SlideShare a Scribd company logo
Delivering the Future of High-Performance Computing
Delivering the Future of High-Performance Computing
Delivering the Future of High-Performance Computing
Delivering the Future of High-Performance Computing
Delivering the Future of High-Performance Computing
▪
▪
▪
▪
▪
▪
Delivering the Future of High-Performance Computing
 AMD Internal
Delivering the Future of High-Performance Computing
Delivering the Future of High-Performance Computing
RETICLE LIMIT
Delivering the Future of High-Performance Computing
Delivering the Future of High-Performance Computing
Delivering the Future of High-Performance Computing
 See endnotes EPYC-07
Components of SPEC CPU®2017_int_ 2017 and 2006 at ISO Frequency
 See endnotes EPYC-07, ROM-236
Delivering the Future of High-Performance Computing
Delivering the Future of High-Performance Computing
Delivering the Future of High-Performance Computing
 See endnotes NAP-170, ROM-92
Delivering the Future of High-Performance Computing
Delivering the Future of High-Performance Computing
Delivering the Future of High-Performance Computing
Delivering the Future of High-Performance Computing
Delivering the Future of High-Performance Computing
▪
▪
▪
Interconnect
BW
Interconnect
Latencies
Interconnect
Topology
Routing
Optimization
Precision
Memory
BW
Memory
Capacity
On-Chip
Cache BW
Compute
Throughput
Delivering the Future of High-Performance Computing
Delivering the Future of High-Performance Computing
 See Endnotes
https://www.top500.org/
 See endnote
 See endnote
▪
▪
▪
 See Endnotes
 See Endnotes
▪
▪
 See Endnotes
Delivering the Future of High-Performance Computing
 See Endnotes
 See Endnotes
Delivering the Future of High-Performance Computing
Delivering the Future of High-Performance Computing
Delivering the Future of High-Performance Computing
Slides 4,5,6,9,10,11
Lisa T. Su, Samuel Naffziger, and Mark Papermaster, “Multi-Chip Technologies to Unleash Computing Performance Gains over the Next Decade,” IEDM Conference 2017.
Slide 11:
Original data up to the year 2010 collected and plotted by M. Horowitz, F Labonte O.Shacham, K. Olukotun, L. Hammond, and C. Batten.
New plot and data collected for 2010-2015 by K. Rupp. https://www.karlrupp.net/2015/06/40-years-of-microprocessor-trend-data/
Slide 16:
Testing by AMD Performance Labs as of 06/03/2019 utilizing 3rd Gen AMD Ryzen™ Processors: 3900X, 3800X, 3700X, 3600X, 3600 and Ryzen™ 7 2700X in Cinebench R20 1T.
Results may vary. RZ3-25
Based on June 8, 2018 AMD internal testing of same-architecture product ported from 14 to 7 nm technology with similar implementation flow/methodology, using performance
from SGEMM. EPYC-07
Based on AMD internal testing, average per thread performance improvement at ISO-frequency on a 32-core, 64-thread, 2nd generation AMD EPYC™ platform as compared to 32-
core 64-thread 1st generation AMD EPYC™ platform measured on a selected set of workloads including sub-components of SPEC CPU® 2017_int and representative server
workloads. ROM-236
Slide 20
The comparison is based on the highest performing results for two-processor servers using AMD EPYC 7601 processors and Intel Xeon Gold 6248 processors published on
www.spec.org as of April 27, 2019.
▪ • Score of 234 using 2 x Intel Xeon Gold 6248 processors.
▪ https://www.spec.org/cpu2017/results/res2019q2/cpu2017-20190318-11225.html
▪ • Score of 301 using 2 x AMD EPYC™ processor model 7601.
▪ https://www.spec.org/cpu2017/results/res2019q1/cpu2017-20190304-11124.html
▪ SPEC® and SPECrate® are registered trademarks of the Standard Performance Evaluation Corporation. Learn more at www.spec.org. NAP-170
Slides 20
A 2P EPYC™ 7742 processor powered server has SPECrate®2017_int_base score of 682, https://spec.org/cpu2017/results/res2019q3/cpu2017-20190722-16242.html as of August 7,
2019 The next highest base score is a 2P Intel Platinum 9282 server with a score of 643, http://spec.org/cpu2017/results/res2019q3/cpu2017-20190624-15369.pdf as of July 28,
2019. SPEC®, SPECrate® and SPEC CPU® are registered trademarks of the Standard Performance Evaluation Corporation. See www.spec.org for more information. ROM-92
Slide 29
Dario Amodei and Danny Hernandez. "AI and Compute." https://openai.com/blog/ai-and-compute/.
Slide 31, 32 , 35 AMD internal performance modeling and analysis
Slide 33 AMD Internal performance modeling and analysis
https://en.wikipedia.org/wiki/Graphics_Core_Next
Slide 38
https://www.top500.org/ and ORNL performance estimate.

