slide-1 lacsi extreme computing mitre isimit lincoln laboratory this work is sponsored by the...

Slide-1LACSI

Extreme Computing

MITRE ISIMIT Lincoln Laboratory

This work is sponsored by the Department of Defense under Army Contract W15P7T-05-C-D001. Opinions, interpretations, conclusions, and recommendations are those of the author and are not necessarily endorsed by the United States Government.

The Path to Extreme Supercomputing—LACSI Workshop

DARPA HPCS

David Koester, Ph.D.DARPA HPCS Productivity Team

12 October 2004Santa Fe, NM

Slide-2LACSI

Extreme Computing


Outline

• Brief DARPA HPCS Overview– Impacts– Programmatics– HPCS Phase II Teams– Program Goals– Productivity Factors — Execution & Development Time– HPCS Productivity Team Benchmarking Working Group

• Panel Theme/Question – How much?– How fast?

Slide-3LACSI

Extreme Computing


High Productivity Computing Systems

Impact:Performance (time-to-solution): speedup critical national

security applications by a factor of 10X to 40XProgrammability (idea-to-first-solution): reduce cost and

time of developing application solutions Portability (transparency): insulate research and

operational application software from systemRobustness (reliability): apply all known techniques to

protect against outside attacks, hardware faults, & programming errors

Fill the Critical Technology and Capability GapToday (late 80’s HPC technology)…..to…..Future (Quantum/Bio Computing)

Fill the Critical Technology and Capability GapToday (late 80’s HPC technology)…..to…..Future (Quantum/Bio Computing)

Applications: Intelligence/surveillance, reconnaissance, cryptanalysis, weapons analysis, airborne contaminant

modeling and biotechnology

HPCS Program Focus Areas

Create a new generation of economically viable computing systems (2010) and a procurement methodology (2007-2010) for the security/industrial community

Slide-4LACSI

Extreme Computing


High Productivity Computing Systems

Phase 1 Phase 2(2003-2005)

Phase 3(2006-2010)

ConceptStudy

AdvancedDesign &Prototypes

Full ScaleDevelopment

Petascale/s Systems

Vendors

New EvaluationFramework

Test EvaluationFramework

Create a new generation of economically viable computing systems (2010) and a procurement methodology (2007-2010) for the security/industrial community

Validated ProcurementEvaluation Methodology

-Program Overview-

Productiv

ity Team

Half-Way PointPhase 2

TechnologyAssessment

Review

Slide-5LACSI

Extreme Computing


HPCS Phase II Teams

Mission Partners

Productivity Team (Lincoln Lead)

PI: SmithPI: Elnozahy PI: Mitchell

MIT Lincoln Laboratory

PI: Kepner PI: Lucas

PI: Koester

PI: Basili PI: Benson & Snavely

PIs: Vetter, Lusk, Post, Bailey PIs: Gilbert, Edelman, Ahalt, Mitchell

CSAIL OhioState

Industry

PI: Dongarra

Slide-6LACSI

Extreme Computing


HPCS Program GoalsProductivity Goals

• HPCS overall productivity goals:– Execution (sustained performance)

1 Petaflop/s (scalable to greater than 4 Petaflop/s)

Reference: Production workflow

– Development 10X over today’s systems Reference: Lone researcher and

Enterprise workflows

10x improvement in time to first solution!10x improvement in time to first solution!

Decide

Observe

Act

Orient

Production

Decide

Observe

Act

Orient

Production

Production

Experiment

Theory

Experiment

Theory

Researcher

Lone ResearcherDesign

Simulation

Visualize

Enterprise

Design

Simulation

Visualize

Enterprise

Port Legacy Software

Enterprise Execution

Development

Slide-7LACSI

Extreme Computing


HPCS Program GoalsProductivity Framework

Development Time

ExecutionTime

ProductivityMetrics

System Parameters

(Examples)BW bytes/flop (Balance)Memory latency Memory size……..

Productivity

Processor flop/cycle Number of processorsClock frequency………

Bisection bandwidthPower/system # of racks……….Code size Restart time Peak flops/sec …

Activity & Purpose

Benchmarks

Actual

System

or

Model

WorkFlows(Utility/Cost)

Development Time

ExecutionTime

ProductivityMetrics

System Parameters

(Examples)BW bytes/flop (Balance)Memory latency Memory size……..

