MLPerf: How Deci and Intel Achieved up to 16.8x Throughput Increase and +1.74% Accuracy Improvement

Authors: Amos Gropp, Deci AI, and Guy Boudoukh, Intel Labs

 

 

 

 

 

 

Figure 1: Deci’s AutoNAC Process

 

 

 

Table 2: Comparison of Current and Previous Submission Results — Throughput per Core

 

Configuration Details

c6i.2xlarge

8 vcpu (Intel® Xeon® Platinum 8375C Processor), 16 GB total memory, bios: SMBIOS 2.7, ucode: 0xd000331, Ubuntu 18.04.5 LTS, 5.4.0–1069-aws, gcc 9.3.0 compiler. Baseline: Resnet50, Accuracy=76.4; DeciNet, Accuracy=78.14

c6i.16xlarge

64 vcpu (Intel® Xeon® Platinum 8375C Processor), 128 GB total memory, bios: SMBIOS 2.7, ucode: 0xd000331, Ubuntu 18.04.5 LTS, 5.4.0–1069-aws, gcc 9.3.0 compiler. Baseline: Resnet50, Accuracy=76.4; DeciNet, Accuracy=78.14

m5zn.6xlarge

24 vcpu (Intel® Xeon® Platinum 8252C Processor), 96 GB total memory, bios: SMBIOS 2.7, ucode: 0x500320a, Ubuntu 18.04.5 LTS, 5.4.0–1069-aws, gcc 9.3.0 compiler. Baseline: Resnet50, Accuracy=76.4; DeciNet, Accuracy=78.14

 

Performance varies by use, configuration and other factors. Learn more at www.Intel.com/PerformanceIndex.

Performance results are based on testing as of dates shown in configurations and may not reflect all publicly available updates. See backup for configuration details. No product or component can be absolutely secure.

Your costs and results may vary.

Intel technologies may require enabled hardware, software or service activation.

© Intel Corporation. Intel, the Intel logo, and other Intel marks are trademarks of Intel Corporation or its subsidiaries. Other names and brands may be claimed as the property of others.

 

This was first published here.