MICRO Graphicionado. A High-Performance and Energy-Efficient Graph Analytics Accelerator

Size: px
Start display at page:

Download "MICRO Graphicionado. A High-Performance and Energy-Efficient Graph Analytics Accelerator"

Transcription

1 MICRO 2016 Graphicinad A High-Perfrmance and Energy-Efficient Graph Analytics Acceleratr Tae Jun Ham Lisa Wu Narayanan Sundaram Nadathur Satish Margaret Martnsi Slide:

2 Graph Analytics A key wrklad f mdern-day cmputing Applicatin Dmains Scial Netwrks, Biinfrmatics, Healthcare, Cybersecurity, Simulatin, etc. Algrithms Path Analysis (e.g., BFS, SSSP) Centrality Analysis (e.g., PageRank, Betweenness Centrality) Recmmendatin systems (e.g., Cllabrative Filtering) Cmmunity Analysis

3 Graph Analytics is difficult t cde Graph Analytics has Irregular data access patterns Extremely lw cmputatin-t-cmmunicatin rati Pr spatial/tempral data lcality Difficult-t-extract parallelism Memry-bund perfrmance Slutin: Sftware Graph Prcessing Framewrks Prgrammers express graph algrithms using a prgramming abstractin (i.e., vertex prgram) Framewrks rchestrate data mvements fr given cmputatin using efficient backend SW Pregel GraphMat PGX Graph Engine Industry ( ) Extended Giraph nvgraph System G GraphLab Create Pegasus (ICDM 09) Galis (SOSP 13) Ligra (PPPP 13) Academia X-Stream (SOSP 13) TurbGraph (KDD 13) GraphX (OSDI 14) Gunrck (PPPP 16)...

4 Sftware Framewrks have limitatins = Insufficient tailring t HW Applicatin-blivius memry system in cnventinal prcessrs Caches blindly cache all data at a carse granularity including... Lw tempral lcality data Lw spatial lcality data (i.e., nly 4B ut f 64B is utilized) Our measurement shws that SW framewrks cnsume 2x-27x mre memry bandwidth than the ptimal cmmunicatin case Expensive data mvement in cnventinal prcessrs Graph analytics has extremely lw cmputatin-t-cmmunicatin rati Mst algrithms: < 6% f instructins are used fr actual cmputatin; 94% used fr data mvements; results in huge energy cnsumptin.

5 Graphicinad Apprach Graphicinad: A high-perfrmance, energy-efficient graph analytics HW acceleratr which vercmes the limitatins f sftware framewrks while retaining the prgrammability benefit f SW framewrks Retains Prgrammability Specialized-while-flexible HW pipeline Overcmes Limitatin Applicatin-specific memry system Applicatin-specific pipeline design

6 Presentatin Outline Why Graphicinad? Vertex Prgramming Abstractin Cnstructing Graphicinad Pipeline frm Abstractin Optimizing Graphicinad Optimizing Data Mvement Scaling On-Chip Memry Usage Parallelizatin Perfrmance/Energy Evaluatin Cnclusins

7 Graph Data Structure Graph cnsists f vertices & edges Each Vertex has an ID and and a prperty (r state) Each Edge is defined as a tuple (SRC ID, DST ID, edge prperty) Graph analytics: an iterative cmputatin t calculate the desired vertex prperty fr each vertex

8 Vertex Prgramming Abstractin 1 Fr each active Vertex U 2 Fr each utging edge E(U,V) Prcessing Phase 3 4 Res = Prcess_Edge (E weight, U prp ) V temp = Reduce (V temp, Res) 5 End 6 End Apply Phase Fr each Vertex V V prp = Apply(V temp, V prp ) End Prcess_Edge Reduce Apply Vertex prgramming abstractin is cmmnly used in SW framewrks (e.g., Intel GraphMat, Ggle Pregel, etc.) Prgrammers express graph analytics algrithm with three custm functins Graphicinad uses the same abstractin t retain the prgrammability f SW framewrks

9 Vertex Prgramming Abstractin Prcessing Phase Apply Phase Fr each active Vertex U End Fr each utging edge E(U,V) End Res = E weight + U prp V temp = min(v temp, Res) Fr each Vertex V End V prp = min(v temp, V prp ) Prcess_Edge Reduce Apply V prp : minimum distance t V n last iteratin V temp : minimum distance t V n this iteratin Single Surce Shrtest Path (SSSP) Example Prcess_Edge: E weight + U prp Reduce: min(v temp, Res) Apply: min(v temp, V prp )

10 Vertex Prgramming Abstractin Prcessing Phase Apply Phase Fr each active Vertex U End Fr each utging edge E(U,V) End Res = E weight + U prp V temp = min(v temp, Res) Fr each Vertex V End V prp = min(v temp, V prp ) Calculate a distance t V thrugh this edge Update the current minimum distance t V Cmpare the current minimum distance with the value frm the last iteratin; If it changes its value, add V t the active vertex set fr the next iteratin V prp : minimum distance t V n last iteratin V temp : minimum distance t V n this iteratin Single Surce Shrtest Path (SSSP) Example Prcess_Edge: E weight + U prp Reduce: min(v temp, Res) Apply: min(v temp, V prp )

11 Frm abstractin t HW fr (i=0; i<activevertex.size(); i++) { vid = ActiveVertexID[i]; vprp = ActiveVertexPrp[i]; eptr = PtrTEdgeList[vid]; fr (e = Edges[eptr]; e.src == vid; e = Edges[++eptr]) { res = Prcess_Edge(e.weight, vprp); temp = TempVertexPrp[e.dst]; temp = Reduce(temp, res); TempVertexPrp[e.dst] = temp; } } // Apply Phase updates ActiveVertexPrp with TempVertexPrp Read the active (SRC) vertex Traversing edges f the given active vertex Updating the destinatin vertex with the prgrammer supplied cmputatin Prcessing Phase Blck Diagram Line 2-3 Line 4 Line 5 Line 6 Read Active SRC Prperty Read Edge Pinter Read Edges fr given SRC Prcess Edge Read Temp. Reduce Write Temp. Line 7 Line 8 Line 9

12 Frm abstractin t HW fr (i=0; i<activevertex.size(); i++) { vid = ActiveVertexID[i]; vprp = ActiveVertexPrp[i]; res = Prcess_Edge(e.weight, vprp); temp = TempVertexPrp[e.dst]; temp = Reduce(temp, res); TempVertexPrp[e.dst] = temp; } } // Apply Phase updates ActiveVertexPrp with TempVertexPrp Custm Cmputatin Prcessing Phase Blck Diagram Line 2-3 Line 4 Line 5 Line 6 Sequential (vertex) Memry Access eptr = PtrTEdgeList[vid]; Nn-sequential (edge ptr) Memry Access fr (e = Edges[eptr]; e.src == vid; e = Edges[++eptr]) { Nn-sequential and then Sequential (edge) Memry Access Nn-sequential (vertex) Memry Access Read Active SRC Prperty Read Edge Pinter Read Edges fr given SRC Prcess Edge Sequential Memry Access Unit Nn-Sequential Memry Access Unit Read Temp. Reduce Write Temp. Custm Cmputatin Unit Line 7 Line 8 Line 9

