penn ese534 spring 2014 -- dehon 1 ese534: computer organization day 11: march 3, 2014 instruction...

35
Penn ESE534 Spring 2014 -- DeHon 1 ESE534: Computer Organization Day 11: March 3, 2014 Instruction Space Modeling 1

Upload: constance-bruce

Post on 29-Dec-2015

213 views

Category:

Documents


0 download

TRANSCRIPT

Penn ESE534 Spring 2014 -- DeHon1

ESE534:Computer Organization

Day 11: March 3, 2014

Instruction Space Modeling 1

Penn ESE534 Spring 2014 -- DeHon2

Last Time

• Instruction Requirements

• Instruction Space

Architecture Taxonomy

Penn ESE534 Spring 2014 -- DeHon3

PCs Pints/PC depth width Architecture

0 N 1 1 FPGA

1 N (48,640) 8 1 Tabula ABAX (A1EC04)

1 1 1024 32 Scalar Processor (RISC)

1 N D W VLIW (superscalar)

1 1 Small W*N SIMD, GPU, Vector

N 1 D W MIMD

16 1 (4?) 2048 64 16-core

Penn ESE534 Spring 2014 -- DeHon4

Today

• Model Architecture from Instruction Parameters – implied costs– gross application characteristics

• Focus on width as example– Instruction depth on Wed.

Penn ESE534 Spring 2014 -- DeHon5

Quotes

• If it can’t be expressed in figures, it is not science; it is opinion. -- Lazarus Long

Penn ESE534 Spring 2014 -- DeHon6

Modeling

• Why do we model?

Penn ESE534 Spring 2014 -- DeHon7

Motivation

• Need to understand– How costly is a solution

• Big, slow, hot, energy hungry….

– How compare to alternatives– Cost and benefit of flexibility

Penn ESE534 Spring 2014 -- DeHon8

What we really want:

• Complete implementation of our application

• For each architectural alternatives– In same implementation technology – w/ multiple area-time points

Penn ESE534 Spring 2014 -- DeHon9

Reality

• Seldom get it packaged that nicely – much work to do so– technology keeps moving

• We must deal with– estimation from components– technology differences– few area-time points

Penn ESE534 Spring 2014 -- DeHon10

Modeling Instruction Effects

• Restrictions from “ideal” +save area and energy– limit usability (yield) of PE

• May cost more energy, area in the end…

• Want to understand effects– Area, energy model– utilization/yield model

Preclass

• Energies? – 8-bit, 16-bit, 32-bit, 64-bit

• 16-bit on 32-bit?– Sources of inefficiency?

• 8-bit operations per 64-bit operation?

• 64-bit on 8-bit?– Sources of inefficiency?

Penn ESE534 Spring 2014 -- DeHon11

Area Target

• Switch focus to area-efficiency– Equivalently computational density

• Come back to energy-efficiency at end of lecture

Penn ESE534 Spring 2014 -- DeHon12

Efficiency/Yield Intuition(Area)

• What happens when– Datapath is too wide?– Datapath is too narrow?– Instruction memory is

too deep?

Penn ESE534 Spring 2014 -- DeHon13

Efficiency/Yield Intuition(Area)

• What happens when– Datapath is too wide?– Datapath is too narrow?– Instruction memory is

too deep?– Instruction memory is

too shallow?

Penn ESE534 Spring 2014 -- DeHon14

Penn ESE534 Spring 2014 -- DeHon15

Computing Device

• Composition– Bit Processing

elements– Interconnect: space– Interconnect: time– Instruction Memory

Tile together to build device

Penn ESE534 Spring 2014 -- DeHon16

Computing Device

Penn ESE534 Spring 2014 -- DeHon17

Computing Device

• Composition– Bit Processing

elements– Interconnect: space– Interconnect: time– Instruction Memory

Tile together to build device

Penn ESE534 Spring 2014 -- DeHon18

Relative Sizes

• Bit Operator 3-5KF2

• Bit Operator Interconnect 200K-250KF2

• Instruction (w/ interconnect) 20KF2

• Memory bit (SRAM) 250-500F2

Penn ESE534 Spring 2014 -- DeHon19

Model Area

Penn ESE534 Spring 2014 -- DeHon20

Architectures Fall in Space

Calibrate Model

Arch. w d c k Area (F2)

FPGA 1 1 1 4 220K

Altera 8K 230K

Xilinx 4K 160K

SIMD 1000 1 0 2 43K

Abacus 48K

Processor 32 32 1024 2 650K

MIPS-X 530K

Penn ESE534 Spring 2014 -- DeHon21

Penn ESE534 Spring 2014 -- DeHon22

Peak Densities from Model

Penn ESE534 Spring 2014 -- DeHon23

Peak Densities from Model

• Only 2 of 4 parameters– small slice of space– 100 density across

• Large difference in peak densities– large design space!

Penn ESE534 Spring 2014 -- DeHon24

Architectural parameters Peak Densities

Penn ESE534 Spring 2014 -- DeHon25

Efficiency

• What do we really want to maximize?– Not peak, “guaranteed not to exceed”

performance, but…– Useful work per unit silicon [per Joule]

• Yield Fraction / Area

• (or minimize (Area/Yielded performance) )

Penn ESE534 Spring 2014 -- DeHon26

Efficiency

• For comparison, look at relative efficiency to ideal.

• Ideal = architecture exactly matched to application requirements

• Efficiency = Aideal/Aarch

• Aarch = Area Op/Yield

Penn ESE534 Spring 2014 -- DeHon27

Width Mismatch Efficiency Calculation

E =Area(Task −on −matched − Architecture)

Area(Task −on − this − Architecture)

arch

task

wwbitelmarch

taskarch

wwbitelmtask

AWW

W

AWE

|

|

Penn ESE534 Spring 2014 -- DeHon28

Efficiency: Width Mismatch

c=1,16K PEs

Application Needs

• What are common application datawidths?

Penn ESE534 Spring 2014 -- DeHon29

Energy Target

• Switch to energy focus for rest of lecture

Penn ESE534 Spring 2014 -- DeHon30

Efficiency for Preclass

• Preclass 6 table

Penn ESE534 Spring 2014 -- DeHon31

E =Energy(Task −on −matched − Architecture)

Energy(Task −on − this − Architecture)

Robust Point

• Can we find an architecture that– Is reasonably efficient regardless of

application needs?– never less efficient than 50% ?

Penn ESE534 Spring 2014 -- DeHon32

Robust Point

• What is Energy Robust Point for preclass model?

Penn ESE534 Spring 2014 -- DeHon33

Penn ESE534 Spring 2014 -- DeHon34

Big Ideas[MSB Ideas]

• Applications typically have structure

• Exploit this structure to reduce resource requirements

• Architecture is about understanding and exploiting structure and costs to reduce requirements

Penn ESE534 Spring 2014 -- DeHon35

Admin• HW4 grading…in progress

• Office hours on Tuesday

• HW5.1 Due Wed.

• Reading for Wed. – Finish chapter