Express Cube Topologies for On-chip

InterconnectsBoris Grot

J. Hestness, S. W. Keckler, O. Mutlu†

The University of Texas at Austin†Carnegie Mellon University

‡Part of this work was performed at Microsoft Research

Feb 17, 2009HPCA ‘09

The Era of Many-core

UTCS 2HPCA ‘09

Intel Larrabee• 16+ cores• Bidirectional

ring interconnect

UT TRIPS• 2x16 exec tiles• 16 NUCA tiles• Multiple networks

Intel Polaris• 80 tiles• 8x10

mesh

Tilera Tile• 64 cores• 5 mesh networks

Networks on a Chip (NOCs) On-chip advantages

No pin constraints Rich wiring resources

On-chip limitations 2D substrates limit implementable topologies Logic area constrains use of wiring resources Energy/power budget caps

Focus Topologies for tomorrow’s many-core CMPs

HPCA ‘09 3UTCS

Outline Introduction Existing topologies Multidrop Express Channels (MECS) Evaluation Generalized Express Cubes Summary

UTCS 4HPCA '09

UTCS 5HPCA '09

2-D Mesh

Pros Low design & layout

complexity Simple, fast routers

Cons Large diameter Energy & latency

impact

UTCS 6HPCA '09

2-D Mesh

Pros Multiple terminals

attached to a router node Fast nearest-neighbor

communication via the crossbar

Hop count reduction proportional to concentration degree

Cons Benefits limited by

crossbar complexity

UTCS 7HPCA '09

Concentration (Balfour & Dally, ICS ‘06)

UTCS 8HPCA '09

Concentration

Side-effects Fewer channels Greater channel width

UTCS 9HPCA ‘09

Replication

CMesh-X2

Benefits Restores bisection

channel count Restores channel width Reduced crossbar

complexity

UTCS 10HPCA '09

Flattened Butterfly (Kim et al., Micro ‘07)

Objectives: Improve connectivity Exploit the wire budget

UTCS 11HPCA '09

Flattened Butterfly (Kim et al., Micro ‘07)

UTCS 12HPCA '09

Flattened Butterfly (Kim et al., Micro ‘07)

UTCS 13HPCA '09

Flattened Butterfly (Kim et al., Micro ‘07)

UTCS 14HPCA '09

Flattened Butterfly (Kim et al., Micro ‘07)

Pros Excellent connectivity Low diameter: 2 hops

Cons High channel count:

k2/2 per row/column Low channel utilization Increased control

(arbitration) complexity

UTCS 15HPCA '09

Flattened Butterfly (Kim et al., Micro ‘07)

UTCS 16HPCA '09

Multidrop Express Channels (MECS)

Objectives: Connectivity More scalable channel

count Better channel

utilization

UTCS 17HPCA '09

Multidrop Express Channels (MECS)

UTCS 18HPCA '09

Multidrop Express Channels (MECS)

UTCS 19HPCA '09

Multidrop Express Channels (MECS)

UTCS 20HPCA '09

Multidrop Express Channels (MECS)

UTCS 21HPCA ‘09

Multidrop Express Channels (MECS)

Pros One-to-many topology Low diameter: 2 hops k channels row/column Asymmetric

Cons Asymmetric Increased control

(arbitration) complexity

UTCS 22HPCA ‘09

Multidrop Express Channels (MECS)

Analytical Comparison

UTCS 23HPCA '09

CMesh FBfly MECS

Network Size 64 256 64 256 64 256

Radix (conctr’d) 4 8 4 8 4 8

Diameter 6 14 2 2 2 2

Channel count 2 2 8 32 4 8

Channel width 576 1152 144 72 288 288

Router inputs 4 4 6 14 6 14

Router outputs 4 4 6 14 4 4

Experimental Methodology

Topologies Mesh, CMesh, CMesh-X2, FBFly, MECS, MECS-X2

Network sizes 64 & 256 terminals

Routing DOR, adaptive

Messages 64 & 576 bits

Synthetic traffic Uniform random, bit complement, transpose, self-similar

PARSECbenchmarks

Blackscholes, Bodytrack, Canneal, Ferret, Fluidanimate, Freqmine, Vip, x264

Full-system config M5 simulator, Alpha ISA, 64 OOO cores

Energy evaluation Orion + CACTI 6

UTCS 24HPCA '09

UTCS 25HPCA '09

64 nodes: Uniform Random

0

10

20

30

40

1 4 7 10 13 16 19 22 25 28 31 34 37 40

Late

ncy

(cyc

les)

injection rate (%)

mesh cmesh cmesh-x2 fbfly mecs mecs-x2

UTCS 26HPCA '09

256 nodes: Uniform Random

0

10

20

30

40

50

60

70

1 4 7 10 13 16 19 22 25

Late

ncy

(cyc

les)

Injection rate (%)

mesh cmesh-x2 fbfly mecs mecs-x2

UTCS 27HPCA '09

Energy (100K pkts, Uniform Random)

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

2.0

Ave

rage

pac

ket e

ne

rgy

(nJ) Link Energy Router Energy

64 nodes 256 nodes

UTCS 28HPCA '09

64 Nodes: PARSEC

0

2

4

6

8

10

12

14

16

18

20

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.40

Router Energy Link Energy latency

Blackscholes Canneal Vip

Tota

l ne

two

rk E

ne

rgy

(J)

Avg

pac

ket

late

ncy

(cyc

les)

x264

Generalized Express Cubes Low-dimensional k-ary n-cube

n = {1,2} Good fit for planar silicon

Express channels Improve connectivity MECS for better wire utilization

Multiple networks Improve throughput Reduce crossbar area & energy overhead

Hierarchical scaling

UTCS 29HPCA '09

Partitioning: a GEC Example

UTCS 30HPCA '09

MECS

MECS-X2

FlattenedButterfly

PartitionedMECS

Summary MECS

A novel one-to-many topology Good fit for planar substrates Excellent connectivity Effective wire utilization

Generalized Express Cubes Framework & taxonomy for NOC topologies Extension of the k-ary n-cube model Useful for understanding and exploring

on-chip interconnect options Future: expand & formalize

UTCS 31HPCA '09

Summary MECS

A novel one-to-many topology Good fit for planar substrates Excellent connectivity Effective wire utilization

Generalized Express Cubes Framework & taxonomy for NOC topologies Extension of the k-ary n-cube model Useful for understanding and exploring

on-chip interconnect options Future: expand & formalize

UTCS 32HPCA '09

UTCS 33HPCA '09