13 Frm abstractin t HW fr (i=0; i<activevertex.size(); i++) { vid = ActiveVertexID[i]; vprp = ActiveVertexPrp[i]; eptr = PtrTEdgeList[vid]; fr (e = Edges[eptr]; e.src == vid; e = Edges[++eptr]) { res = Prcess_Edge(e.weight, vprp); temp = TempVertexPrp[e.dst]; temp = Reduce(temp, res); TempVertexPrp[e.dst] = temp; Read-Mdify-Write Update needs t be atmic fr the same destinatin vertex (e.dst) } } // Apply Phase updates ActiveVertexPrp with TempVertexPrp Prcessing Phase Pipeline Line 2-3 Line 4 Line 5 Line 6 Read Active SRC Prperty Read Edge Pinter Read Edges fr given SRC Prcess Edge Sequential Memry Access Unit Nn-Sequential Memry Access Unit Cntrl Read Temp. Reduce Write Temp. Line 7 Line 8 Line 9 Custm Cmputatin Unit Cntrl Unit

14 Optimizing Data Mvement Prcessing Phase Pipeline S1: Read Active SRC Prperty S2: Read Edge Pinter S3: Read Edges fr given SRC S4: Prcess Edge Sequential Memry Access Unit Nn-Sequential Memry Access Unit S5: Cntrl S6: Read Temp. S7: Reduce S8: Write Temp. Custm Cmputatin Unit Cntrl Unit 1. Nn-sequential access; lw-spatial-lcality PtrtEdgeList[vid] TempVertexPrp[e.dst] Only use 4-16B ut f 64B ff-chip memry access granularity; leads t ff-chip memry BW waste 2. High-tempral lcality TempVertexPrp[e.dst] Data will be used again fr edges with the same destinatin vertex Use fine-grained n-chip SPM fr data used fr these mdules

15 Optimizing Data Mvement Prcessing Phase Pipeline S1: Read Active SRC Prperty S2: Read Edge Pinter S3: Read Edges fr given SRC S4: Prcess Edge On-chip SPM Access Unit Sequential Memry Access Unit Nn-Sequential Memry Access Unit S5: Cntrl S6: Read Temp. S7: Reduce S8: Write Temp. Custm Cmputatin Unit Cntrl Unit 1. Nn-sequential access; lw-spatial-lcality PtrtEdgeList[vid] TempVertexPrp[e.dst] Only use 4-16B ut f 64B ff-chip memry access granularity; leads t ff-chip memry BW waste 2. High-tempral lcality TempVertexPrp[e.dst] Data will be used again fr edges with the same destinatin vertex Use fine-grained n-chip SPM fr data used in these mdules

16 Optimizing Data Mvement Prcessing Phase Pipeline S1: Read Active SRC Prperty S2: Read Edge Pinter S3: Read Edges fr given SRC S4: Prcess Edge On-chip SPM Access Unit Sequential Memry Access Unit Nn-Sequential Memry Access Unit S5: Cntrl S6: Read Temp. S7: Reduce S8: Write Temp. Custm Cmputatin Unit Cntrl Unit 1. Sequentially accessed; high-spatial lcality Has predictable access pattern ActiveVertexPrp[i] Edges[eptr++] All 64B laded frm the ff-chip memry will be used 2. Lw-tempral lcality Each active vertex r an edge is prcessed nly nce. This data will nt be accessed again fr the current iteratin. Use prefetcher; D nt use n-chip SPM fr data used fr these mdules

17 Scaling On-chip Memry Usage Graphicinad utilizes an n-chip strage t avid wasting ff-chip BW Often requires 4-16B n-chip strage per vertex Nt enugh n-chip strage fr 10M+ vertices Slutin: Partitin a graph befre the executin; Then, prcess each subgraph at a time Assign edges t different slices based n their destinatin vertices Original Graph

18 *Paper has additinal scaling techniques Scaling On-chip Memry Usage Graphicinad utilizes an n-chip strage t avid wasting ff-chip BW Often requires 4-16B n-chip strage per vertex Nt enugh n-chip strage fr 10M+ vertices Slutin: Partitin a graph befre the executin; Then, prcess each subgraph at a time Assign edges t different slices based n their destinatin vertices Slice 1 (Edges with DST vertices 1,2,3) Original Graph Slice 2 (Edges with DST vertices 4,5,6) Only half f the vertices t be stred in n-chip strage at a time

19 Parallelizing Graphicinad Pipeline S1: Read SRC Active SRC Vertex Prperty 0 S2: Read Edge ID Table S3: Read Edges fr Edge given (0,3) SRC S4: Prcess Edge S5: Cntrl S6: Read Temp. S7: Reduce S8: Write Temp. S1: Read Active SRC Prperty S1: Read SRC Active SRC Vertex Prperty 2 S2: Read Edge ID Table S2: Read Edge ID Table S3: Read Edges fr given SRC S3: Read Edges Edge (2,3) fr given SRC N NxN x N Switch S4: Prcess Edge S4: Prcess Edge S5: Cntrl S5: Cntrl S6: Read Temp. S6: Read Temp. S7: Reduce S7: Reduce S8: Write Temp. S8: Write Temp. S1: Read Active SRC Prperty S2: Read Edge ID Table S3: Read Edges fr given SRC S4: Prcess Edge S5: Cntrl S6: Read Temp. S7: Reduce S8: Write Temp. Naive Parallelizatin Apprach Replicate Pipeline Tw mdules trying t read the same address (TempVertexPrp[3]) Each pipeline stream prcesses a prtin f the active (SRC) vertices (e.g., a mdul f the SRCID determines which stream fr prcessing) Different streams will try t read/write the same memry address (TempVertexPrp[]) at the same time n S6 & S8 Design cmplexity (SPM prt cunt) and perfrmance degradatin

20 Parallelizing Graphicinad Pipeline SRC Access Prtin DST Access Prtin S1: Read SRC Active SRC Vertex Prperty 0 S2: Read Edge ID Table S3: Read Edges Edge (0,3) fr given SRC S4: Prcess Edge S5: Cntrl S6: Read Temp. S7: Reduce S8: Write Temp. S1: Read Active SRC Prperty S1: Read SRC Active SRC Vertex Prperty 2 S2: Read Edge ID Table S2: Read Edge ID Table S3: Read Edges fr given SRC S3: Read Edges fr Edge given (2,3) SRC N NxN x N Switch S4: Prcess Edge S4: Prcess Edge S5: Cntrl S5: Cntrl S6: Read Temp. S6: Read Temp. S7: Reduce S7: Reduce S8: Write Temp. S8: Write Temp. S1: Read Active SRC Prperty S2: Read Edge ID Table S3: Read Edges fr given SRC S4: Prcess Edge S5: Cntrl S6: Read Temp. S7: Reduce S8: Write Temp. Graphicinad Apprach Each pipeline stream is separated t tw pieces: SRC access / DST access prtin Nw each unit access an exclusive prtin f the scratchpad memry Each unit accesses an exclusive memry regin N x N switch rutes data t the crrect DST stream (e.g., a mdul f the DSTID) Reductin in design cmplexity and imprvement in thrughput