More Related Content

PDF
The Path to "Zen 2"
 
PPTX
3D V-Cache
 
PPTX
Hot Chips: AMD Next Gen 7nm Ryzen 4000 APU
 
PDF
Delivering a new level of visual performance in an SoC AMD "Raven Ridge" APU
 
PDF
7nm "Navi" GPU - A GPU Built For Performance
 
PPTX
“Zen 3”: AMD 2nd Generation 7nm x86-64 Microprocessor Core
 
PPTX
AMD Chiplet Architecture for High-Performance Server and Desktop Products
 
PPTX
AMD Hot Chips Bulldozer & Bobcat Presentation
 
The Path to "Zen 2"
 
3D V-Cache
 
Hot Chips: AMD Next Gen 7nm Ryzen 4000 APU
 
Delivering a new level of visual performance in an SoC AMD "Raven Ridge" APU
 
7nm "Navi" GPU - A GPU Built For Performance
 
“Zen 3”: AMD 2nd Generation 7nm x86-64 Microprocessor Core
 
AMD Chiplet Architecture for High-Performance Server and Desktop Products
 
AMD Hot Chips Bulldozer & Bobcat Presentation
 

What's hot (20)

PPTX
Zen 2: The AMD 7nm Energy-Efficient High-Performance x86-64 Microprocessor Core
 
PDF
ISSCC 2018: "Zeppelin": an SoC for Multi-chip Architectures
 
PPTX
Heterogeneous Integration with 3D Packaging
 
PDF
AMD EPYC™ Microprocessor Architecture
 
PDF
AMD and the new “Zen” High Performance x86 Core at Hot Chips 28
 
PDF
AMD Ryzen CPU Zen Cores Architecture
PDF
Kernel Recipes 2017 - An introduction to the Linux DRM subsystem - Maxime Ripard
PPTX
GCC for ARMv8 Aarch64
PDF
"Snapdragon Hybrid Computer Vision/Deep Learning Architecture for Imaging App...
PDF
AMD: Where Gaming Begins
 
PDF
BKK16-315 Graphics Stack Update
PDF
Method of NUMA-Aware Resource Management for Kubernetes 5G NFV Cluster
PDF
System On Chip
PPTX
Digital Design Flow
PPTX
Lightelligence: Optical CXL Interconnect for Large Scale Memory Pooling
PPTX
Low latency in java 8 v5
PPT
An Overview on Programmable System on Chip: PSoC-5
PDF
Message Signaled Interrupts
PPTX
Modular by Design: Supermicro’s New Standards-Based Universal GPU Server
PPTX
Graphic Processing Unit (GPU)
Zen 2: The AMD 7nm Energy-Efficient High-Performance x86-64 Microprocessor Core
 
