The Dell MLPerf v1.1 submission included multinode results. This blog showcases performance across multiple nodes on Dell PowerEdge R750xa and XE8545 servers and demonstrates that the multinode scaling performance was excellent.

The compute requirement for deep learning training is growing at a rapid pace. It is imperative to train models across multiple nodes to attain a faster time-to-solution. Therefore, it is critical to showcase the scaling performance across multiple nodes. To demonstrate to customers the performance that they can expect across multiple nodes, our v1.1 submission includes multinode results. The following figures show multinode results for PowerEdge R750xa and XE8545 systems.

Figure 1: One-, two-, four-, and eight-node results with PowerEdge R750xa Resnet50 MLPerf v1.1 scaling performance

Figure 1 shows the performance of the PowerEdge R750xa server with Resnet50 training. These numbers scale from one node to eight nodes, from four NVIDIA  A100-PCIE-80GB GPUs to 32 NVIDIA A100-PCIE-80GB GPUs. We can see that the scaling is almost linear across nodes. MLPerf training requires passing Reference Convergence Points (RCP) for compliance. These RCPs were inhibitors to show linear scaling for the 8x scaling case. The near linear scaling makes a PowerEdge R750xa node an excellent choice for multinode training setup.

The workload was distributed by using singularity on PowerEdge R750xa servers. Singularity is a secure containerization solution that is primarily used in traditional HPC GPU clusters. Our submission includes setup scripts with singularity that help traditional HPC customers run workloads without the need to fully restructure their existing cluster setup. The submission also includes Slurm Docker-based scripts.

Figure 2: Multinode submission results for PowerEdge XE8545 server with BERT, MaskRCNN, Resnet50, SSD, and RNNT

Figure 2 shows the submitted performance of the PowerEdge XE8545 server with BERT, MaskRCNN, Resnet50, SSD, and RNNT training. These numbers scale from one node to two nodes, from four NVIDIA A100-SXM-80GB GPUs to eight NVIDIA  A100-SXM-80GB GPUs. All GPUs operate at 500W TDP for maximum performance. They were distributed using Slurm and Docker on PowerEdge XE8545 servers. The performance is nearly linear.

Note: The RNN-T single node results submitted for the PowerEdge  XE8545x4A100-SXM-80GB system used a different set of hyperparameters than for two nodes. After the submission, we ran the RNN-T benchmark again on the PowerEdge XE8545x4A100-SXM-80GB system with the same hyperparameters and found that the new time to converge is approximately 77.37 minutes. Because we only had the resources to update the results for the 2xXE8545x4A100-SXM-80GB system before the submission deadline, the MLCommons results show 105.6 minutes for a single node XE8545x4100-SXM-80GB system.

The following figure shows the adjusted representation of performance for the PowerEdge XE8545x4A100-SXM-80GB system. RNN-T provides an unverified score of 77.31 minutes[1]:

Figure 3: Revised multinode results with PowerEdge XE8545 BERT, MaskRCNN, Resnet50, SSD, and RNNT

Figure 3 shows the linear scaling abilities of the PowerEdge XE8545 server across different workloads such as BERT, MaskRCNN, ResNet, SSD, and RNNT. This linear scaling ability makes the PowerEdge XE8545 server an excellent choice to run large-scale multinode workloads.

Note: This rnnt.zip file includes log files for 10 runs that show that the averaged performance is 77.31 minutes.

Conclusion

• It is critical to measure deep learning performance across multiple nodes to assess the scalability component of training as deep learning workloads are growing rapidly.
• Our MLPerf training v1.1 submission includes multinode results that are linear and perform extremely well.
• Scaling numbers for the PowerEdge XE8545 and PowerEdge R750xa server make them excellent platform choices for enabling large scale deep learning training workloads across different areas and tasks.

Dell Technologies has completed the successful submission of MLPerf Training, which marks the seventh round of submission to MLCommons™. This blog provides an overview and highlights the performance of the Dell PowerEdge R750xa, XE8545, and DSS8440 servers that were used for the submission.

What’s new in MLPerf Training v2.1?

This round of submission does not include new benchmarks or changes in the existing benchmarks. A change is introduced in the submission compliance checker. 

This round adds one-sided normalization to the checker to reduce variance in the number of steps to converge. This change means that if a result converges faster than the RCP mean within a certain range, the checker normalizes the results to the RCP mean. This normalization was not available in earlier rounds of submission.

What’s new in MLPerf Training v2.1 with Dell submissions?

For Dell submission for MLPerf Training v2.1, we included:

• Improved performance with BERT and Mask R-CNN models
• Minigo submission results on Dell PowerEdge R750xa server with A100 PCIe GPUs

Overall Dell Submissions

Figure 1.     Overall submissions for all Dell PowerEdge servers in MLPerf Training v2.1

Figure 1 shows our submission in which the workloads span across image classification, lightweight and heavy object detection, speech recognition, natural language processing, recommender systems, medical image segmentation, and reinforcement learning. There were different NVIDIA GPUs including the A100, with PCIe and SXM4 form factors having 40 GB and 80 GB VRAM and A30.

