Forecast
Facebook Twitter
  • Embed
  • Copy Link
    •
  • Approval
  • Set Reminder
  • Share private question
  • Unsubscribe from question
  • View Metaculus Prediction
  • {{qctrl.question.communityIsHidden ? 'Show' : 'Hide'}} Community Prediction
    •

    Your submission is now in Draft mode.

    Once it's ready, please submit your draft for review by our team of Community Moderators. Thank you!

    Submit Essay

    Once you submit your essay, you can no longer edit it.

    Pending

    This content now needs to be approved by community moderators.

    Submitted

    This essay was submitted and is waiting for review.

    How much computation did AlphaStar use for training?

    Created by jacobklagerros on {{qctrl.question.created_time | dateStr}}
    Published by jacobklagerros {{"onDate" | translate:({date: (qctrl.question.publish_time | dateStr)})}}

    Question

    This is related to the AlphaFold computation guesstimate question. Respecting Anthony's idiosyncratic linguistic preferences, this is also phrased not using "compute", but "computation" which is a noun rather than a verb.

    Computation in PFLOP/s-days used to train the versions of AlphaStar that beat top players TLO and MaNa with 5-0 in games. This includes all the computation necessary to run the AlphaStar League, but does not include hyperparameter tuning and architecture search.

    Announcement is here.

    Resolution

    Resolution by paper or other reliable announcement, closes 1 day prior to the public announcement.

    The method of calculations should be as similar as possible to that used in the "AI and compute" article.

    Note also that this article estimates actual rather than theoretical FLOPS, assuming a GPU utilization at 33% and CPU utilization at 17%.

    Categories:
    Compute guesstimates

    Make a Prediction

    Prediction

    Note: this question resolved before its original close time. All of your predictions came after the resolution, so you did not gain (or lose) any points for it.

    Note: this question resolved before its original close time. You earned points up until the question resolution, but not afterwards.

    Current points depend on your prediction, the community's prediction, and the result. Your total earned points are averaged over the lifetime of the question, so predict early to get as many points as possible! See the FAQ.