ISSCC 2018: "Zeppelin": an SoC for Multi-chip Architectures
 
Heterogeneous Integration with 3D Packaging
 
AMD EPYC™ Microprocessor Architecture
 
AMD and the new “Zen” High Performance x86 Core at Hot Chips 28
 
AMD Ryzen CPU Zen Cores Architecture
Kernel Recipes 2017 - An introduction to the Linux DRM subsystem - Maxime Ripard
GCC for ARMv8 Aarch64
"Snapdragon Hybrid Computer Vision/Deep Learning Architecture for Imaging App...
AMD: Where Gaming Begins
 
BKK16-315 Graphics Stack Update
Method of NUMA-Aware Resource Management for Kubernetes 5G NFV Cluster
System On Chip
Digital Design Flow
Lightelligence: Optical CXL Interconnect for Large Scale Memory Pooling
Low latency in java 8 v5
An Overview on Programmable System on Chip: PSoC-5
Message Signaled Interrupts
Modular by Design: Supermicro’s New Standards-Based Universal GPU Server
Graphic Processing Unit (GPU)
Ad

Similar to Delivering the Future of High-Performance Computing (20)

PDF
Amd ces tech day 2018 lisa su
PPTX
AMD EPYC 7002 Launch World Records
 
PDF
PCCC22:日本AMD株式会社 テーマ1「第4世代AMD EPYC™ プロセッサー (Genoa) の概要」
PPTX
SQLintersection keynote a tale of two teams
PPTX
OSS-10mins-7th2.pptx
PPTX
M|18 Intel and MariaDB: Strategic Collaboration to Enhance MariaDB Functional...
PPTX
Hardware and Software Co-optimization to Make Sure Oracle Fusion Middleware R...
PDF
Known basic of NFV Features
PDF
Ceph Day Beijing - SPDK for Ceph
PDF
Ceph Day Beijing - SPDK in Ceph
PDF
Intel® Xeon® Processor E5-2600 v3 Product Family Application Showcase - Fin...
PDF
DATA IS THE NEW OIL
PDF
Hp sizer for microsoft share point
PDF
Intel® Xeon® Processor E5-2600 v3 Product Family Application Showcase - Core ...
PPTX
AMD EPYC 7002 World Records
 
PPTX
Optimize DR and Cloning with Logical Hostnames in Oracle E-Business Suite (OA...
PPTX
Retour d'expérience d'un environnement base de données multitenant
PPTX
Ceph Day Taipei - Accelerate Ceph via SPDK
PDF
JBoss EAP 7 & JDG 7 최신 기술 소개
PPTX
22by7 and DellEMC Tech Day July 20 2017 - Power Edge
Amd ces tech day 2018 lisa su
AMD EPYC 7002 Launch World Records
 
PCCC22:日本AMD株式会社 テーマ1「第4世代AMD EPYC™ プロセッサー (Genoa) の概要」
SQLintersection keynote a tale of two teams
OSS-10mins-7th2.pptx
M|18 Intel and MariaDB: Strategic Collaboration to Enhance MariaDB Functional...
Hardware and Software Co-optimization to Make Sure Oracle Fusion Middleware R...
Known basic of NFV Features
Ceph Day Beijing - SPDK for Ceph
Ceph Day Beijing - SPDK in Ceph
Intel® Xeon® Processor E5-2600 v3 Product Family Application Showcase - Fin...
DATA IS THE NEW OIL
Hp sizer for microsoft share point
Intel® Xeon® Processor E5-2600 v3 Product Family Application Showcase - Core ...
AMD EPYC 7002 World Records
 
Optimize DR and Cloning with Logical Hostnames in Oracle E-Business Suite (OA...
Retour d'expérience d'un environnement base de données multitenant
Ceph Day Taipei - Accelerate Ceph via SPDK
JBoss EAP 7 & JDG 7 최신 기술 소개
22by7 and DellEMC Tech Day July 20 2017 - Power Edge
Ad

More from AMD (15)