The Minigo on the PowerEdge R750xa server is a first-time submission, and it takes around 516 minutes to run to target quality. That submission has 4x A100 PCIe 80 GB GPUs.

Our results have increased in count from 41 to 45. This increased number of submissions helps customers see the performance of the systems using different PowerEdge servers, GPUs, and CPUs. With more results, customers can expect to see the influence of using different hardware settings that can play a vital role in time to convergence.

We have several procured winning titles that demonstrate the higher performance of our systems in relation to other submitters, starting with the highest number of results across all the submitters. Some other titles include the top position in the time to converge for BERT, ResNet, and Mask R-CNN with our PowerEdge XE8545 server powered by NVIDIA A100-40GB GPUs.

Improvement in Performance for BERT and Mask R-CNN

Figure 2.     Performance gains from MLPerf v2.0 to MLPerf v2.1 running BERT

Figure 2 shows the improvements seen with the PowerEdge R750xa and PowerEdge XE8545 servers with A100 GPUs from MLPerf training v2.0 to MLPerf training v2.1 running BERT language model workload. The PowerEdge XE8545 server with A100-80GB has the fastest time to convergence and the highest improvement at 13.1 percent, whereas the PowerEdge XE8545 server with A100-40GB has 7.74 percent followed by the PowerEdge R750xa server with A100-PCIe at 5.35 percent.

Figure 3.     Performance gains from MLPerf v2.0 to MLPerf v2.1 running Mask R-CNN

Figure 3 shows the improvements seen with the PowerEdge XE8545 server with A100 GPUs. There is a 3.31 percent improvement in time to convergence with MLPerf v2.1.

For both BERT and Mask R-CNN, the improvements are software-based. These results show that software-only improvements can reduce convergence time. Customers can benefit from similar improvements without any changes in their hardware environment.

The following sections compare the performance differences between SXM and PCIe form factor GPUs.

Performance Difference Between PCIe and SXM4 Form Factor with A100 GPUs

Figure 4.     SXM4 form factor compared to PCIe for the BERT

Figure 5.     SXM4 form factor compared to PCIe for Resnet50 v1.5

Figure 6.     SXM4 form factor compared to PCIe for the RNN-T

Table 1:

Figures 4, 5, and 6 and Table 1 show that SXM form factor is faster than the PCIe form factor for BERT, Resnet50 v1.5, and RNN-T workloads.

The SXM form factor typically consumes more power and is faster than PCIe. For the above workloads, the minimum percentage improvement in convergence that customers can expect is in double digits, ranging from approximately 12 percent to 40 percent, depending on the workload.

Multinode Results Comparison

Multinode performance assessment is more important than ever. With the advent of large models and different parallelism techniques, customers have an ever-increasing need to find results faster. Therefore, we have submitted several multinode results to assess scaling performance.

Figure 7.     BERT multinode results with PowerEdge R750xa and XE8545 servers

Figure 7 indicates multinode results from three different systems with the following configurations:

1. R750xa with 4 A100-PCIe-80GB GPUs
2. XE8545 with 4 A100-SXM-40GB GPUs
3. XE8545 with 4 A100-SXM-80GB GPUs

Every node of the above system has four GPUs each. When the graph shows eight GPUs, it means that the performance results are derived from two nodes. Similarly, for 16 GPUs the results are derived from four nodes, and so on.

Figure 8.     Resnet50 multinode results with R750xa and XE8545 servers

Figure 9.     Mask R-CNN multinode results with R750xa and XE8545 servers

As shown in Figures 7, 8, and 9, the multinode scaling results of the BERT, Resnet50, and Mask R-CNN are linear or nearly linear scaled. This shows that Dell servers offer outstanding performance with single-node and multinode scaling.

Conclusion

The findings described in this blog show that:

• Dell servers can run all types of workloads in the MLPerf Training submission.
• Software-only enhancements reduce time to solution for our customers, as shown in our MLPerf Training v2.1 submission, and customers can expect to see improvements in their environments.
• Dell PowerEdge XE8545 and PowerEdge R750xa servers with NVIDIA A100 with PCIe and SXM4 form factors are both great selections for all deep learning models.
• PCIe-based PowerEdge R750xa servers can deliver reinforcement learning workloads in addition to other classes of workloads, such as image classification, lightweight and heavy object detection, speech recognition, natural language processing, and medical image segmentation.
• The single-node results of our submission indicate that Dell servers deliver outstanding performance and that multinode run scales well and helps to reduce time to solution across a distinct set of workload types, making Dell servers apt for single-node and multinode deep learning training workloads.
• The single-node results of our submission indicate that Dell servers deliver outstanding performance and that multinode results show a well-scaled performance that helps to reduce time to solution across a distinct set of workload types. This makes Dell servers apt for both small training workloads on single nodes and large deep learning training workloads on multinodes.

Appendix

System Under Test

MLPerf system configurations for PowerEdge XE8545 systems

MLPerf system configurations for Dell PowerEdge R750xa servers