21 Presentatin Outline Why Graphicinad? Vertex Prgramming Abstractin Cnstructing Graphicinad Pipeline frm Abstractin Optimizing Graphicinad Optimizing Data Mvement Scaling On-Chip Memry Usage Parallelizatin Perfrmance/Energy Evaluatin Cnclusins

22 Overall Graphicinad Perfrmance 7 6 Sftware Framewrk Graphicinad Graphicinad Sftware Framewrk Speedup ver Sftware Framewrk FR FB Wiki LJ RMAT TW AVG FR FB Wiki LJ RMAT TW AVG FR FB Wiki LJ RMAT TW AVG NF SB AVG PageRank (PR) Breadth-First Search (BFS) (BFS) Single Surce Single-Surce Shrtest Path Cllabrative (SSSP) (CF) Shrtest Path (SSSP) Filtering (CF) Graphicinad achieves ~3x speedup ver state-f-the-art sftware framewrk (Intel GraphMat) acrss different wrklads Sftware framewrk was run n a 32-cre Haswell Xen server Bth system were prvisined the same theretical peak memry BW (78GB/s)

23 Graphicinad Energy Cnsumptin Nrmalized Energy Cnsumptin edram Static Energy edram Dynamic Energy edram Static Energy edram Dynamic Energy Graphicinad Pipeline Static Energy Graphicinad Pipeline Dynamic Energy Custm Cmputatin Energy Graphicinad Pipeline Static Energy Graphicinad Pipeline Dynamic Energy Custm Cmputatin Energy 0 FR FB Wiki LJ RMAT TW FR FB Wiki LJ RMAT TW FR FB Wiki LJ RMAT TW NF SB PageRank (PR) Breadth-First Search (BFS) Single Surce Single-Surce Shrtest Path (SSSP) Cllabrative Shrtest Path (SSSP) Filtering (CF) Graphicinad cnsumes < 2% f the energy (50x-100x) cmpared t the sftware prcessing framewrk Mst f the energy (70%+) was spent fr the edram static energy 20-25x pwer saving & 2-5x speedup

24 Effect f Parallelizatin & Optimizatin Nrmalized Perfrmance Baseline Baseline CFG 1 = Stream Parallelizatin(x2) CFG 2 = Stream Parallelizatin(x4) CFG 3 = Stream Parallelizatin(x8) CFG1 = Stream Parallelizatin (2x) CFG2 = Stream Parallelizatin (4x) CFG3 = Stream Parallelizatin (8x) CFG4 = CFG3 + Prefetching CFG 4 = CFG 3 + Prefetching CFG 5 = CFG 4 + Pipeline Balancing* PageRank (PR) Breadth-First Search (BFS) Single Surce Shrtest Path (SSSP) Cllabrative Filtering (CF) PageRank (PR) Breadth-First Search (BFS) Single-Surce Shrtest Path (SSSP) CFG5 = CFG4 + Pipeline Balancing* Cllabrative Filtering (CF) The parallelizatin and ptimizatin f Graphicinad achieves up t 27x speedup ver the baseline single-stream pipeline Withut effective parallelizatins and ptimizatins, simply using HW des nt necessarily bring any benefit

25 Cnclusins Sftware graph prcessing framewrks has limitatins Inefficient n-chip memry (cache) usage Expensive data mvement Graphicinad: Specialized HW acceleratr fr graph analytics Efficient use f n-chip memry Specialized pipeline tailred fr graph analytics data mvement Effective parallelism ~3x speedup and 50x+ energy saving ver state-f-the-art sftware framewrk

26 MICRO 2016 Graphicinad A High-Perfrmance and Energy-Efficient Graph Analytics Acceleratr Tae Jun Ham Lisa Wu Narayanan Sundaram Nadathur Satish Margaret Martnsi

NVIDIA Tesla K20X GPU Accelerator. Breton Minnehan, Beau Sattora

NVIDIA Tesla K20X GPU Accelerator. Breton Minnehan, Beau Sattora NVIDIA Tesla K20X GPU Acceleratr Bretn Minnehan, Beau Sattra Overview Jb f the GPU Histry What is the K20X GK110 Benchmark Perfrmance Jb f the GPU Vertex Shader Applies transfrms n each vertex Applies

More information

Exploring the Viability of the Cell Broadband Engine for Bioinformatics Applications

Exploring the Viability of the Cell Broadband Engine for Bioinformatics Applications Explring the Viability f the Cell Bradband Engine fr Biinfrmatics Applicatins Authrs: Vipin Sachdeva, Michael Kistler, Evan Speight and Tzy-Hwa Kathy Tzeng Presentatin by: Keyur Malaviya Dept f Cmputer

More information

It has hardware. It has application software.

It has hardware. It has application software. Q.1 What is System? Explain with an example A system is an arrangement in which all its unit assemble wrk tgether accrding t a set f rules. It can als be defined as a way f wrking, rganizing r ding ne

More information

Using SPLAY Tree s for state-full packet classification

Using SPLAY Tree s for state-full packet classification Curse Prject Using SPLAY Tree s fr state-full packet classificatin 1- What is a Splay Tree? These ntes discuss the splay tree, a frm f self-adjusting search tree in which the amrtized time fr an access,

More information

Operating systems. Module 15 kernel I/O subsystem. Tami Sorgente 1

Operating systems. Module 15 kernel I/O subsystem. Tami Sorgente 1 Operating systems Mdule 15 kernel I/O subsystem Tami Srgente 1 SWAP SPACE MANAGEMENT Swap space can be defined as a temprary strage lcatin that is used when system s memry requirements exceed the size

More information

A Characterization of Data Mining Algorithms on a Modern Processor

A Characterization of Data Mining Algorithms on a Modern Processor A Characterizatin f Data Mining Algrithms n a Mdern Prcessr Aml Ghting, Gregry Buehrer, and Srinivasan Parthasarathy Data Mining Research Labratry, The Ohi State University Daehyun Kim, Anthny Nguyen,

More information

Memory Hierarchy. Goal of a memory hierarchy. Typical numbers. Processor-Memory Performance Gap. Principle of locality. Caches

Memory Hierarchy. Goal of a memory hierarchy. Typical numbers. Processor-Memory Performance Gap. Principle of locality. Caches Memry Hierarchy Gal f a memry hierarchy Memry: hierarchy f cmpnents f varius speeds and capacities Hierarchy driven by cst and perfrmance In early days Primary memry = main memry Secndary memry = disks

More information

Speculative Parallelization. Devarshi Ghoshal

Speculative Parallelization. Devarshi Ghoshal Speculative Parallelizatin Devarshi Ghshal Indiana University, Blmingtn 10/10/2011 1 Agenda Speculative Parallelizatin FastFrward-A Speculatin using Checkpint/Restart System Design Sftware-based Speculatin

More information

SW-G using new DryadLINQ(Argentia)

SW-G using new DryadLINQ(Argentia) SW-G using new DryadLINQ(Argentia) DRYADLINQ: Dryad is a high-perfrmance, general-purpse distributed cmputing engine that is designed t manage executin f large-scale applicatins n varius cluster technlgies,

More information

Computer Organization and Architecture

Computer Organization and Architecture Campus de Gualtar 4710-057 Braga UNIVERSIDADE DO MINHO ESCOLA DE ENGENHARIA Departament de Infrmática Cmputer Organizatin and Architecture 5th Editin, 2000 by William Stallings Table f Cntents I. OVERVIEW.

