Interconnect Your Future
Supercomputing Nov 2015
Paving the Road to Exascale
© 2015 Mellanox Technologies 2
The Ever Growing Demand for Higher Performance
2000 202020102005
“Roadrunner”
1st
2015
Terascale Petascale Exascale
Single-Core to Many-CoreSMP to Clusters
Performance Development
Co-Design
HW SW
APP
Hardware
Software
Application
The Interconnect is the Enabling Technology
© 2015 Mellanox Technologies 3
Co-Design Architecture to Enable Exascale Performance
CPU-Centric Co-Design
Limited to Main CPU Usage
Results in Performance Limitation
Creating Synergies
Enables Higher Performance and Scale
Software
Software
In-CPU
Computing
In-Network
Computing
In-Storage
Computing
© 2015 Mellanox Technologies 4
The Intelligence is Moving to the Interconnect
CPU
Interconnect
Past Future
© 2015 Mellanox Technologies 5
Co-Design Architecture Depends on Offloading Technologies
Programmability
RDMA GPUDirect Virtualization
Backward and Future Compatibility
Direct Communications
Applications (Innovations, Scalability, Performance)
Software-Defined
Network (SDN)
Co-Design Requires Intelligent Interconnect
Offloading Technologies: Intelligent Interconnect
© 2015 Mellanox Technologies 6
Breaking the Application Latency Wall
§ Today: Network device latencies are on the order of 100 nanoseconds
§ Challenge: Enabling the next order of magnitude improvement in application performance
§ Solution: Creating synergies between software and hardware – intelligent interconnect
Intelligent Interconnect Paves the Road to Exascale Performance
10 years ago
~10
microsecond
~100
microsecond
NetworkCommunication
Framework
Today
~10
microsecond
Communication
Framework
~0.1
microsecond
Network
~1
microsecond
Communication
Framework
Future
~0.05
microsecond
Co-Design
Network
© 2015 Mellanox Technologies 7
Introducing Switch-IB 2 World’s First Smart Switch
© 2015 Mellanox Technologies 8
Introducing Switch-IB 2 World’s First Smart Switch
§ The world fastest switch with <90 nanosecond latency
§ 36-ports, 100Gb/s per port, 7.2Tb/s throughput, 7.02 Billion messages/sec
§ Adaptive Routing, Congestion control, support for multiple topologies
World’s First Smart Switch
Build for Scalable Compute and Storage Infrastructures
10X Higher Performance with The New Switch SHArP Technology
© 2015 Mellanox Technologies 9
SHArP (Scalable Hierarchical Aggregation Protocol) Technology
Delivering 10X Performance Improvement
for MPI and SHMEM/PAGS Applications
Switch-IB 2 Enables the Switch Network to
Operate as a Co-Processor
SHArP Enables Switch-IB 2 to Manage and
Execute MPI Operations in the Network
© 2015 Mellanox Technologies 10
The Intelligence is Moving to the Interconnect
Communication Frameworks (MPI, SHMEM/PGAS)
The Only Approach to Deliver 10X Performance Improvements
Applications Transport
RDMA
SR-IOV
Collectives
Peer-Direct
GPUDirect
More…
MPI / SHMEM Offloads
© 2015 Mellanox Technologies 11
High-Performance Designed 100Gb/s Interconnect Solutions
Transceivers
Active Optical and Copper Cables
(10 / 25 / 40 / 50 / 56 / 100Gb/s)
VCSELs, Silicon Photonics and Copper
36 EDR (100Gb/s) Ports, <90ns Latency
Throughput of 7.2Tb/s
7.02 Billion msg/sec (195M msg/sec/port)
100Gb/s Adapter, 0.7us latency
150 million messages per second
(10 / 25 / 40 / 50 / 56 / 100Gb/s)
32 100GbE Ports, 64 25/50GbE Ports
(10 / 25 / 40 / 50 / 100GbE)
Throughput of 6.4Tb/s
© 2015 Mellanox Technologies 12
Intelligent Interconnect Delivers Higher Datacenter ROI
Users
NETWORK
COMPUTING
NETWORK
Users
Intelligence
Network Offloads
Computing for applications
Smart Network
Increase Datacenter Value
Network functions
On CPU
COMPUTING
© 2015 Mellanox Technologies 13
Mellanox InfiniBand Proven and Most Scalable HPC Interconnect
“Summit” System “Sierra” System
Paving the Road to Exascale
© 2015 Mellanox Technologies 14
Technology Roadmap – One-Generation Lead over the Competition
2000 202020102005
“Roadrunner”
Mellanox Connected
1st3rd
TOP500 2003
Virginia Tech (Apple)
2015
Terascale Petascale Exascale
10G 20G 40G 56G 100G 200G 400G
TrueScale
No-Offload Network
40G (InfiniBand)
Same PathScale Technology
Under Intel Logo
InfiniPath
No-Offload Network
20G (InfiniBand)
OmniPath
No-Offload Network
100G (Proprietary)
Same PathScale Technology
Under QLogic Logo
© 2015 Mellanox Technologies 15
Mellanox InfiniBand Solutions Deliver Highest ROI for Any Scale
100
Gb/s
Link Speed
Higher Performance
Unlimited Scalability
Higher Resiliency
Proven!
