• No results found

High Performance Multi Paradigm Messaging Runtime Integrating Grids and Multicore Systems

N/A
N/A
Protected

Academic year: 2020

Share "High Performance Multi Paradigm Messaging Runtime Integrating Grids and Multicore Systems"

Copied!
40
0
0

Loading.... (view fulltext now)

Full text

(1)

1

High Performanc

Multi-Paradigm

Messaging Runtime Integrating

Grids and Multicore Systems

e-Science 2007 Conference Bangalore India

December 13 2007

Geoffrey Fox, Huapeng Yuan, Seung-Hee Bae

Community Grids Laboratory, Indiana University Bloomington IN 47404

Xiaohong Qiu

Research Computing UITS

,

Indiana University Bloomington IN

George Chrysanthakopoulos, Henrik Frystyk Nielsen

Microsoft Research, Redmond WA

(2)

2

Abstract of Multicore Sals

Parallel Programming 2.0

n eScience applications need to use distributed Grid environments

where each component is an individual or cluster of multicore machines. These are expected to have 64-128 cores 5 years from now and need to support scalable parallelism.

n Users will want to compose heterogeneous components into single

jobs and run seamlessly in both distributed fashion and on a future “Grid on a chip” with different subsets of cores supporting individual components.

n We support this with a simple programming model made up of two

layers supporting traditional parallel and Grid programming paradigms (workflow) respectively.

n We examine for a parallel clustering application, the Concurrency

and Coordination Runtime CCR from Microsoft as a multi-paradigm runtime that integrates the two layers.

n Our work uses managed code (C#) and for AMD and Intel

(3)

Some links

n

See

http://www.connotea.org/user/crmc

for

references

-- select tag

oldies

for venerable links; tags like

MPI

Applications Compiler

have obvious significance

n

h

ttp://www.infomall.org/salsa f

or recent work

including publications

n

My tutorial on parallel computing

(4)

Too much Computing?

n Historically both grids and parallel computing have tried to

increase computing capabilities by

Optimizing performance of codes at cost of re-usabilityExploiting all possible CPU’s such as Graphics

co-processors and “idle cycles” (across administrative domains)

Linking central computers together such as NSF/DoE/DoD

supercomputer networks without clear user requirements

n Next Crisis in technology area will be the opposite problem

commodity chips will be 32-128way parallel in 5 years time and we currently have no idea how to use them on commodity systems – especially on clients

Only 2 releases of standard software (e.g. Office) in this

time span so need solutions that can be implemented in next 3-5 years

n Intel RMS analysis: Gaming and Generalized decision

(5)
(6)

Too much Data to the Rescue?

n Multicore servers have clear “universal parallelism” as many

users can access and use machines simultaneously

n Maybe also need application parallelism (e.g. datamining) as

needed on client machines

n Over next years, we will be submerged of course in data

deluge

Scientific observations for e-ScienceLocal (video, environmental) sensors

Data fetched from Internet defining users interests

n Maybe data-mining of this “too much data” will use up the

“too much computing” both for science and commodity PC’s

PC will use this data(-mining) to be intelligent user

assistant?

(7)

Parallel Programming Model

n If multicore technology is to succeed, mere mortals must be able to

build effective parallel programs on commodity machines

n There are interesting new developments – especially the new Darpa

HPCS Languages X10, Chapel and Fortress

n However if mortals are to program the 64-256 core chips expected in 5-7

years, then we must use near term technology and we must make it easy

This rules out radical new approaches such as new languages

n Remember that the important applications are not scientific computing

but most of the algorithms needed are similar to those explored in scientific parallel computing

n We can divide problem into two parts:

“Micro-parallelism”: High Performance scalable (in number of

cores) parallel kernels or libraries

Macro-parallelism: Composition of kernels into complete

applications

n We currently assume that the kernels of the scalable parallel

algorithms/applications/libraries will be built by experts with a

n Broader group of programmers (mere mortals) composing library

(8)

Multicore SALSA at CGL

n Service Aggregated Linked Sequential Activities

n Aims to link parallel and distributed (Grid) computing by

developing parallel applications as services and not as programs or libraries

Improve traditionally poor parallel programming

development environments

n Developing set of services (library) of multicore parallel data

mining algorithms

n Looking at Intel list of algorithms (and all previous experience),

we find there are two styles of “micro-parallelism”

Dynamic search as in integer programming, Hidden Markov Methods

(and computer chess); irregular synchronization with dynamic threads

“MPI Style” i.e. several threads running typically in SPMD (Single

Program Multiple Data); collective synchronization of all threads together n Most Intel RMS are “MPI Style” and very close to scientific

(9)
(10)

Scalable Parallel Components

n How do we implement micro-parallelism?