More information

NVIDIA S KEPLER ARCHITECTURE. Tony Chen 2015

NVIDIA S KEPLER ARCHITECTURE. Tony Chen 2015 NVIDIA S KEPLER ARCHITECTURE Tny Chen 2015 Overview 1. Fermi 2. Kepler a. SMX Architecture b. Memry Hierarchy c. Features 3. Imprvements 4. Cnclusin 5. Brief verlk int Maxwell Fermi ~2010 40 nm TSMC (sme

More information

KIRA-EMTA. Why KIRA-EMTA Multi-Threaded Supercomputer. Inside KIRA-EMTA. Elastic Supercomputing Architecture. Exa-Converged Architecture

KIRA-EMTA. Why KIRA-EMTA Multi-Threaded Supercomputer. Inside KIRA-EMTA. Elastic Supercomputing Architecture. Exa-Converged Architecture Explring New Supercmputer Perfrmance Scenaris Fr New Emerging Applicatins Why KIRA-EMTA Multi-Threaded Supercmputer The A3Cube KIRA-EMTA supercmputer realizes an adaptive supercmputing architecture that

More information

NQueens Problem with CUDA

NQueens Problem with CUDA NQueens Prblem with CUDA Shuqing Chen (21760448) [MAP] [S.Chen] [NQueens Prblem with CUDA] Table f Cntents Backgrund Apprach Evaluatin Discussin Cnclusin [MAP] [S.Chen] [NQueens Prblem with CUDA] Backgrund

More information

CS4500/5500 Operating Systems Computer and Operating Systems Overview

CS4500/5500 Operating Systems Computer and Operating Systems Overview Operating Systems Cmputer and Operating Systems Overview Yanyan Zhuang Department f Cmputer Science http://www.cs.uccs.edu/~yzhuang UC. Clrad Springs Ref. MOS4E, OS@Austin, Clumbia, UWisc Overview Recap

More information

Operating systems. Module 7 IPC (Interprocess communication) PART I. Tami Sorgente 1

Operating systems. Module 7 IPC (Interprocess communication) PART I. Tami Sorgente 1 Operating systems Mdule 7 IPC (Interprcess cmmunicatin) PART I Tami Srgente 1 INTERPROCESS COMMUNICATION Prcesses within a system may be independent r cperating Cperating prcess can affect r be affected

More information

Because this underlying hardware is dedicated to processing graphics commands, OpenGL drawing is typically very fast.

Because this underlying hardware is dedicated to processing graphics commands, OpenGL drawing is typically very fast. The Open Graphics Library (OpenGL) is used fr visualizing 2D and 3D data. It is a multipurpse pen-standard graphics library that supprts applicatins fr 2D and 3D digital cntent creatin, mechanical and

More information

B Tech Project First Stage Report on

B Tech Project First Stage Report on B Tech Prject First Stage Reprt n GPU Based Image Prcessing Submitted by Sumit Shekhar (05007028) Under the guidance f Prf Subhasis Chaudhari 1. Intrductin 1.1 Graphic Prcessr Units A graphic prcessr unit

More information

ALU Design. ENG2410 Digital Design Datapath Design. Parts of CPU. Memory and I/O. Resources. Week #9 Topics. School of Engineering 1

ALU Design. ENG2410 Digital Design Datapath Design. Parts of CPU. Memory and I/O. Resources. Week #9 Topics. School of Engineering 1 ENG2410 Digital Design Datapath Design Datapath cnsists f: Parts f CPU Registers, Multiplexrs, Adders, Subtractrs and lgic t perfrm peratins n data (Cmb Lgic) Cntrl unit Generates signals t cntrl data-path

More information

KIRA-CS. Why KIRA-CS Supercomputer. Inside KIRA-CS. Elastic Supercomputing Architecture. Exa-Converged Architecture

KIRA-CS. Why KIRA-CS Supercomputer. Inside KIRA-CS. Elastic Supercomputing Architecture. Exa-Converged Architecture Explring Exa-Supercmputing Scenaris Fr New Emerging Challenges Why Supercmputer The series delivers n A3Cube s cmmitment t an adaptive supercmputing architecture that prvides bth extreme scalability and

More information

Eastern Mediterranean University School of Computing and Technology Information Technology Lecture2 Functions

Eastern Mediterranean University School of Computing and Technology Information Technology Lecture2 Functions Eastern Mediterranean University Schl f Cmputing and Technlgy Infrmatin Technlgy Lecture2 Functins User Defined Functins Why d we need functins? T make yur prgram readable and rganized T reduce repeated

More information

CSE 3320 Operating Systems Synchronization Jia Rao

CSE 3320 Operating Systems Synchronization Jia Rao CSE 3320 Operating Systems Synchrnizatin Jia Ra Department f Cmputer Science and Engineering http://ranger.uta.edu/~jra Recap f the Last Class Multiprcessr scheduling Tw implementatins f the ready queue

More information

Xilinx Answer Xilinx PCI Express DMA Drivers and Software Guide

Xilinx Answer Xilinx PCI Express DMA Drivers and Software Guide Xilinx Answer 65444 Xilinx PCI Express DMA Drivers and Sftware Guide Imprtant Nte: This dwnladable PDF f an Answer Recrd is prvided t enhance its usability and readability. It is imprtant t nte that Answer

More information

Extensible Query Processing in Starburst

Extensible Query Processing in Starburst Extensible Query Prcessing in Starburst Laura M. Haas, J.C. Freytag, G.M. Lhman, and H.Pirahesh IBM Almaden Research Center CS848 Instructr: David Tman Presented By Yunpeng James Liu Outline Intrductin

More information

Greg Dias, Alex Wilson. Fermi

Greg Dias, Alex Wilson. Fermi Greg Dias, Alex Wilsn Fermi Agenda Cmputer Graphics and Parallelism Basic GPU structure NVIDIA Fermi and CUDA Graphics Rendering Methds Plygn Rendering Ray Tracing Plygn Rendering Apprximatin f cmplex

More information

Impala: A Modern SQL Engine for Hadoop. James Kinley, Solutions Architect Cloudera, Inc.

Impala: A Modern SQL Engine for Hadoop. James Kinley, Solutions Architect Cloudera, Inc. Impala: A Mdern SQL Engine fr Hadp James Kinley, Slutins Architect Cludera, Inc. Agenda Gals User view f Impala Impala internals Cmparing Impala t ther systems Impala Overview: Gals General-purpse SQL

More information

CSE 3320 Operating Systems Computer and Operating Systems Overview Jia Rao

CSE 3320 Operating Systems Computer and Operating Systems Overview Jia Rao CSE 3320 Operating Systems Cmputer and Operating Systems Overview Jia Ra Department f Cmputer Science and Engineering http://ranger.uta.edu/~jra Overview Recap f last class What is an perating system?