Smart Network For Smart Systems
RDMA, Acceleration Engines, Programmability 200
Gb/s
Link Speed
2014
Gain Competitive Advantage Today, Protect Your Future
1 Mellanox Estimations
25%
Lower
20%
Lower
Scalability
CPU efficiency1
Power Consumption
Per Switch Port
Switch Latency
2X
Higher
Message Rate
44%
Higher
2017
© 2015 Mellanox Technologies 16
Mellanox Delivers Best Interconnect
§  100Gb/s throughput at 0% CPU utilization
§  Adapter: 150 Million messages/sec on today’s systems, 44% higher
§  Switch: 7.02 Billion messages/sec (195 Million per port)
§  20% lower switch latency, with deterministic latency!
Higher
Performance
Lower
TCO
Higher
Reliability
§  25% lower power consumption per switch port
§  Standard-based solutions, large eco-system support
§  Backward and future compatibility – protect investments
§  Offloading Architecture (RDMA, GPUDirect etc.) delivers highest system efficiency
§  1,000X higher reliability - Mellanox delivers Bit Error Rate of 10-15 versus 10-12
§  Superior signal integrity
§  Support for Multiple data integrity mechanisms (FEC1, LLR2, COD3 and more)
1- Forward Error Correction; 2 – Link Level Retransmission; 3- Correction on Demand
© 2015 Mellanox Technologies 17
End-to-End Interconnect Solutions for All Platforms
Highest Performance and Scalability for
X86, Power, GPU, ARM and FPGA-based Compute and Storage Platforms
10, 20, 25, 40, 50, 56 and 100Gb/s Speeds
X86
Open
POWER
GPU ARM FPGA
Smart Interconnect to Unleash The Power of All Compute Architectures
© 2015 Mellanox Technologies 18
Mellanox Solutions Interconnect Your Future
The Interconnect is The Competitive Advantage
Mellanox Interconnect
Performance, Scalability, Reliability, Proven
Speed-Up Your Present, Protect Your Future
Paving The Road to Exascale Computing Together
Thank You

Mellanox Announcements at SC15

  • 1.
    Interconnect Your Future SupercomputingNov 2015 Paving the Road to Exascale
  • 2.
    © 2015 MellanoxTechnologies 2 The Ever Growing Demand for Higher Performance 2000 202020102005 “Roadrunner” 1st 2015 Terascale Petascale Exascale Single-Core to Many-CoreSMP to Clusters Performance Development Co-Design HW SW APP Hardware Software Application The Interconnect is the Enabling Technology
  • 3.
    © 2015 MellanoxTechnologies 3 Co-Design Architecture to Enable Exascale Performance CPU-Centric Co-Design Limited to Main CPU Usage Results in Performance Limitation Creating Synergies Enables Higher Performance and Scale Software Software In-CPU Computing In-Network Computing In-Storage Computing
  • 4.
    © 2015 MellanoxTechnologies 4 The Intelligence is Moving to the Interconnect CPU Interconnect Past Future
  • 5.
    © 2015 MellanoxTechnologies 5 Co-Design Architecture Depends on Offloading Technologies Programmability RDMA GPUDirect Virtualization Backward and Future Compatibility Direct Communications Applications (Innovations, Scalability, Performance) Software-Defined Network (SDN) Co-Design Requires Intelligent Interconnect Offloading Technologies: Intelligent Interconnect
  • 6.
    © 2015 MellanoxTechnologies 6 Breaking the Application Latency Wall § Today: Network device latencies are on the order of 100 nanoseconds § Challenge: Enabling the next order of magnitude improvement in application performance § Solution: Creating synergies between software and hardware – intelligent interconnect Intelligent Interconnect Paves the Road to Exascale Performance 10 years ago ~10 microsecond ~100 microsecond NetworkCommunication Framework Today ~10 microsecond Communication Framework ~0.1 microsecond Network ~1 microsecond Communication Framework Future ~0.05 microsecond Co-Design Network
  • 7.
    © 2015 MellanoxTechnologies 7 Introducing Switch-IB 2 World’s First Smart Switch
  • 8.
    © 2015 MellanoxTechnologies 8 Introducing Switch-IB 2 World’s First Smart Switch § The world fastest switch with <90 nanosecond latency § 36-ports, 100Gb/s per port, 7.2Tb/s throughput, 7.02 Billion messages/sec § Adaptive Routing, Congestion control, support for multiple topologies World’s First Smart Switch Build for Scalable Compute and Storage Infrastructures 10X Higher Performance with The New Switch SHArP Technology
  • 9.