n There are no agreed high-level programming environments for

building library members that are broadly applicable.

n However lower level approaches where experts define

parallelism explicitly are available and have clear performance models.

n These include MPI for messaging or just locks within a single

shared memory.

n There are several patterns to support here including the

collective synchronization of MPI, dynamic irregular thread parallelism needed in search algorithms, and more specialized cases like discrete event simulation.

n We use Microsoft CC

(11)

There is MPI style messaging and ..

n OpenMP annotation or Automatic Parallelism of existing

software is practical way to use those pesky cores with existing code

As parallelism is typically not expressed precisely, one needs luck to get

good performance

Remember writing in Fortran, C, C#, Java … throws away information

about parallelism

n HPCS Languages should be able to properly express parallelism

but we do not know how efficient and reliable compilers will be

High Performance Fortran failed as language expressed a subset of

parallelism and compilers did not give predictable performance

n PGAS (Partitioned Global Address Space) like UPC, Co-array

Fortran, Titanium, HPJava

One decomposes application into parts and writes the code for each

component but use some form of global index

Compiler generates synchronization and messaging

PGAS approach should work but has never been widely used – presumably

(12)

Summary of micro-parallelism

n

On

new applications

, use MPI/locks with explicit

user decomposition

n

A subset of applications can use “

data parallel

compilers which follow in HPF footsteps

Graphics Chips and Cell processor motivate such

special compilers but not clear how many

applications can be done this way

n

OpenMP and/or Compiler-based Automatic

(13)

Composition of Parallel Components

n The composition (macro-parallelism) step has many excellent solutions

as this does not have the same drastic synchronization and correctness constraints as one has for scalable kernels

Unlike micro-parallelism step which has no very good solutions n Task parallelism in languages such as C++, C#, Java and Fortran90; n General scripting languages like PHP Perl Python

n Domain specific environments like Matlab and Mathematica n Functional Languages like MapReduce, F#

n HeNCE, AVS and Khoros from the past and CCA from DoE

n Web Service/Grid Workflow like Taverna, Kepler, InforSense KDE,

Pipeline Pilot (from SciTegic) and the LEAD environment built at Indiana University.

n Web solutions like Mash-ups and DSS

n Many scientific applications use MPI for the coarse grain composition

as well as fine grain parallelism but this doesn’t seem elegant

n The new languages from Darpa’s HPCS program support task

parallelism (composition of parallel components) decoupling

(14)

Integration of Services and “MPI”/Threads

n Kernels and Composition must be supported both inside chips (the multicore

problem) and between machines in clusters (the traditional parallel computing problem) or Grids.

n The scalable parallelism (kernel) problem is typically only interesting on true

parallel computers (rather than grids) as the algorithms require low communication latency.

n However composition is similar in both parallel and distributed scenarios and it

seems useful to allow the use of Grid and Web composition tools for the parallel problem.

This should allow parallel computing to exploit large investment in service

programming environments

n Thus in SALSA we express parallel kernels not as traditional libraries but as (some

variant of) services so they can be used by non expert programmers

n Bottom Line: We need a runtime that supports inter-service linkage and micro-parallelism linkage

n CCR and DSS have this property

Does it work and what are performance costs of the universality of runtime?Messaging need not be explicit for large data sets inside multicore node.

(15)

CICC Chemical Informatics and Cyberinfrastructure Collaboratory Web Service Infrastructure

Portal Services

RSS Feeds User Profiles

Collaboration as in Sakai

Core Grid Services

Service Registry

Job Submission and Management

Local Clusters

IU Big Red, TeraGrid, Open Science Grid

Varuna.net

Quantum Chemistry OSCAR Document Analysis

InChI Generation/Search

Computational Chemistry (Gamess, Jaguar etc.)

(16)

Deterministic Annealing for Data Mining

n We are looking at deterministic annealing algorithms because

although heuristic

They have clear scalable parallelism (e.g. use parallel BLAS)They avoid (some) local minima and regularize ill defined

problems in an intuitively clear fashion

They are fast (no Monte Carlo)

I understand them and Google Scholar likes them n Developed first by Durbin as Elastic Net for TSP

n Extended by Rose (my student then; now at UCSB)) and Gurewitz

(visitor to C3P) at Caltech for signal processing and applied later to

many optimization and supervised and unsupervised learning

methods.

n See K. Rose, "Deterministic Annealing for Clustering, Compression,

(17)

High Level Theory

n

Deterministic Annealing

can be looked at from a

Physics, Statistics and/or Information theoretic point of

view

n

Consider a function (e.g. a

likelihood

)

L({y})

that we

want to operate on (e.g.

maximize

)

n

Set

L

({y

},T) =

L({y}) exp(- ({y

} - {y})