More information

Low-Cost Solutions for Video Compression Systems

Low-Cost Solutions for Video Compression Systems Overview Lw-Cst Slutins fr Cmpressin Systems Brian Jentz Altera Crpratin 101 Innvatin Drive San Jse, CA 9505, USA (08) 5-7709 bjentz@altera.cm Many device applicatins utilize vide cmpressin t reduce the

More information

Dynamic Storage (ECS)

Dynamic Storage (ECS) User Guide Dynamic Strage (ECS) Swisscm (Schweiz) AG 1 / 10 Cntent 1 Abut Dynamic Strage... 3 2 Virtual drive, the EMC CIFS-ECS Tl... 4 3 Amazn S3 Brwer... 6 4 Strage Gateway Appliance... 9 5 Amazn S3

More information

SSDNow vs. HDD and Use Cases/Scenarios. U.S.T.S. Tech. Comm

SSDNow vs. HDD and Use Cases/Scenarios. U.S.T.S. Tech. Comm SSDNw vs. HDD and Use Cases/Scenaris U.S.T.S. Tech. Cmm Intrductin This white paper examines the technlgy f SSDNw and its ptential applicatin scenaris. Features and benefits f emplying Slid State Drives

More information

CS510 Concurrent Systems Class 2. A Lock-Free Multiprocessor OS Kernel

CS510 Concurrent Systems Class 2. A Lock-Free Multiprocessor OS Kernel CS510 Cncurrent Systems Class 2 A Lck-Free Multiprcessr OS Kernel The Synthesis kernel A research prject at Clumbia University Synthesis V.0 ( 68020 Uniprcessr (Mtrla N virtual memry 1991 - Synthesis V.1

More information

CS4500/5500 Operating Systems Synchronization

CS4500/5500 Operating Systems Synchronization Operating Systems Synchrnizatin Yanyan Zhuang Department f Cmputer Science http://www.cs.uccs.edu/~yzhuang UC. Clrad Springs Recap f the Last Class Multiprcessr scheduling Tw implementatins f the ready

More information

CodeSlice. o Software Requirements. o Features. View CodeSlice Live Documentation

CodeSlice. o Software Requirements. o Features. View CodeSlice Live Documentation CdeSlice View CdeSlice Live Dcumentatin Scripting is ne f the mst pwerful extensibility features in SSIS, allwing develpers the ability t extend the native functinality within SSIS t accmmdate their specific

More information

Application Notes for Stratus ftserver 6310 with VMWare and Avaya Aura Contact Center Release 6.2 Issue 1.0

Application Notes for Stratus ftserver 6310 with VMWare and Avaya Aura Contact Center Release 6.2 Issue 1.0 Avaya Slutin & Interperability Test Lab Applicatin Ntes fr Stratus ftserver 6310 with VMWare and Avaya Aura Cntact Center Release 6.2 Issue 1.0 Abstract These Applicatin Ntes describes the integratin,

More information

On the road again. The network layer. Data and control planes. Router forwarding tables. The network layer data plane. CS242 Computer Networks

On the road again. The network layer. Data and control planes. Router forwarding tables. The network layer data plane. CS242 Computer Networks On the rad again The netwrk layer data plane CS242 Cmputer Netwrks The netwrk layer The transprt layer is respnsible fr applicatin t applicatin transprt. The netwrk layer is respnsible fr hst t hst transprt.

More information

Retrieval Effectiveness Measures. Overview

Retrieval Effectiveness Measures. Overview Retrieval Effectiveness Measures Vasu Sathu 25th March 2001 Overview Evaluatin in IR Types f Evaluatin Retrieval Perfrmance Evaluatin Measures f Retrieval Effectiveness Single Valued Measures Alternative

More information

Tekmos. TK68020 Microprocessor. Features. General Description. 9/03/14 1

Tekmos. TK68020 Microprocessor. Features. General Description. 9/03/14   1 Tekms TK68020 Micrprcessr September 3, 2014 Prduct Overview Features Addressing Mde Extensins fr Enhanced Supprt f High-Level Languages Object-Cde Cmpatible with Earlier M68000 Micrprcessrs Addressing

More information

An Introduction to Crescendo s Maestro Application Delivery Platform

An Introduction to Crescendo s Maestro Application Delivery Platform An Intrductin t Crescend s Maestr Applicatin Delivery Platfrm Intrductin This dcument is intended t serve as a shrt intrductin t Crescend s Maestr Platfrm and its cre features/benefits. The dcument will

More information

Elasticity : Advanced Cloud Computing. Garth Gibson Greg Ganger Majd Sakr. Feb 6, Adv. Cloud Computing 1

Elasticity : Advanced Cloud Computing. Garth Gibson Greg Ganger Majd Sakr. Feb 6, Adv. Cloud Computing 1 Elasticity 15-719: Advanced Clud Cmputing Garth Gibsn Greg Ganger Majd Sakr Feb 6, 2017 15719 Adv. Clud Cmputing 1 Advanced Clud Cmputing Elasticity Readings Req d Ref 1: Dynamically Scaling Applicatins

More information

$ARCSIGHT_HOME/current/user/agent/map. The files are named in sequential order such as:

$ARCSIGHT_HOME/current/user/agent/map. The files are named in sequential order such as: Lcatin f the map.x.prperties files $ARCSIGHT_HOME/current/user/agent/map File naming cnventin The files are named in sequential rder such as: Sme examples: 1. map.1.prperties 2. map.2.prperties 3. map.3.prperties

More information

Parallel Processing in NCAR Command Language for Performance Improvement

Parallel Processing in NCAR Command Language for Performance Improvement Parallel Prcessing in NCAR Cmmand Language fr Perfrmance Imprvement Ping Gu, University f Wyming Mentr: Wei Huang, NCAR C- Mentr: Dave Brwn, NCAR August 1, 2013 Intrductin and Mtivatin ² The NCAR Cmmand

More information

Admin Report Kit for Exchange Server

Admin Report Kit for Exchange Server Admin Reprt Kit fr Exchange Server Reprting tl fr Micrsft Exchange Server Prduct Overview Admin Reprt Kit fr Exchange Server (ARKES) is an Exchange Server Management and Reprting slutin that addresses

More information

VEE CompSC: Live Migration with Pass-through Devices ZHENHAO PAN &, YAOZU DONG *, YU CHEN &, LEI ZHANG &, ZHIJIAO ZHANG &,

VEE CompSC: Live Migration with Pass-through Devices ZHENHAO PAN &, YAOZU DONG *, YU CHEN &, LEI ZHANG &, ZHIJIAO ZHANG &, CmpSC: Live Migratin with Pass-thrugh Devices ZHENHAO PAN &, YAOZU DONG *, YU CHEN &, LEI ZHANG &, ZHIJIAO ZHANG &, & Tsinghua University, *Intel Asia-Pacific Research and Develpment Ltd. VEE 2012 Outline

More information

Software Engineering

Software Engineering Sftware Engineering Chapter #1 Intrductin Sftware systems are abstract and intangible. Sftware engineering is an engineering discipline that is cncerned with all aspects f sftware prductin. Sftware Prducts

More information

Mobility Support by the Common API for Transparent Hybrid Multicast

Mobility Support by the Common API for Transparent Hybrid Multicast Mbility Supprt by the Cmmn API fr Transparent Hybrid Multicast draft-irtf-samrg-cmmn-api-03 Prject http://hamcast.realmv6.rg Matthias Wählisch, Thmas C. Schmidt, Stig Venaas {waehlisch, t.schmidt}@ieee.rg,