    © 2015 MellanoxTechnologies 9 SHArP (Scalable Hierarchical Aggregation Protocol) Technology Delivering 10X Performance Improvement for MPI and SHMEM/PAGS Applications Switch-IB 2 Enables the Switch Network to Operate as a Co-Processor SHArP Enables Switch-IB 2 to Manage and Execute MPI Operations in the Network
  • 10.
    © 2015 MellanoxTechnologies 10 The Intelligence is Moving to the Interconnect Communication Frameworks (MPI, SHMEM/PGAS) The Only Approach to Deliver 10X Performance Improvements Applications Transport RDMA SR-IOV Collectives Peer-Direct GPUDirect More… MPI / SHMEM Offloads
  • 11.
    © 2015 MellanoxTechnologies 11 High-Performance Designed 100Gb/s Interconnect Solutions Transceivers Active Optical and Copper Cables (10 / 25 / 40 / 50 / 56 / 100Gb/s) VCSELs, Silicon Photonics and Copper 36 EDR (100Gb/s) Ports, <90ns Latency Throughput of 7.2Tb/s 7.02 Billion msg/sec (195M msg/sec/port) 100Gb/s Adapter, 0.7us latency 150 million messages per second (10 / 25 / 40 / 50 / 56 / 100Gb/s) 32 100GbE Ports, 64 25/50GbE Ports (10 / 25 / 40 / 50 / 100GbE) Throughput of 6.4Tb/s
  • 12.
    © 2015 MellanoxTechnologies 12 Intelligent Interconnect Delivers Higher Datacenter ROI Users NETWORK COMPUTING NETWORK Users Intelligence Network Offloads Computing for applications Smart Network Increase Datacenter Value Network functions On CPU COMPUTING
  • 13.
    © 2015 MellanoxTechnologies 13 Mellanox InfiniBand Proven and Most Scalable HPC Interconnect “Summit” System “Sierra” System Paving the Road to Exascale
  • 14.
    © 2015 MellanoxTechnologies 14 Technology Roadmap – One-Generation Lead over the Competition 2000 202020102005 “Roadrunner” Mellanox Connected 1st3rd TOP500 2003 Virginia Tech (Apple) 2015 Terascale Petascale Exascale 10G 20G 40G 56G 100G 200G 400G TrueScale No-Offload Network 40G (InfiniBand) Same PathScale Technology Under Intel Logo InfiniPath No-Offload Network 20G (InfiniBand) OmniPath No-Offload Network 100G (Proprietary) Same PathScale Technology Under QLogic Logo
  • 15.
    © 2015 MellanoxTechnologies 15 Mellanox InfiniBand Solutions Deliver Highest ROI for Any Scale 100 Gb/s Link Speed Higher Performance Unlimited Scalability Higher Resiliency Proven! Smart Network For Smart Systems RDMA, Acceleration Engines, Programmability 200 Gb/s Link Speed 2014 Gain Competitive Advantage Today, Protect Your Future 1 Mellanox Estimations 25% Lower 20% Lower Scalability CPU efficiency1 Power Consumption Per Switch Port Switch Latency 2X Higher Message Rate 44% Higher 2017
  • 16.
    © 2015 MellanoxTechnologies 16 Mellanox Delivers Best Interconnect §  100Gb/s throughput at 0% CPU utilization §  Adapter: 150 Million messages/sec on today’s systems, 44% higher §  Switch: 7.02 Billion messages/sec (195 Million per port) §  20% lower switch latency, with deterministic latency! Higher Performance Lower TCO Higher Reliability §  25% lower power consumption per switch port §  Standard-based solutions, large eco-system support §  Backward and future compatibility – protect investments §  Offloading Architecture (RDMA, GPUDirect etc.) delivers highest system efficiency §  1,000X higher reliability - Mellanox delivers Bit Error Rate of 10-15 versus 10-12 §  Superior signal integrity §  Support for Multiple data integrity mechanisms (FEC1, LLR2, COD3 and more) 1- Forward Error Correction; 2 – Link Level Retransmission; 3- Correction on Demand
  • 17.
    © 2015 MellanoxTechnologies 17 End-to-End Interconnect Solutions for All Platforms Highest Performance and Scalability for X86, Power, GPU, ARM and FPGA-based Compute and Storage Platforms 10, 20, 25, 40, 50, 56 and 100Gb/s Speeds X86 Open POWER GPU ARM FPGA Smart Interconnect to Unleash The Power of All Compute Architectures
  • 18.
    © 2015 MellanoxTechnologies 18 Mellanox Solutions Interconnect Your Future The Interconnect is The Competitive Advantage Mellanox Interconnect Performance, Scalability, Reliability, Proven Speed-Up Your Present, Protect Your Future Paving The Road to Exascale Computing Together
  • 19.