Skip to navigation

Performance

View as Markdown

The AlphaFold2-Multimer NIM has been tuned to improve performance. However, the Multiple Sequence Alignment (MSA) and structure prediction processes are still computationally expensive. Performance also varies depending on:

  • Which NVIDIA GPUs are attached and available to the NIM.
  • How many CPU cores are available to the NIM.
  • The speed of the Solid State Drive (SSD) available to the NIM.
  • The parameters used to configure the NIM at runtime.
  • The number and composition of amino acid sequences provided to the NIM.

Performance expectations and general tips follow. These are not meant to be indicative of expected performance, and performance on your system will vary from these values. Supported hardware is in Prerequisites.

  • At least 1x NVIDIA A100 GPU.
  • At least 1 terabyte (1,000 gigabytes) of SSD storage (preferably NVMe Gen 4 with a speed of more than 3,500 MB/s).
  • At least 24 CPU cores (physical cores).
  • An internet connection with a download speed of at least 200 Mbps.

Performance Guidelines

In general, users should expect the following performance:

  • A100 and at least 64 CPU cores and at least 180 GB of RAM (fastest)
    • Sequences up to 3,000 amino acids should succeed, but the NIM can take several hours to run. This is the fastest MSA configuration.
  • A100 and at least 32 CPU cores and at least 128 GB of RAM (good)
    • Similar speed to the fastest configuration, but with more limited MSA performance. Use fewer runners per MSA and fewer cores per MSA (refer to Sequence to MSA).
  • L40S and at least 32 CPU cores and at least 128 GB of RAM (good)
    • Reduced structural prediction speed compared to A100. Similar MSA performance compared to using the same number of cores with an A100 GPU.
  • NVIDIA CUDA GPU with at least 32 GB of VRAM and at least 12 CPU cores and at least 64 GB of RAM (minimum)
    • Short sequences will run on this configuration, but sequences longer than 600 amino acids might fail in either MSA or structural prediction. MSA will be very slow, and the number of MSA runners should be set to 1 (refer to Sequence to MSA).

Sequence to MSA

  • MSA performance is largely dependent on disk speed and CPU cores. Make sure your NIM cache is on a fast SSD, and that you have allocated at least 24 cores to the NIM.
  • Increase the threads used per MSA process. This can be done using the NIM_PARALLEL_THREADS_PER_MSA environment variable.
  • Increase the number of concurrent MSA processes. This can be done using the NIM_PARALLEL_MSA_RUNNERS environment variable.
  • Refer to Configuring the NIM at Runtime for these parameters.

MSA runtime for short sequences can vary dramatically, but generally runtime scales linearly with sequence length and the number of input sequences to align. For sequences with many hits in the database (that is, sequences with motifs observed in many other proteins), the MSA runtime can grow significantly regardless of sequence length.

In general, expect MSA to take between one minute and six hours per query. These are approximate runtime estimates, and performance will vary between machines and input queries.

MSA to Structure and Sequence to Structure

  • Structure prediction performance is mostly dependent on GPU capability. If structure prediction is a bottleneck, consider a GPU with more compute and memory (for example, an A100 instead of an A30).

The time required for structure prediction grows exponentially with sequence length. As a rule of thumb, a sequence that is twice as long will take at least four times as long during the structural prediction phase.

In general, expect the structural prediction component to take between four minutes and 24 hours. These are approximate runtime estimates, and performance will vary between machines and input queries.