More information

Distributed Data Structures xfs: Serverless Network File System

Distributed Data Structures xfs: Serverless Network File System Advanced Tpics in Cmputer Systems, CS262B Prf Eric A. Brewer Distributed Data Structures xfs: Serverless Netwrk File System February 3, 2004 I. DDS Gal: new persistent strage layer that is a better fit

More information

Standardizing On A Set Of Radio Set APIs To Ensure Waveform Portability

Standardizing On A Set Of Radio Set APIs To Ensure Waveform Portability Standardizing On A Set Of Radi Set APIs T Ensure Wavefrm Prtability Glenn Fgarty Embedded Sftware Engineer The Being Cmpany, Anaheim, CA glenn.a.fgarty@being.cm (714) 762-0137 Wavefrm Applicatin Prtability

More information

Datacenter Traffic Measurement and Classification

Datacenter Traffic Measurement and Classification Datacenter Traffic Measurement and Classificatin Speaker: Lin Wang Research Advisr: Biswanath Mukherjee Grup meeting 6/15/2017 Datacenter Traffic Measurement and Analysis Data Cllectin Cllect netwrk events

More information

CSE 3320 Operating Systems Page Replacement Algorithms and Segmentation Jia Rao

CSE 3320 Operating Systems Page Replacement Algorithms and Segmentation Jia Rao CSE 0 Operating Systems Page Replacement Algrithms and Segmentatin Jia Ra Department f Cmputer Science and Engineering http://ranger.uta.edu/~jra Recap f last Class Virtual memry Memry verlad What if the

More information

Multi-Device Basic. Sample User's Guide. Intel SDK for OpenCL* Applications - Samples. Document Number: US

Multi-Device Basic. Sample User's Guide. Intel SDK for OpenCL* Applications - Samples. Document Number: US Sample User's Guide Intel SDK fr OpenCL* Applicatins - Samples Dcument Number: 329763-004US Cntents Legal Infrmatin... 3 Abut Multi-Device Basic Sample... 4 Algrithm... 4 OpenCL* Implementatin... 5 System-Level

More information

Aloha Offshore SDLC Process

Aloha Offshore SDLC Process Alha Sftware Develpment Life Cycle Alha Offshre SDLC Prcess Alha Technlgy fllws a sftware develpment methdlgy that is derived frm Micrsft Slutins Framewrk and Ratinal Unified Prcess (RUP). Our prcess methdlgy

More information

CS4500/5500 Operating Systems Page Replacement Algorithms and Segmentation

CS4500/5500 Operating Systems Page Replacement Algorithms and Segmentation Operating Systems Page Replacement Algrithms and Segmentatin Yanyan Zhuang Department f Cmputer Science http://www.cs.uccs.edu/~yzhuang UC. Clrad Springs Ref. MOSE, OS@Austin, Clumbia, Rchester Recap f

More information

Setac: A Phased Deterministic Testing Framework for Scala Actors. Samira Tasharofi Jun 02, 2011 Stanford, CA

Setac: A Phased Deterministic Testing Framework for Scala Actors. Samira Tasharofi Jun 02, 2011 Stanford, CA Setac: A Phased Deterministic Testing Framewrk fr Scala Actrs Samira Tasharfi Jun 02, 2011 Stanfrd, CA Mtivatin Schedule is a surce f nn-determinism in cncurrent prgrams Shared memry: rder f accesses Message-passing:

More information

Contents: Module. Objectives. Lesson 1: Lesson 2: appropriately. As benefit of good. with almost any planning. it places on the.

Contents: Module. Objectives. Lesson 1: Lesson 2: appropriately. As benefit of good. with almost any planning. it places on the. 1 f 22 26/09/2016 15:58 Mdule Cnsideratins Cntents: Lessn 1: Lessn 2: Mdule Befre yu start with almst any planning. apprpriately. As benefit f gd T appreciate architecture. it places n the understanding

More information

ARM Programmer s Model

ARM Programmer s Model ARM Prgrammer s Mdel Hsung-Pin Chang Department f Cmputer Science Natinal Chung Hsing University PDF created with FinePrint pdffactry Pr trial versin www.pdffactry.cm Outline ARM Data Types ARM Prcessr

More information

Understanding the Backward Slices of Performance Degrading Instructions

Understanding the Backward Slices of Performance Degrading Instructions Understanding the Backward Slices of Performance Degrading Instructions Craig Zilles and Guri Sohi University of Wisconsin - Madison International Symposium on Computer Architecture June, 2000 Motivation

More information

Infrastructure Series

Infrastructure Series Infrastructure Series TechDc WebSphere Message Brker / IBM Integratin Bus Parallel Prcessing (Aggregatin) (Message Flw Develpment) February 2015 Authr(s): - IBM Message Brker - Develpment Parallel Prcessing

More information

Chapter 5. The Network Layer IP

Chapter 5. The Network Layer IP Chapter 5 The Netwrk Layer IP These slides are taken frm the bk Cmputer etwrking, A Tp Dwn Apprach Featuring the Internet by Kurse & Rss and frm the bk Cmputer etwrks by Andrew Tanenbaum. The Netwrk Layer

More information

FLEXPOD A Scale-Out Converged System for the Next-Generation Data Center

FLEXPOD A Scale-Out Converged System for the Next-Generation Data Center FLEXPOD A Scale-Out Cnverged System fr the Next-Generatin Data Center A Scale-Out Cnverged System fr the Next-Generatin Data Center By Lee Hward Welcme t the age f scale-ut cnverged systems made pssible

More information

Definiens XD Release Notes

Definiens XD Release Notes Definiens XD 1.1.2 Release Ntes Errr! N text f specified style in dcument. Definiens XD 1.1.2 - Release Ntes Imprint and Versin Dcument Versin XD 1.1.2 Cpyright 2009 Definiens AG. All rights reserved.

More information

Programming Heterogeneous Many-cores Using Directives

Programming Heterogeneous Many-cores Using Directives Prgramming Hetergeneus Many-cres Using Directives HMPP - OpenAcc Françis Bdin, CAPS entreprise Intrductin Prgramming many-cre systems faces the fllwing dilemma The cnstraint f keeping a unique versin f

More information

Parallel error-correcting output codes classification in volume visualization: parallelism for AI and AI for parallelism

Parallel error-correcting output codes classification in volume visualization: parallelism for AI and AI for parallelism Parallel errr-crrecting utput cdes classificatin in vlume visualizatin: parallelism fr AI and AI fr parallelism Oscar Amrós Huguet Advisrs: Sergi Escalera, Anna Puig UPC-UB-URV Intrductin Main Gal: explre

More information

UNIT-II INTRODUCTION TO EMBEDDED SYSTEMS

UNIT-II INTRODUCTION TO EMBEDDED SYSTEMS UNIT-II INTRODUCTION TO EMBEDDED SYSTEMS Embedded System: Def: An Embedded system is ne that has cmputer hardware with sftware embedded in it as ne f its mst imprtant cmpnent. The three main cmpnents f

More information

Andrid prgramming curse Asynchrnus Techniques Intrductin Sessin bjectives Intrductin Asynchrnus Techniques Executr Handler AsyncTask Service & IntentService AsyncQueryHandler Lader By Võ Văn Hải Faculty

