Leandro Soares Indrusiak http://www-users.cs.york.ac.uk/lsi Dagstuhl - - PowerPoint PPT Presentation

leandro soares indrusiak http users cs york ac uk lsi
SMART_READER_LITE
LIVE PREVIEW

Leandro Soares Indrusiak http://www-users.cs.york.ac.uk/lsi Dagstuhl - - PowerPoint PPT Presentation

Department of Computer Science Mixed Criticality Support on Networks-on-Chip Leandro Soares Indrusiak http://www-users.cs.york.ac.uk/lsi Dagstuhl Seminar 15121 March 2015 Mixed Criticality Support on NoCs | L. S. Indrusiak Many-Core


slide-1
SLIDE 1

Mixed Criticality Support on Networks-on-Chip

Leandro Soares Indrusiak

http://www-users.cs.york.ac.uk/lsi

Dagstuhl Seminar 15121 – March 2015

Department of Computer Science

slide-2
SLIDE 2

2 Mixed Criticality Support on NoCs | L. S. Indrusiak

Many-Core Systems

  • Many-core systems present a shift towards

communication-centric design

abundant computation resources shared communication media

  • Inter-core communication architectures

point-to-point on-chip bus network-on-chip

slide-3
SLIDE 3

3 Mixed Criticality Support on NoCs | L. S. Indrusiak

Router Link Core

Networks-on-Chip

C R R R R C R R C R R C R C C C C C Low capacitive load and short wires Scalable throughput Point to point connectivity

 Communication parallelism

Shared media

  • Reusability
  • Communication infrastructure based on links and routers

that interconnect cores providing packet-based data transfer

slide-4
SLIDE 4

4 Mixed Criticality Support on NoCs | L. S. Indrusiak

NoC parallelism and scalability

CPU CPU CPU CPU RAM CPU I/O CPU

Multiple connections simultaneously

slide-5
SLIDE 5

5 Mixed Criticality Support on NoCs | L. S. Indrusiak

NoC communication interference

link contention leads to latency variability

CPU CPU CPU CPU RAM CPU I/O CPU

slide-6
SLIDE 6

6 Mixed Criticality Support on NoCs | L. S. Indrusiak

Time predictability and isolation

  • Full traffic separation (i.e. no packet blocking)

deterministic routing, fully disjoint routes (e.g. Hermes) multiple overlay networks (e.g. Tilera), contention over NIs and memory still possible circuit switching (e.g. PNoC), unpredictable circuit setup time

  • Virtual traffic separation

fixed TDM traffic slotting (e.g. Aethereal, AElite) rate controlling (e.g. Nostrum, IDAMC) priority-arbitrated virtual channels (e.g. QNoC)

slide-7
SLIDE 7

7 Mixed Criticality Support on NoCs | L. S. Indrusiak

C R R R R C R R C R R C R C C C C C

Priority preemptive virtual channels

highest priority with remaining credit data_in credit_out data_out credit_in

routing & transmission control priority ID

highest priority with remaining credit

routing & transmission control

slide-8
SLIDE 8

8 Mixed Criticality Support on NoCs | L. S. Indrusiak

Packet latency

  • Packet flows suffer interference from other flows

that have higher priority and share at least one link

indirect interference also plays a role

  • Worst case latency can be found by an application
  • f Response Time Analysis
  • Z. Shi, A. Burns: Real-Time Communication Analysis for On-Chip Networks with Wormhole Switching. NOCS

2008: 161-170

slide-9
SLIDE 9

9 Mixed Criticality Support on NoCs | L. S. Indrusiak

Mixed criticality packet flows

  • Possible sources of uncertainty

packet length packet flow period jitter

  • All packet flows must be

schedulable under normal mode

  • Runtime monitoring detects

when packets go “beyond normal”

C R R R R C R R C R R C R C C C C C

network interface

slide-10
SLIDE 10

10 Mixed Criticality Support on NoCs | L. S. Indrusiak

Mixed criticality packet flows

  • Runtime monitoring detects

when packets go “beyond normal”

if it is a LO-CRIT packet exceeding its normal budget, reject it if it is a HI-CRIT packet exceeding its normal budget, signalise a mode change to the NoC, aiming to notify that a service degradation to LO-CRIT packets is needed so that HI-CRIT packets can still be scheduled despite

