdeburg, Berlin, Hagen, Koblenz,

Kaiserslautern, Erlangen, TUM, IPSI FhG, Karlsruhe, Ulm, ...

NYU, Purdue, Dartmouth, Hawaii, …

Modena, Ferrara, Siena, Pisa, Salerno, Napoli, …

Tech De Monterrey, Tech De Durango, UNAM, La Salle, …

St. Mary’s, Dalhousie, …

RIT, Skoevde, Karlskrona, Karlstadt, …

Valencia, Madrid, Oviedo, Ciudad Real, …

9/77

Some RecentSome Recent R&D R&D ProjectsProjects

NCR, Encore/Compaq/HP, SUN, Intel, …

Comshare, Zycad, QSI, Virtual, …

TechnologyConnect, BioPop, eT, MainStreetNetworks, …

DowJones, WallStreetJournal, …

Maxeler, FinSoft, …

Ericsson, STC, …

Ulm, Darmstadt, …

Salerno, Pisa, Siena, L’Aquila, ...

10/77

R&D Methodology

• Introduction• Problem Statement• Criticism of Existing Solutions• Proposed Solution• Conditions and Assumptions• Details (1+k)• Mathematical Analysis• Simulation Analysis• Implementation Analysis• Conclusion

11/77

Internet Servers

12/77

NCR: NextGen PC for E-Business Cache coherence maintenance: Hardware approach

Cache coherence maintenance: Software approach

Accelerator chip for windowing

Accelerator board for dbase applications

Prefetching on the "silence" for disk cacheing

Accelerator chip for text compression

Accelerator chip for JPEG/MPEG

13/77

SMP in Action

MM

PP

14/77

ENCORE/COMPAQ/HP

Improved RMS for PC, and its prototype

The RM/MC for PC approach, and its analysis

Simulation of selected DSM approaches, and their comparison (RMS, KSR, and SCI)

Search for the optimal RMS inteconnect technology

15/77

DSM in Action

In t e r co n n e c tio nC o n t ro l le r

In te rc o n n e c t io nC o n tro lle r

In te rc o n n e c tio nC o n tro ll e r

D ire c to ry D i rec to ry D ire c to ryP ro c e s s o rs P ro c e s s o rs P ro c es s o rs

C a c h e s C ac h e s C a c h e sD S M

C luster 1 C luster 2 C luste r N

IC N

D SMSharedA ddressSpace

D S MD S M

16/77

Prologue

17/77

Epilogue

18/77

Cutting the Edge

Top Down Technologies i860

Selected current microprocessor models

QSI An ATM router chip with intelligence

In-memory processing with intelligence

19/77

N.B.

ERRORS

MADE

&

LESSONS

LEARNED

20/77

1

21/65

2

22/77

3

23/77

The Split Temporal/Spatial Cache

Veljko Milutinović, Boris Marković*, Milo Tomašević, Aleksandar Milenković, and MarkTremblay**

IFACT

Department of Computer Engineering School of Electrical Engineering University of Belgrade POB 35-54 11120 Beograde, Serbia

___________________________________________________________________________* Boris Marković is with the University of Montenegro, Podgorica, Montenegro** Mark Tremblay is with the SUN Microsystems, Palo Alto, California, USA

24/77

C1.spat

COMPILE.time

C2.temp

C1.temp

RUN.time

PFB

SPLIT TEMPORAL/SPATIAL CACHE

MM

25/77

The Injection Cache

Veljko Milutinović, Aleksandar Milenković, Davor Magdić,and Gad Sheaffer*

IFACTDepartment of Computer Engineering School of Electrical Engineering University of Belgrade POB 35-54 11120 Beograde, Serbia

________________________________________________________________________ * Gad Sheaffer is with the Intel Corporation, Beverton, Oregon, USA

26/77

IN

PRODUCER

CONSUMER

C2

C1

P

EARLY LATE

tc

t

CACHE INJECTION

27/77

Integrated Systems

28/77

VLSI Detection for

Internet/Telephony Interfaces

Goran Davidović, Miljan Vuletić, Veljko Milutinović,

Tom Chen, and Tom Brunett

* eT

29/77

INTERNET

SERVICE

PROVIDER

REMOTESITE

Superposition/DETECTION Superposition/DETECTION

. . .

