high performance computing on laptops with multicores & gpus

18
High Performance Computing On Laptops With Multicores & GPUs Sushil K. Prasad Computer Science [email protected]

Upload: channing-vega

Post on 31-Dec-2015

23 views

Category:

Documents


0 download

DESCRIPTION

High Performance Computing On Laptops With Multicores & GPUs. Sushil K. Prasad Computer Science [email protected]. About me. Research Area: Parallel and Distributed Algorithms and Systems - over multicores , GPUs, clusters, sensors, handhelds, web services, … - PowerPoint PPT Presentation

TRANSCRIPT

Page 1: High Performance Computing On Laptops With  Multicores  & GPUs

High Performance Computing On Laptops With Multicores & GPUs

Sushil K. PrasadComputer [email protected]

Page 2: High Performance Computing On Laptops With  Multicores  & GPUs

About meResearch Area: Parallel and Distributed

Algorithms and Systems- over multicores, GPUs, clusters, sensors, handhelds, web services, …

Lab: Distributed and Mobile Systems (DiMoS)

at Ga. Tech campus, 5 PhD students, 2 M.S. students

IEEE TCPP Chair (elected) 2 NSF grants – currently looking for

PhD/MS/undergraduate students Distributed Algorithms High Performance Cloud Computing

Page 3: High Performance Computing On Laptops With  Multicores  & GPUs

Multicore & GPU Chips Inside a Laptop - 100s of processors

Page 4: High Performance Computing On Laptops With  Multicores  & GPUs

GPUs Vs Multicores

• Combined power exceeds 180 GFLOPs

Page 5: High Performance Computing On Laptops With  Multicores  & GPUs

Difficult to parallelize

Memory hierarchy is a barrier:

1 cycle core 3 cycles L1

cache 14 cycles L2 250 cycles

RAM

Intel Core-2 Duo Multicore

Page 6: High Performance Computing On Laptops With  Multicores  & GPUs

GPU: Graphics Processing Unit

Nvidia 280 GTX• 240 cores• Extreme memory hierarchy• Registers• Local memory• Shared memory/8

cores• Off chip Global

Memory• bottleneck bus to CPU

Page 7: High Performance Computing On Laptops With  Multicores  & GPUs

• Nvidia 8800 GTX

• Smith Waterman Seq Alignment, Fasta, and Blast

• Database: SwissProt

• Manavski and Valle 2008

Page 8: High Performance Computing On Laptops With  Multicores  & GPUs

Parallel Data Structures -Priority Queues

5 3 1 2

6 8 7 6 5 8 9 7

19 21

12 1423 34 25 3816 13 14 6510 12 15 9

• Large Scale Event Simulation• Immune System Simulation • VLSI Logic simulation

• Branch and Bound• Task Scheduling• Challenge: Fine Grained

Systems • Students: Dinesh Agarwal, Nick Mancuso

Page 9: High Performance Computing On Laptops With  Multicores  & GPUs

Parallel Priority Queues on Multicore

Effect of Prefetching on PPQs

00.020.040.060.080.1

0.120.140.160.180.2

1 2 3 4 5 6 7 8

# of processors

Tim

e ta

ken

(se

con

ds)

With PF Without PF

Page 10: High Performance Computing On Laptops With  Multicores  & GPUs

Legacy-Code to GPUs(Student: Chad Christopher)

Page 11: High Performance Computing On Laptops With  Multicores  & GPUs

Distributed Algorithms for Lifetime of Wireless Sensor Networks (Student: Akshaye Dhawan)

Page 12: High Performance Computing On Laptops With  Multicores  & GPUs

NP-Hard Distributed Problems in Networks

NSF Grant Minimum Vertex/Target Cover Minimum Triangle Packing Optimum mobile sensor network target

tracking Minimum channel assignment in mobile

ad-hoc networks Students: John Daigle, Thamer Sulaiman

Page 13: High Performance Computing On Laptops With  Multicores  & GPUs

April 19, 2023UM-Morris

13

Middleware for Mobile Ad–hoc Applications

Process Requests

Listener

1. Register

2. Lookup

3. p2p communication

Directory

Process Requests

Listener

Process Requests

Listener

Process Requests

Listener

Applications

Applications

Applications

Groupware

Deviceware

Deviceware

Deviceware

Mobile Support Station

Page 14: High Performance Computing On Laptops With  Multicores  & GPUs

BondFlow: Distributed Workflow over Web Services

(Student: Janaka Balasooriya)

Web service interface module

Proxy object generator module

Workflow configuration module

Execution module.

Mobile Web Services

Web Service Interface Module

WSDL Parser

WS Locator

Proxy Object Generator

Module

Workflow Configuration

Module

Web Services Registry (UDDI)

SOAP

Lookup for Web services

WSDL

Parsed WSDL

Workflow Execution Module

SOAP/ SyD

Web Bond Runtime

JVM

Page 15: High Performance Computing On Laptops With  Multicores  & GPUs

Current Belief

P2P Search based on Bayesian Decision and Value of Information (VOI) – (Student: Rasanjalee)

Peer Selection: Sending/forwarding query at

each node along query path = series of decision making steps based on incomplete data

A decision step: query the node that will reduce the uncertainty of current belief most.

Experimental Results:

A Priori Uncertainty : U1

A Posterior Uncertainty : U2

U1 –U2 = Information

The meaning of Uncertainty based Information

Decision step 1..

Decision step n

The reduction in uncertainty at each decision step

1 2 3 4 5

0102030405060708090

100

Success Ratio

SR-APS

SR-UCBPS

# walkers

Su

cces

s ra

tio

%

.

.

Page 16: High Performance Computing On Laptops With  Multicores  & GPUs

Middleware on Distributed Smart Cameras

cmucam3

Middleware on DSC networks provide a high-level programming

interface for applications. simplify the development of distributed

applications on DSC networks. provide networking functionality as part of

the middleware Student: Jayampathi Sampat

Page 17: High Performance Computing On Laptops With  Multicores  & GPUs

About meResearch Area: Parallel and Distributed

Algorithms and Systems- over multicores, GPUs, clusters, sensors, handhelds, web services, …

Lab: Distributed and Mobile Systems (DiMoS)

at Ga. Tech campus, 5 PhD students, 2 M.S. students

IEEE TCPP Chair (elected) 2 NSF grants – currently looking for

PhD/MS/undergraduate students Distributed Algorithms High Performance Cloud Computing

Page 18: High Performance Computing On Laptops With  Multicores  & GPUs

High Performance Computing On Laptops With Multicores & GPUs

Sushil K. PrasadComputer [email protected]