More information

Comparing Boosted Cascades to Deep Learning Architectures for Fast and Robust Coconut Tree Detection in Aerial Images

Comparing Boosted Cascades to Deep Learning Architectures for Fast and Robust Coconut Tree Detection in Aerial Images Cmparing Bsted Cascades t Deep Learning Architectures fr Fast and Rbust Ccnut Tree Detectin in Aerial Images VISAPP2018, 27-29 January 2018 Steven Puttemans*, Kristf Van Beeck* and Tn Gedemé Intrductin

More information

Due Date: Lab report is due on Mar 6 (PRA 01) or Mar 7 (PRA 02)

Due Date: Lab report is due on Mar 6 (PRA 01) or Mar 7 (PRA 02) Lab 3 Packet Scheduling Due Date: Lab reprt is due n Mar 6 (PRA 01) r Mar 7 (PRA 02) Teams: This lab may be cmpleted in teams f 2 students (Teams f three r mre are nt permitted. All members receive the

More information

Data Link Layer 10/28/2013. Data Link Layer. Adaptors Communicating. Data Link Layer. Session Transport Network Data Link Physical

Data Link Layer 10/28/2013. Data Link Layer. Adaptors Communicating. Data Link Layer. Session Transport Network Data Link Physical /28/23 Data Link Layer Department f Cmputer and IT Engineering University f Kurdistan Data Cmmunicatin Netwtks (Graduate level) Data Link Layer By: Dr. Alireza Abdllahpuri 2 Data Link Layer Applicatin

More information

Data Structure Interview Questions

Data Structure Interview Questions Data Structure Interview Questins A list f tp frequently asked Data Structure interview questins and answers are given belw. 1) What is Data Structure? Explain. Data structure is a way that specifies hw

More information

Transmission Control Protocol Introduction

Transmission Control Protocol Introduction Transmissin Cntrl Prtcl Intrductin TCP is ne f the mst imprtant prtcls f Internet Prtcls suite. It is mst widely used prtcl fr data transmissin in cmmunicatin netwrk such as Internet. Features TCP is reliable

More information

ARDUINO NANO INTRODUCTION

ARDUINO NANO INTRODUCTION ARDUINO NANO INTRODUCTION Arduin nan differ frm ther Arduin as it very small s it suitable fr small sized prjects and it supprts breadbards s it can be plugged with ther cmpnents in nly ne breadbard. ARDUINO

More information

RELEASE NOTES FOR PHOTOMESH 7.3.1

RELEASE NOTES FOR PHOTOMESH 7.3.1 RELEASE NOTES FOR PHOTOMESH 7.3.1 Abut PhtMesh Skyline s PhtMesh fully autmates the generatin f high-reslutin, textured, 3D mesh mdels frm standard 2D phtgraphs, ffering a significant reductin in cst and

More information

Data Communications over Context-Based WMNs Delay Performance Evaluation

Data Communications over Context-Based WMNs Delay Performance Evaluation 3rd IFIP/IEEE Internatinal Wrkshp n Bandwidth n Demand and Federatin Ecnmics (BD), C-lcated with NOMS Sympsium Osaka, Japan, 23rd April 2010 Data Cmmunicatins ver Cntext-Based WMNs Delay Perfrmance Evaluatin

More information

Dynamic Instruction Reuse

Dynamic Instruction Reuse Dynamic Instructin Reuse Guri Shi Cmputer Sciences Department University f Wiscnsin Madisn 1 Mtivatin Prgrams cnsist f static instructins Executin sees static instructin many times - ften with same inputs

More information

Verigy V93000 HSM HSM3G

Verigy V93000 HSM HSM3G Verigy V93000 HSM HSM3G Prduct Overview Industry Challenges High-end wrkstatin, desktp and laptp PCs, cmputer servers, perfrmance graphics cards, dynamic game cnsles, high-end vide/hdtv, cmputer netwrking

More information

Troubleshooting of network problems is find and solve with the help of hardware and software is called troubleshooting tools.

Troubleshooting of network problems is find and solve with the help of hardware and software is called troubleshooting tools. Q.1 What is Trubleshting Tls? List their types? Trubleshting f netwrk prblems is find and slve with the help f hardware and sftware is called trubleshting tls. Trubleshting Tls - Hardware Tls They are

More information

CS 309: Autonomous Intelligent Robotics. Instructor: Jivko Sinapov

CS 309: Autonomous Intelligent Robotics. Instructor: Jivko Sinapov CS 309: Autnmus Intelligent Rbtics Instructr: Jivk Sinapv http://www.cs.uteas.edu/~jsinapv/teaching/cs309_spring2017/ Machine Learning Annuncements Final Prject Presentatins Saturday, May 13, 7:00-10:00

More information

CS510 Concurrent Systems Class 1a. Linux Kernel Locking Techniques

CS510 Concurrent Systems Class 1a. Linux Kernel Locking Techniques CS510 Cncurrent Systems Class 1a Linux Kernel Lcking Techniques Intr t kernel lcking techniques (Linux) Why d we need lcking in the kernel? Which prblems are we trying t slve? What implementatin chices

More information

Telecommunication Protocols Laboratory Course

Telecommunication Protocols Laboratory Course Telecmmunicatin Prtcls Labratry Curse Lecture 2 March 11, 2004 http://www.ab.fi/~lpetre/teleprt/teleprt.html 1 Last time We examined sme key terms: prtcl, service, layer, netwrk architecture We examined

More information

CCNA 1 Chapter v5.1 Answers 100%

CCNA 1 Chapter v5.1 Answers 100% CCNA 1 Chapter 5 2016 v5.1 Answers 100% 1. What happens t runt frames received by a Cisc Ethernet switch? The frame is drpped. The frame is returned t the riginating netwrk device. The frame is bradcast

More information

Experience With Processes and Monitors in Mesa

Experience With Processes and Monitors in Mesa Advanced Tpics in Cmputer Systems, CS262A Prf. Eric Brewer Experience With Prcesses and Mnitrs in Mesa I. Experience With Prcesses and Mnitrs in Mesa Fcus f this paper: light-weight prcesses (threads in

More information

Course 6368A: Programming with the Microsoft.NET Framework Using Microsoft Visual Studio 2008

Course 6368A: Programming with the Microsoft.NET Framework Using Microsoft Visual Studio 2008 Curse 6368A: Prgramming with the Micrsft.NET Framewrk Using Micrsft Visual Studi 2008 5 Days Abut this Curse This five-day, instructr-led curse prvides an intrductin t develping n-tier applicatins fr the

More information

CSE 361S Intro to Systems Software Lab #2

CSE 361S Intro to Systems Software Lab #2 Due: Thursday, September 22, 2011 CSE 361S Intr t Systems Sftware Lab #2 Intrductin This lab will intrduce yu t the GNU tls in the Linux prgramming envirnment we will be using fr CSE 361S this semester,

More information

Performance and Scalability Benchmark: Siebel CRM Release 7.7 Industry Applications on IBM eserver p690 and IBM DB2 UDB on eserver p5 570