USERS...

HOME/OFFICE/FACTORY AUTOMATION ON THE INTERNET

SPECIALIZED

30/77

Reconfigurable FPGA for EBI

Božidar Radunović, Predrag Knežević, Veljko Milutinović,

Steve Casselman, and John Schewel*

* Virtual

31/77

INTERNET

PROVIDER

. . .

USERS

VCC VCC

CUSTOMER SATISFACTION vs CUSTOMER PROFILE

SPECIALIZED

SERVICE

32/77

Browser Acceleration

Gvozden Marinković, Dragan Jandrić, Vladimir Ivanović,

Veljko Milutinović, and Tom Chen

*MainStreetNetworks

33/77

What is the Major Bottleneck?Rendering!Send the "IN FO R M A T IO N A B O U T E T F B E L G R A D E "

H T T P

T h e c lie n t se n d s H T T P m e ssa g e to a c o m p u te rru n n in g a W e b S e rv e r p ro g ra m a n d a sk s fo r a d o c u m e n t

T he Inform ationabout

E T F B E L G R A D E

T h e W eb server sen ds th e h yperm edia H T M L docum en ts to th e clien t.Y ou en d up seein g th e docum en t on your screen

C lie n t S e rv e r

34/77

BioPoP

Veljko Milutinovic, Vladimir Jovicic, Milan Simic,

Bratislav Milic, Milan Savic, Veljko Jovanovic,

Stevo Ilic, Djordje Veljkovic, Stojan Omorac,

Nebojsa Uskokovic, and Fred Darnell

•isItWorking.com

35/77

Testing the Infrastructure for EBI

• Phones

• Faxes

• Email

• Web links

• Servers

• Routers

• Software

• Statistics

• Correlation

• Innovation

36/77

CNUCEIntegration and Datamining

on Ad-Hoc Networks and the Internet

Veljko Milutinović, Luca Simoncini,

and Enrico Gregory

*University of Pisa, Santanna, CNUCE

37/77

GSM

DMAd-Hoc

InternetAd-Hoc

38/77

Intelligent Search

39/77

Genetic Search with Spatial/Temporal Mutations

Jelena Mirković, Dragana Cvetković,and Veljko Milutinović

*Comshare

40/77

Drawbacks of INDEX-BASED: Time to index + ranking

Advantages of LINKS-BASED: Mission critical applications + customer tuned ranking

Provider

Well organized markets: Best first searchIf elements of disorder: G w DB mutationsChaotic markets: G w S/T mutations

41/77

University of UlmReverse Engineering of

GeForce2-4

Sasa Jandric, Zaharije Radivojevic, Milos Cvetanovic, and Veljko Milutinovic

42/65

010101000001101101

• Developing system control programs (drivers) for GeForce 2-4, for the Plurix operating system

• Main advantage of the GeForce chip (called graphical processor) is the use of 3D accelerating functions

• Reversed engineering is used as a technology for finding previous information on the GeForce chip

43/77

SummaryThe world’s best journals - IEEE: Balkan record in ICT (50)

Books with Nobel Laureates (7):

Kenneth Wilson, Ohio (North-Holland) Leon Cooper, Brown (Prentice-Hall) Robert Richardson, Cornell (Kluwer-Academics) Jerome Friedman, MIT (IOS Press)

Herb Simon (Kluwer-Academics) Harold Kroto (Academic Mind Press) Arno Penzias (Academic Mind Press)