2

/T ) d{y}

Incorporating entropy term ensuring that one looks for most

likely states at temperature T

If {y} is a distance, replacing L by Lcorresponds to smearing

or smoothing it over resolutionT

n

Minimize

Free Energy F = -Ln L

({y

},T)

rather than

energy E = -Ln L ({y})

Use mean field approximation to avoid Monte Carlo

(18)

Deterministic Annealing for Clustering I

n Illustrating similarity between clustering and Gaussian mixtures n Deterministic annealing for mixtures replaces by

(19)

Deterministic Annealing for Clustering II

n This is an extended K-means algorithm guaranteed not to diverge n Start with a single cluster giving as solution y1 as centroid

n For some annealing schedule for T, iterate above algorithm

testing correlation matrix in xi about each cluster center to see if

“elongated”

n Split cluster if elongation “long enough”; splitting is a phase

transition in physics view

n You do not need to assume number of clusters but rather a final

resolutionT or equivalent

(20)

n Minimum evolving as temperature decreases n Movement at fixed temperature going to local

minima if not initialized “correctly

Solve Linear Equations for each temperature

Nonlinearity removed by approximating with solution at previous higher temperature

Deterministi

Annealing

F({y}, T)

(21)

Clustering Data

n Cheminformatics was tested successfully with small datasets and

compared to commercial tools

n Cluster on properties of chemicals from high throughput

screening results to chemical properties (structure, molecular weight etc.)

n Applying to PubChem (and commercial databases) that have

6-20 million compounds

Comparing traditional fingerprint (binary properties) with real-valued

properties

n GIS uses publicly available Census data; in particular the 2000

Census aggregated in 200,000 Census Blocks covering Indiana

100MB of data

n Initial clustering done on simple attributes given in this data

Total population and number of Asian, Hispanic and Renters

n Working with POLIS Center at Indianapolis on clustering of

SAVI (Social Assets and Vulnerabilities Indicators) attributes at http://www.savi.org) for community and decision makers

(22)

30 Clusters

Renters

Asian

Hispanic

Total

30 Clusters

(23)
(24)

Where are we?

n We have deterministically annealed clustering running well on

8-core (2-processor quad 8-core) Intel systems using C# and Microsoft Robotics Studio CCR/DSS

n Could also run on multicore-based parallel machines but didn’t

do this (is there a large Windows quad core cluster on TeraGrid?)

This would also be efficient on large problems

n Applied to Geographical Information Systems (GIS) and census

data

Could be an interesting application on future broadly deployed PC’sVisualize nicely on Google Maps (and presumably Microsoft Virtual Earth)

n Applied to several Cheminformatics problems and have parallel

efficiency but visualization harder as in 150-1024 (or more) dimensions

n Will develop a family of such parallel annealing data-mining

tools where basic approach known for

Clustering

(25)

25

Microsoft CCR

Supports exchange of messages between threads using named ports

Fewer more general primitives than MPI

FromHandler: Spawn threads without reading ports

Receive: Each handler reads one item from a single port

MultipleItemReceive: Each handler reads a prescribed number of items of a given type from a given port. Note items in a port can be general structures but all must have same type.

MultiplePortReceive: Each handler reads a one item of a given type from multiple ports.

JoinedReceive: Each handler reads one item from each of two ports. The items can be of different type.

Choice: Execute a choice of two or more port-handler pairingsInterleave: Consists of a set of arbiters (port -- handler pairs) of 3

types that are Concurrent, Exclusive or Teardown (called at end for clean up). Concurrent arbiters are run concurrently but

exclusive handlers are

(26)

Preliminary Results

Parallel Deterministic Annealing Clustering

in

C# with

speed-up of 7

on Intel 2 quadcore

systems

Analysis of performance of

Java, C, C# in

MPI

and dynamic threading with XP, Vista,

Windows Server, Fedora, Redhat

on

Intel/AMD systems

Study of

cache effects

coming with MPI

thread-based parallelism

Study of

execution time fluctuations

in

(27)

Parallel Multicor

Deterministic Annealing Clustering

Parallel Overhea on 8 Threads Intel 8b

Speedup = 8/(1+Overhead)

10000/(Grain Size n = points per core) Overhead = Constant1 + Constant2/n

Constant1 = 0.05 to 0.1 (Client Windows) due to threa

runtime fluctuations

10 Clusters

(28)

Parallel Multicore

Deterministic Annealing Clustering

“Constant1”

Increasing number of clusters decreases communication/memory bandwidth overheads

Parallel Overhead for large (2M points) Indiana Census clusterin on 8 Threads Intel 8

(29)

Parallel Multicore

Deterministic Annealing Clustering

“Constant1”