Performance and Scalability Benchmark: Siebel CRM Release 7.7 Industry Applications on IBM eserver p690 and IBM DB2 UDB on eserver p5 570 Perfrmance and Scalability Benchmark: Siebel CRM Release 7.7 Industry Applicatins n IBM eserver p690 and IBM DB2 UDB n eserver p5 570 An Oracle White Paper Released March 2005 Perfrmance and Scalability

More information

QC Clearance 4-Channel Surveillance Bundle

QC Clearance 4-Channel Surveillance Bundle QC444-411-5 Clearance 4-Channel Surveillance Bundle AT A GLANCE: 4 Channel H.264 CIF/D1 DVR 500GB Hard Drive Fur cameras with 400TV Lines f Reslutin Weatherprf cameras fr indr r utdr use 40 feet f night

More information

ENG2410 Digital Design Programmable Logic Technologies

ENG2410 Digital Design Programmable Logic Technologies ENG2410 Digital Design Prgrammable Lgic Technlgies Fall 2017 S. Areibi Schl f Engineering University f Guelph The Vn Neumann Cmputer Principle In 1945, the mathematician Vn Neumann (VN) demnstrated in

More information

Illusionist: Transforming Lightweight Cores into Aggressive Cores on Demand

Illusionist: Transforming Lightweight Cores into Aggressive Cores on Demand Illusinist: Transfrming s int Aggressive s n Demand Amin Ansari 1, Shuguang Feng 2, Shantanu Gupta 3, Jsep Trrellas 1, and Sctt Mahlke 4 1 University f Illinis, Urbana-Champaign 2 Nrthrp Grumman Crp. 3

More information

Profiling & Debugging

Profiling & Debugging Prfiling & Debugging CISC 879 Tristan Vanderbruggen & Jhn Cavazs Dept f Cmputer & Infrmatin Sciences University f Delaware 1 Lecture Overview Prfiling and Debugging Why? Tls Data sets Race Cnditin and

More information

CCNA 1 Chapter v5.1 Answers 100%

CCNA 1 Chapter v5.1 Answers 100% CCNA 1 Chapter 6 2016 v5.1 Answers 100% 1. Which characteristic f the netwrk layer in the OSI mdel allws carrying packets fr multiple types f cmmunicatins amng many hsts? the de-encapsulatin f headers

More information

RICOH IMAGING COMPANY, LTD. wishes to announce the release of the Windows Updater for update the

RICOH IMAGING COMPANY, LTD. wishes to announce the release of the Windows Updater for update the Dwnlads: Update Sftware IMAGE Transmitter 2 Update fr Windws Thank yu fr using PENTAX digital SLR camera. RICOH IMAGING COMPANY, LTD. wishes t annunce the release f the Windws Updater fr update the IMAGE

More information

Chapter 14. Basic Planning Methodology

Chapter 14. Basic Planning Methodology Chapter 14 Basic Planning Methdlgy This chapter prvides a basic and generic methdlgy fr planning prtectin requirements. It fcuses n the primary cnsideratins fr designing and implementing a basic strage

More information

Rust: using linear and region types to make the Internet more secure

Rust: using linear and region types to make the Internet more secure Rust: using linear and regin types t make the Internet mre secure Keith Wansbrugh, Metaswitch Netwrks Ltd. 2017-11-20 Metaswitch Netwrks 2017 1 Outline Cmmn vulnerabilities in sftware tday Linear types

More information

SAP Business One Hardware Requirements Guide

SAP Business One Hardware Requirements Guide Hardware Requirements Guide Dcument Versin: 1.13 2018-02-02 Release Family 9 Typgraphic Cnventins Type Style Example Descriptin Wrds r characters quted frm the screen. These include field names, screen

More information

The Current Forensic Workflow

The Current Forensic Workflow Vassil Russev The Current Frensic Wrkflw Frensic Target (3TB) Clne @150MB/s ~5.5 hrs Prcess @10MB/s ~82.5 hrs 129 * We can start wrking n the case after 88 hurs. * http://accessdata.cm/distributed-prcessing

More information

Practical Exercises in Computer Networks and Distributed Systems

Practical Exercises in Computer Networks and Distributed Systems (V..6, Nv 2) Practical Exercises in Cmputer Netwrks and Distributed Systems Stream Sckets and the Client/Server mdel (C language, W) 2-, Jsé María F Mrán This practical illustrates basic cncepts prtcl

More information

Chapter 2. The OSI Model and TCP/IP Protocol Suite. PDF created with FinePrint pdffactory Pro trial version

Chapter 2. The OSI Model and TCP/IP Protocol Suite. PDF created with FinePrint pdffactory Pro trial version Chapter 2 The OSI Mdel and TCP/IP Prtcl Suite PDF created with FinePrint pdffactry Pr trial versin www.pdffactry.cm Outline THE OSI MODEL LAYERS IN THE OSI MODEL TCP/IP PROTOCOL SUITE ADDRESSING TCP/IP

More information

Computer Organization and Architecture

Computer Organization and Architecture Campus de Gualtar 4710-057 Braga UNIVERSIDADE DO MINHO ESCOLA DE ENGENHARIA Departament de Infrmática Cmputer Organizatin and Architecture 5th Editin, 2000 by William Stallings Table f Cntents I. OVERVIEW.

More information

Advances in Real-Time Voxel-Based GI

Advances in Real-Time Voxel-Based GI Advances in Real-Time Vxel-Based GI Alexey Panteleev, Senir Develper Technlgy Engineer Rahul Sathe, Senir Develper Technlgy Engineer March 21, 2018 Bth #223 - Suth Hall www.nvidia.cm/gdc Recap n VXGI Vxel

More information

2. When logging is used, which severity level indicates that a device is unusable?

2. When logging is used, which severity level indicates that a device is unusable? CCNA 4 Chapter 8 v5.0 Exam Answers 2015 (100%) 1. What are the mst cmmn syslg messages? thse that ccur when a packet matches a parameter cnditin in an access cntrl list link up and link dwn messages utput

More information

COSC 4397 Parallel Computation. Performance Modeling. Edgar Gabriel. Spring Motivation

COSC 4397 Parallel Computation. Performance Modeling. Edgar Gabriel. Spring Motivation COSC 4397 Perfrmance Mdeling Spring 2010 Mtivatin Can we estimate the csts fr a parallel cde in rder t Evaluate quantitative and qualitative differences between different implementatin alternatives Understand

More information

Principles of Programming Languages

Principles of Programming Languages Principles f Prgramming Languages Slides by Dana Fisman based n bk by Mira Balaban and lecuture ntes by Michael Elhadad Dana Fisman Lessn 16 Type Inference System www.cs.bgu.ac.il/~ppl172 1 Type Inference

More information

COSC 6374 Parallel Computation. Performance Modeling (II) Edgar Gabriel. Spring Hockney s Model. t(s) = l + s/b

COSC 6374 Parallel Computation. Performance Modeling (II) Edgar Gabriel. Spring Hockney s Model. t(s) = l + s/b COSC 6374 Parallel Cmputatin Perfrmance Mdeling (II) Spring 2007 Hckney s Mdel t(s) = l + s/b (1) l: latency f the netwrk b: bandwidth f the netwrk Hw can we determine the latency and the bandwidth? Ping-png

More information