44/77

Recent and On-Going Projects

• StorageTek, Colorado, USA• Panthesis (exBoeing), Oregon, USA• Wall Street Journal, New York, USA• HP, Palo Alto, USA• STC, Uppsala, Sweden• Fraunhofer, Darmstadt, Germany• Finsoft, London, England• Maxeler, London, England

45/77

DSFS - Digital Sealed File System

• Goal: System for Data Modification Detection

• StampTool = Smart Card– Input: Hash – Output: Stamp + Public Key

• Verification Body– Calculation of Hash Value– Comparison with Encrypted Stamp

• PublishingTool – Public Key Distribution– Several Possible Solutions

Depending on a Purpose

StorageTek/A

StampTool

Server

Signer PublishingTool

Verification Body

46/77

Panthesis/A

• E-Learning with SWAN

FinSoft/GTech

• The art of testing in a fragile environment

• Application tuned testing methodologies

47/73

Maxeler

• Compilation into FPGA

• Performance over the top supercomputers up to 240 times.

48/73

FP7

• WeGo

• Balcon

• ARTreat

• ProSense

49/77

50/77

http://home.etf.rs/~vmhttp://www.ipsibelgrade.net/

e-mail: vm@etf.rsoffice@ipsibelgrade.net

51/77

The Pro Life Management Course

Prof. Dr. Veljko Milutinović, Fellow of the IEEE

How to Maximize the $1M Probability

52/40

53/77

The Pro Life Management Course is Divided in 10 Parts

Till 40:• For those students who like to do their PhD in the USA or Canada• For those students who like to continue professional work in EU or WB

Till 60:• For those students who plan to do MBA,

so they can become managers in companies of others, or their own

Till 80:• For those students who plan to do PhD,

so they can teach at public or private universities

Till 100:• For those students who have to make business plans for investors• For those students who have to make investment-related decisions

Also:• For all students, no matter which way they like to continue: efficiency• For all students, no matter which way they like to continue: effectiveness• For all students, no matter which way they like to continue: inventiveness • For all students, no matter which way they like to continue: creativity

54/77

Pro Life Management 1st

The first part is for those students who like to do their PhD in the USA or Canada(the stress is on how to maximize the score on GRE or GMAT).This is a good option after your 20th birthday,i.e., after you complete your undergraduate work.

55/77

Form of the Exam

• Verbal reasoning (input),important for teaching schools

• Quantitative reasoning (CPU);tough to compete with Chinese and Indians

• Analytical writing (output),important for research schools

56/77

Pro Life Management 2nd

The second part is for those students who like to continue professional work in EU or WB(the stress is on how to apply for EU FP projects).This is a good option after your 20th birthday,i.e., after you complete your undergraduate work.

57/77

PP (proposal) ST

OBJ (objectives) OBJ1-OBJY [SC (success criteria)1-N]

WP (work packages) WP0-WPY [MS (milestones)1-N]

Tasks

Deliverables

15=ST(5)+MF(5)+ED(5)13.510

CH (challenges)

1:1

1:M

1:K

1:1

FP/FP7/CALLn/ClosingDate/E1M->E4M(STREP)/E40M(IP)

1:1

If x>=10 and each{ST,MF,ED}>=3 Then: E5K, if leader

E3K, if WP leaderE1K, if partner

58/77

Example ST

Objectives:

1. Development of Use Cases [e.g., TMG, SLO]2. Development of System Architecture [e.g., ETF, SRB]3. HW Details [e.g., PADOVA, ITA]4. SW Details [e.g., BSC, ESP]5. AI Details [e.g., INRIA, FRA]6. PoC Implementation [e.g., XYZ, AUS]7. Testing #1 [e.g., IGALO, MNE]8. Testing #2 [e.g., KOPER, SLO]9. Dissemination: Conferences + Media + Investors [MediaPro, ESP]10. QoQ [e.g., Philips, NED]

WP: Almost each partner in almost each WP!