Increasing number of clusters decreases communication/memory bandwidth overheads

Parallel Overhead for subset of PubChem clustering on 8 Threads (Intel 8b

The fluctuating overhead is reduced to 2% (as bits not doubles

(30)

Intel 8-core C# with 80 Clusters: Vista Run

Time Fluctuations for Clustering Kernel

• 2 Quadcore Processors

• This is average of standard deviation of run time of the 8 threads between messaging synchronization points

(31)

Intel 8 core with 80 Clusters: Redhat Run

Time Fluctuations for Clustering Kernel

This is average of standard deviation of run time of the

8 threads between messaging synchronization points

(32)
(33)

MPI Exchange Latency in µs (20-30 µs computation between messaging)

Machine OS Runtime Grains Parallelis

m MPI ExchangeLatency

Intel8c:gf12

(8 core 2.33 Ghz)

(in 2 chips)

Redhat MPJE

(Java) Process 8 181

MPICH2

(C) Process 8 40.0

MPICH2:

Fast Process 8 39.3

Nemesis Proces

s 8 4.21

Intel8c:gf20

(8 core 2.33 Ghz)

Fedora MPJE Proces

s 8 157

mpiJava Proces

s 8 111

MPICH2 Proces

s 8 64.2

Intel8b

(8 core 2.66 Ghz)

Vista MPJE Proces

s 8 170

Fedora MPJE Proces

s 8 142

Fedora mpiJava Proces

s 8 100

Vista CCR (C#) Threa

d 8 20.2

AMD4

(4 core 2.19 Ghz)

XP MPJE Proces

s 4 185

Redhat MPJE Proces

s 4 152

mpiJava Proces

s 4 99.4

MPICH2 Proces

s 4 39.3

XP CCR Threa

d 4 16.3

Intel4 (4 core 2.8

(34)

CCR Overhead for a computation of

23.76 µs between messaging

Rende vous

Intel8b: 8 Core Number of Parallel Computations

(μs) 1 2 3 4 7 8

Spawned

Pipeline 1.58 2.44 3 2.94 4.5 5.06

Shift 2.42 3.2 3.38 5.26 5.14

Two Shifts 4.94 5.9 6.84 14.32 19.44

MPI

Pipeline 2.48 3.96 4.52 5.78 6.82 7.18

Shift 4.46 6.42 5.86 10.86 11.74

Exchange As

Two Shifts 7.4 11.64 14.16 31.86 35.62

(35)
(36)

Cache Line Interference

Early implementations of our clustering algorithm

showed large fluctuations due to the cache line

interference effect discussed here and on next slide

in a simple case

We have one thread on each core each calculating a

sum of same complexity storing result in a common

array A with different cores using different array

locations

Thread i stores sum in A(i) is separation 1 – no

variable access interference but cache line

interference

Thread i stores sum in A(X*i) is separation X

Serious degradation if X < 8 (64 bytes) with Windows

Note A is a double (8 bytes)

(37)

Cache Line Interference

Note measurements at a separation X of 8 (and values between 8 and 1024 not shown) are essentially identical

Measurements at 7 (not shown) are higher than that at 8 (except for Red Hat which shows essentially no enhancement at X<8)

If effects due to co-location of thread variables in a 64 byte cache line, the array must be aligned with cache boundaries

(38)

DSS Section

• We view system as a collection of services

– in this case

One to supply data

One to run parallel clustering

One to visualize results – in this by spawning

a Google maps browser

Note we are clustering Indiana census data

(39)

39

Timing of HP Opteron Multicore as a function of number of simultaneous two-way service messages processed (November 2006 DSS Release) n Measurements of Axis 2 shows about 500 microseconds – DSS is 10 times better

(40)

References

Related documents

Psychology and Counseling (Christian Perspective) (3 credits) Description: This standard introduction to psychology and counseling offers an explanation of classic theory and

This finding does not coincide with the conventional wisdom that individual investors have a stronger impact on small stock returns, at least if company size is a proxy for the ratio

Skeide, Dilation theory and continuous tensor product systems of Hilbert modules, PQQP: Quantum Probability and White Noise Analysis XV (2003), World Scientific..

We found that the concentrations of radon for these samples and the percentage of excess allowable limits for these samples was 11% for the city of Ramadi and transmission

3. A market analysis of the future potential of organic trout in selected European countries. The first approach looked into the possibility for organic trout to achieve a price

In the context of a breakdown of the distribution network, coliforms can also be used as indicators of the integrity and the state of this network like other

Time Series plot for Forcast using ar (2). Looking at Figs. 4, it shows that the P-values for the Ljung-Box statistics are not significant. The forecast as shown in Fig. 7 does