Productivity

Processor flop/cycle Number of processorsClock frequency………

Bisection bandwidthPower/system # of racks……….Code size Restart time Peak flops/sec …

Activity & Purpose

Benchmarks

Actual

System

or

Model

WorkFlows(Utility/Cost)

U

C

U(T)

CS +CO +CM

Productivity = Utility/Cost

Slide-8LACSI

Extreme Computing


HPCS Program GoalsHardware Challenges

HPCS Program Goals &The HPCchallenge Benchmarks

HighLow

Low

PTRANS

FFT

MissionPartner

Applications

Spa

tial L

ocal

ity

Temporal Locality

RandomAccess

STREAMHPL

HighHigh

Low

Low

PTRANS

FFT

MissionPartner

Applications

Spa

tial L

ocal

ity

Temporal Locality

RandomAccess

STREAMHPL

High

• General purpose architecture capable of:

Subsystem Performance Indicators

1) 2+ PF/s LINPACK 2) 6.5 PB/sec data

STREAM bandwidth3) 3.2 PB/sec bisection

bandwidth4) 64,000 GUPS

Slide-9LACSI

Extreme Computing


Productivity FactorsExecution Time & Development Time

• Utility and some Costs are relative to– Workflow (WkFlow)– Execution Time (ExecTime)– Development Time (DevTime)

U

C

U(T)

CS +CO +CM

Productivity = Utility/Cost

Utility Cost Utility = f(WkFlow,ExecTime, DevTime)

Utility Operating Costs Procurement Costs

U(T)

Low (Bad)

ExecTime

De

vTim

e

Low

Low

High (Good)

CS

High (Bad)

Low (Good)

ExecTime

De

vTim

e

Low

Low

CO

High (Bad)

Low (Good)

ExecTime

De

vTim

e

Low

Low

• However, systems that will provide increased utility and decreased operating costs may have a higher initial procurement cost

– Need productivity metrics to justify the higher initial cost

• Reductions in both Execution Time and Development Time contribute to positive decreases in operating costs

– Reduction in programmer costs– More work performed over a

period

• Reductions in both Execution Time and Development Time contribute to positive increases in Utility

– Utility generally is inversely related to time

– Quicker is better

Slide-10LACSI

Extreme Computing


HPCS Benchmark Spectrum

8 HPCchallengeBenchmarks

Many (~40)Micro & KernelBenchmarks

9 SimulationApplications

LocalDGEMMSTREAM

RandomAcces1D FFT

GlobalLinpackPTRANS

RandomAccess1D FFT

Exi

stin

g A

pp

lica

tio

ns

Em

erg

ing

Ap

pli

cati

on

s

Fu

ture

Ap

pli

cati

on

s

SpanningKernels

DiscreteMath

…Graph

Analysis…

LinearSolvers

…Signal

Processing…

Simulation…I/O

ExecutionBounds

ExecutionIndicators

6 ScalableCompact Apps

Pattern MatchingGraph Analysis

SimulationSimulationSimulation

Signal Processing

PurposeBenchmarks

…

Others…

DevelopmentIndicators

Several (~10)Small ScaleApplications

System Bounds

• Spectrum of benchmarks provide different views of system– HPCchallenge pushes spatial and temporal boundaries; sets performance bounds– Applications drive system issues; set legacy code performance bounds

• Kernels and Compact Apps for deeper analysis of execution and development time

• Spectrum of benchmarks provide different views of system– HPCchallenge pushes spatial and temporal boundaries; sets performance bounds– Applications drive system issues; set legacy code performance bounds

• Kernels and Compact Apps for deeper analysis of execution and development time

CurrentUM2000GAMESS

OVERFLOWLBMHDRFCTHHYCOM

Near-FutureNWChemALEGRA

CCSM

Slide-11LACSI

Extreme Computing


Panel Theme/Question

• “How much should we change supercomputing to enable the applications that are important to us, and how fast?”

• How much? — HPCS is intended to “Fill the Critical Technology and Capability Gap between Today’s (late 80’s HPC technology)…..to…..Future (Quantum/Bio Computing)

• How fast?– Meaning when — HPCS SN-001 in 2010– Meaning performance — Petascale/s sustained

• I’m here to listen to you — the HPCS Mission Partners – Scope out emerging and future applications for 2010+

delivery (What applications will be important to you?)

– Collect data for the HPCS Vendors on future Applications Kernels Application characterizations and models

Slide-12LACSI

Extreme Computing


Statements

• Moore’s Law cannot go on forever

Proof: 2x →x→∞ ∞

So what?

• Moore’s Law doesn’t matter as long as we need to invest the increase in transistors into machine state — i.e., overhead — instead of real use

slide-1 lacsi extreme computing mitre isimit lincoln laboratory this work is sponsored by the...

Documents