59/77

Pro Life Management 3rd

The third part is for those students who plan to do MBA,

so they can become managers in companies of others or their own

(the stress is on MBA in the nutshell).

This is a good option after the 40th birthday,

unless you like the young professionals to run over you.

60/77

MBA – What is it?

• Experiences from Sloan, MIT, USA• Work and management experience required

(at least 4/2 years)• MBA fee + logistics = FYS• Significant investment into scholarship,

but great return on investment

61/77

Pro Life Management 4th

The forth part is for those students who plan to do PhD,

so they can teach at public or private universities

(the stress is on publishing in SCI journals).

This might be a good option after the 60th birthday,

unless you start with academia right away.

62/77

Important Issues:1) Shanghai Top 500 (6),

or how many SCI papers for PhD?

2) Conditio sine qua non: Two stories

3) Best Method, .doc (over 1M)

4) Good Method, .ppt (over 1B)

5) Three types of SCI papers:Survey, Comparison, and Research

6) Response to reviewers must have a structure

63/77

Pro Life Management 5th

The fifth part is for those students

who have to make business plans for investors

(the main question is

how to present the basic idea in a clear way).

This is a good option after the 80th birthday.

64/77

Harvard Business Plan Template

65/77

Pro Life Management 6th

The sixth part is for those students

who have to make investment-related decisions

(the main question is how to know

which one of the options is the best one for investing into it).

This is a good option after the 80th birthday.

66/77

Clemson BizPlan Evaluation Template

67/77

Pro Life Management 7th

The seventh part is for all students,

no matter which way they like to continue

(the stress is on how to become effective:

to organize time).

This part is based on the book by Drucker.

68/77

The Effective Executive

69/77

Pro Life Management 8th

The eight part is for all students,

no matter which way they like to continue

(the stress is on how to become efficient:

to sell ideas, products, services, etc.).

This part is based on the famous book by Lloyd.

70/77

Lloyd,

SELLING, Dorling Kindersley, UK

Questions?

71/77

Pro Life Management 9th

The ninth part is for all students,

no matter which way they like to continue

(the stress is on inventiveness).

This part is based on a VIPSI conference talk,

by Nobel Laureate Jerome Friedman of MIT.

72/77

Qs (NL)• 1. What is the essence of the contribution for which you received the Nobel Prize? • 2. What are the impacts of this contribution? • 3. What are the applications of your contribution that may change the everyday life? • 4. We learned a lot from your lectures in Belgrade. Can you tell us, what are the issues that we have

to teach our kids, so they become inventive when they finish studies? • 5. What are the major things to keep in mind, when you form a team for a scientific experiment, or

similar • 6. What are the people to avoid, when trying to generate a break-through achievement? • 7. When targeting a major breakthrough, how sensitive one has to be about the direct interests of

tax-payers? • 8. What is the major driving force that motivates a Nobel Laureate to continue to create and generate

results after he-she receives the Nobel Prize? • 9. For small nations like Serbian, what is your advice, which road to take, when it comes to science? • 10. What road to take, when it comes to its general future development plans?

73/77

Pro Life Management 10th

The tenth part is for all students,

no matter which way they like to continue

(the stress is on creativity).

This part is based on a VIPSI conference talk,

by Nobel Laureate Martin Perl of Stanford.

74/77

• Creativity in Science and Engineering

• How to Get a Good Idea

• Colleagues

• The Art of Obsession

• The Technology You Use

• Future Technology

Topics

75/77

Literature

• GRE (2 weeks) • FP7 (2 weeks) • MBA in the Nutshell (1 week) • PhD in the Nutshell (1 week) • Harvard Business Plan Template (1 week) • Harvard Decision Making Methodology (1 week) • Drucker (2 weeks) • LLOYD (2 weeks) • Nobel Laureates About Inventiveness:

Friedman, Arrow, Glashow, Maskin (1 week) • Nobel Laureate Perl About Creativity (1 week)

76/40

77/77

Q?