PPTX
AMD EPYC Family World Record Performance Summary Mar 2022
 
PPTX
AMD EPYC Family of Processors World Record
 
PPTX
AMD EPYC Family of Processors World Record
 
PPTX
AMD EPYC World Records
 
PPTX
Hot Chips: AMD Next Gen 7nm Ryzen 4000 APU
 
PPTX
AMD EPYC 7002 World Records
 
PPTX
AMD Radeon™ RX 5700 Series 7nm Energy-Efficient High-Performance GPUs
 
PPTX
AMD EPYC 100 World Records and Counting
 
PPTX
AMD Next Horizon
 
PPTX
AMD Next Horizon
 
PDF
AMD Next Horizon
 
PDF
Race to Reality: The Next Billion-People Market Opportunity
 
PDF
GPU Compute in Medical and Print Imaging
 
PPTX
Enabling ARM® Server Technology for the Datacenter
 
PPTX
Lessons From MineCraft: Building the Right SMB Network
 
AMD EPYC Family World Record Performance Summary Mar 2022
 
AMD EPYC Family of Processors World Record
 
AMD EPYC Family of Processors World Record
 
AMD EPYC World Records
 
Hot Chips: AMD Next Gen 7nm Ryzen 4000 APU
 
AMD EPYC 7002 World Records
 
AMD Radeon™ RX 5700 Series 7nm Energy-Efficient High-Performance GPUs
 
AMD EPYC 100 World Records and Counting
 
AMD Next Horizon
 
AMD Next Horizon
 
AMD Next Horizon
 
Race to Reality: The Next Billion-People Market Opportunity
 
GPU Compute in Medical and Print Imaging
 
Enabling ARM® Server Technology for the Datacenter
 
Lessons From MineCraft: Building the Right SMB Network
 

Recently uploaded (20)

PDF
Build a system with the filesystem maintained by OSTree @ COSCUP 2025
PPTX
Detection-First SIEM: Rule Types, Dashboards, and Threat-Informed Strategy
PDF
Electronic commerce courselecture one. Pdf
PDF
Bridging biosciences and deep learning for revolutionary discoveries: a compr...
PPTX
A Presentation on Artificial Intelligence
PDF
7 ChatGPT Prompts to Help You Define Your Ideal Customer Profile.pdf
PPTX
Cloud computing and distributed systems.
PDF
Encapsulation theory and applications.pdf
PDF
Advanced methodologies resolving dimensionality complications for autism neur...
PDF
KodekX | Application Modernization Development
PDF
Per capita expenditure prediction using model stacking based on satellite ima...
PDF
Encapsulation_ Review paper, used for researhc scholars
PDF
Peak of Data & AI Encore- AI for Metadata and Smarter Workflows
PDF
Agricultural_Statistics_at_a_Glance_2022_0.pdf
PDF
The Rise and Fall of 3GPP – Time for a Sabbatical?
PDF
Mobile App Security Testing_ A Comprehensive Guide.pdf
PDF
Review of recent advances in non-invasive hemoglobin estimation
PDF
Reach Out and Touch Someone: Haptics and Empathic Computing
PPTX
PA Analog/Digital System: The Backbone of Modern Surveillance and Communication
PDF
TokAI - TikTok AI Agent : The First AI Application That Analyzes 10,000+ Vira...
Build a system with the filesystem maintained by OSTree @ COSCUP 2025
Detection-First SIEM: Rule Types, Dashboards, and Threat-Informed Strategy
Electronic commerce courselecture one. Pdf
Bridging biosciences and deep learning for revolutionary discoveries: a compr...
A Presentation on Artificial Intelligence
7 ChatGPT Prompts to Help You Define Your Ideal Customer Profile.pdf
Cloud computing and distributed systems.
Encapsulation theory and applications.pdf
Advanced methodologies resolving dimensionality complications for autism neur...
KodekX | Application Modernization Development
Per capita expenditure prediction using model stacking based on satellite ima...
Encapsulation_ Review paper, used for researhc scholars
Peak of Data & AI Encore- AI for Metadata and Smarter Workflows
Agricultural_Statistics_at_a_Glance_2022_0.pdf
The Rise and Fall of 3GPP – Time for a Sabbatical?
Mobile App Security Testing_ A Comprehensive Guide.pdf
Review of recent advances in non-invasive hemoglobin estimation
Reach Out and Touch Someone: Haptics and Empathic Computing
PA Analog/Digital System: The Backbone of Modern Surveillance and Communication
TokAI - TikTok AI Agent : The First AI Application That Analyzes 10,000+ Vira...