  • f potential increase of interference

due to overbudget packets C R R R R C R R C R R C R C C C C C

mode change notification

slide-11
SLIDE 11

11 Mixed Criticality Support on NoCs | L. S. Indrusiak

Mixed criticality packet flows

  • Two mode change

propagation protocols

WPMC: mode change flag “piggybacked” on packets that pass through a router that has changed mode WPMC-FLOOD: mode change is flooded to the entire NoC C R R R R C R R C R R C R C C C C C

  • A. Burns, J. Harbin, L. S. Indrusiak: A Wormhole NoC Protocol for Mixed Criticality Systems. RTSS 2014: 184-195
slide-12
SLIDE 12

12 Mixed Criticality Support on NoCs | L. S. Indrusiak

Mixed criticality packet flows

  • Two mode change

propagation protocols

WPMC: mode change flag “piggybacked” on packets that pass through a router that has changed mode WPMC-FLOOD: mode change is flooded to the entire NoC C R R R R C R R C R R C R C C C C C

  • L. S. Indrusiak , J. Harbin, A. Burns: Average and Worst-Case Latency Improvements in

Mixed-Criticality Wormhole Networks-on-Chip. ECRTS 2015 (submitted).

slide-13
SLIDE 13

13 Mixed Criticality Support on NoCs | L. S. Indrusiak

Mixed criticality packet flows

  • Two HI-CRIT mode arbitration

schemes

routers that change mode ignore arbitration requests of LO-CRIT packets routers that change mode arbitrate links in criticality order (HI-CRIT then LO-CRIT), and in priority order within the same criticality C R R R R C R R C R R C R C C C C C

  • A. Burns, J. Harbin, L. S. Indrusiak: A Wormhole NoC Protocol for Mixed Criticality Systems. RTSS 2014: 184-195
slide-14
SLIDE 14

14 Mixed Criticality Support on NoCs | L. S. Indrusiak

Mixed criticality packet flows

C R R R R C R R C R R C R C C C C C

  • L. S. Indrusiak , J. Harbin, A. Burns: Average and Worst-Case Latency Improvements in

Mixed-Criticality Wormhole Networks-on-Chip. ECRTS 2015 (submitted).

  • Two HI-CRIT mode arbitration

schemes

routers that change mode ignore arbitration requests of LO-CRIT packets routers that change mode arbitrate links in criticality order (HI-CRIT then LO-CRIT), and in priority order within the same criticality

slide-15
SLIDE 15

15 Mixed Criticality Support on NoCs | L. S. Indrusiak

Mixed criticality packet flows

  • Response Time Analysis formulations for each of

the protocols were developed

  • Evaluation with synthetic flowsets (against no

criticality awareness and criticality-monotonic arbitration) and cycle-accurate NoC simulation

WPMC-FLOOD slightly better in general, significantly better in stress scenarios

  • Less restrictive arbitration allows LO-CRIT packets

to flow when there are no HI-CRIT packets or when they are blocked due to interferences

slide-16
SLIDE 16

16 Mixed Criticality Support on NoCs | L. S. Indrusiak

Mixed criticality packet flows

slide-17
SLIDE 17

17 Mixed Criticality Support on NoCs | L. S. Indrusiak

Mixed criticality packet flows

slide-18
SLIDE 18

18 Mixed Criticality Support on NoCs | L. S. Indrusiak

Open issues

  • Handle recovery

how to detect that there are no further overbudget packets in the network? how to make sure their impact on the network (i.e. additional interference) is no longer there? how to notify all routers to return to normal mode?

  • Explore optimisations on task allocation and packet

routing

  • Improved experimental work

how many packet flows are HI-CRIT and how many are LO-CRIT? how much overbudget can HI-CRIT reasonably be?

slide-19
SLIDE 19

Mixed Criticality Support on Networks-on-Chip

Leandro Soares Indrusiak Alan Burns James Harbin

Dagstuhl Seminar 15121 – March 2015

Department of Computer Science