Download - Heterogeneous Exascale Computing
© Copyright 2021 XilinxCopyright 2021 Xilinx©
Heterogeneous Exascale Computing
Ivo Bolsens, CTO
March 25, 2021
1
© Copyright 2021 XilinxCopyright 2021 Xilinx©
Digitalization
Towards Exascale Computing
Compute
P C E R A
2000 2010 2020
EXA
PETA
TERA
ASIC
RTL developers
Democratize hardware designCloud
Accelerator
Hardware savvy SW developers
HW/SW co-design
Virtex
Peer
Data scientists
Machine Learning
AI
Zynq
Alveo
Zen
Radeon
Versal
SCALAR VECTOR MATRIX SPATIAL
Mobile
Internet
© Copyright 2021 XilinxCopyright 2021 Xilinx©
3
Historic: CPU OnlyTop of Rack Switch (TOR)
© Copyright 2021 XilinxCopyright 2021 Xilinx©
4
Scale-Up: CPU Host With GPU and FPGA Accelerators
© Copyright 2021 XilinxCopyright 2021 Xilinx©
5
Scale-Out: CPU Host With GPU, FPGA and SmartNIC
HPC Ethernet
(Cray)
© Copyright 2021 XilinxCopyright 2021 Xilinx©
6
Scale-Up: Heterogeneous Fabric of Peers (CPU, GPU, FPGA, SSD)
SmartSSD
Communication
bottleneck through CPU
Communication
between peers
© Copyright 2021 XilinxCopyright 2021 Xilinx©
Communication
bottleneck through CPU
7
Scale-up: Heterogeneous fabric of Peers + Smart SSD
Accelerator
memory
Customizable
accelerator
5th gen
Samsung
VNAND
© Copyright 2021 XilinxCopyright 2021 Xilinx©
8
Scale-out : Compute On Data-In-Use, Data-In-Motion, Data-At-Rest
Data-In-Motion
Data-In-Use Data-At-Rest
© Copyright 2021 XilinxCopyright 2021 Xilinx©
Scale-Out : Disaggregation of Compute And Storage
1
1
2
9
1
2
FPGA# =Workload
CPU
GPU
Storage
NIC
© Copyright 2021 XilinxCopyright 2021 Xilinx©
10
Fabric
Data “In-Use” In Distributed Memory
Fabric
© Copyright 2021 XilinxCopyright 2021 Xilinx©
11
Communication using message passing
Software view : Local memory & message passing
© Copyright 2021 XilinxCopyright 2021 Xilinx©
12
Memory
Fabric
Future Programmer’s View: Unified Memory
Unified Distributed System Memory
Memory
Fabric
Any To Any Shared Memory Fabric
© Copyright 2021 XilinxCopyright 2021 Xilinx©
Memory
Fabric
Future: Shared Virtual Address Spaces
Memory
Fabric
13
Virtual Resource Group A
Virtual Resource Group B
© Copyright 2021 XilinxCopyright 2021 Xilinx©
14
Software Abstraction for Exascale Compute
Uniform System Memory
Any To Any Memory Fabric
Virtualized Address Spaces
© Copyright 2021 XilinxCopyright 2021 Xilinx©
OS
S h e l l
VITIS Software StackEnables Smart NIC|SSD + Heterogeneous Compute
SMART
NIC
F R A M E W OR K S
L I B R A R I E S
V++ dev ice code + XRT API
X R T
SMART
SSD
VIRTUALIZATION
ORCHESTRATION
COMPUTE
AIE
COMPUTE
PL
© Copyright 2021 XilinxCopyright 2021 Xilinx©
Conclusions
16
© Copyright 2021 XilinxCopyright 2021 Xilinx©
17
Future
Exascale computing and data - driven by new workloads such as AI
Increase compute power by
Scale-up of compute nodes
Scale-out of network
Enabled by Accelerators, SmartNIC and SmartSSD
Compute on data-in-use
Compute on data-in-motion
Compute on data-at-rest
Software stack :
Higher level of programming abstraction
Runtime enabling heterogeneous compute fabrics
Shared virtual address spaces
© Copyright 2021 Xilinx
Thank You
Copyright 2021 Xilinx©
18