Delivering the Future of High-Performance Computing

  • 15.  See endnotes EPYC-07
  • 16. Components of SPEC CPU®2017_int_ 2017 and 2006 at ISO Frequency  See endnotes EPYC-07, ROM-236
  • 20.  See endnotes NAP-170, ROM-92
  • 42. Slides 4,5,6,9,10,11 Lisa T. Su, Samuel Naffziger, and Mark Papermaster, “Multi-Chip Technologies to Unleash Computing Performance Gains over the Next Decade,” IEDM Conference 2017. Slide 11: Original data up to the year 2010 collected and plotted by M. Horowitz, F Labonte O.Shacham, K. Olukotun, L. Hammond, and C. Batten. New plot and data collected for 2010-2015 by K. Rupp. https://www.karlrupp.net/2015/06/40-years-of-microprocessor-trend-data/ Slide 16: Testing by AMD Performance Labs as of 06/03/2019 utilizing 3rd Gen AMD Ryzen™ Processors: 3900X, 3800X, 3700X, 3600X, 3600 and Ryzen™ 7 2700X in Cinebench R20 1T. Results may vary. RZ3-25 Based on June 8, 2018 AMD internal testing of same-architecture product ported from 14 to 7 nm technology with similar implementation flow/methodology, using performance from SGEMM. EPYC-07 Based on AMD internal testing, average per thread performance improvement at ISO-frequency on a 32-core, 64-thread, 2nd generation AMD EPYC™ platform as compared to 32- core 64-thread 1st generation AMD EPYC™ platform measured on a selected set of workloads including sub-components of SPEC CPU® 2017_int and representative server workloads. ROM-236 Slide 20 The comparison is based on the highest performing results for two-processor servers using AMD EPYC 7601 processors and Intel Xeon Gold 6248 processors published on www.spec.org as of April 27, 2019. ▪ • Score of 234 using 2 x Intel Xeon Gold 6248 processors. ▪ https://www.spec.org/cpu2017/results/res2019q2/cpu2017-20190318-11225.html ▪ • Score of 301 using 2 x AMD EPYC™ processor model 7601. ▪ https://www.spec.org/cpu2017/results/res2019q1/cpu2017-20190304-11124.html ▪ SPEC® and SPECrate® are registered trademarks of the Standard Performance Evaluation Corporation. Learn more at www.spec.org. NAP-170
  • 43. Slides 20 A 2P EPYC™ 7742 processor powered server has SPECrate®2017_int_base score of 682, https://spec.org/cpu2017/results/res2019q3/cpu2017-20190722-16242.html as of August 7, 2019 The next highest base score is a 2P Intel Platinum 9282 server with a score of 643, http://spec.org/cpu2017/results/res2019q3/cpu2017-20190624-15369.pdf as of July 28, 2019. SPEC®, SPECrate® and SPEC CPU® are registered trademarks of the Standard Performance Evaluation Corporation. See www.spec.org for more information. ROM-92 Slide 29 Dario Amodei and Danny Hernandez. "AI and Compute." https://openai.com/blog/ai-and-compute/. Slide 31, 32 , 35 AMD internal performance modeling and analysis Slide 33 AMD Internal performance modeling and analysis https://en.wikipedia.org/wiki/Graphics_Core_Next Slide 38 https://www.top500.org/ and ORNL performance estimate.