Will the TOP500 use lower-precision for its primary ranking methodology on 1/1/26?
3
27
Ṁ36Ṁ110
2025
35%
chance
1D
1W
1M
ALL
Right now the way the TOP500 measures speed is in FP64. But lower-precision computing (e.g., FP8, FP16) is becoming more relevant for ML workloads.
This resolves positively iff the TOP500 switches its primary ranking methodology to measure computational speed in something less precise than FP64. "Primary" will refer to the most prominently displayed list on top500.org, so if a less prominent methodology uses FP(n<64), but the most prominent list still uses FP64, this resolves no. I will use my judgment in assessing prominence.
I will not bet in this market.
Get Ṁ200 play money
Related questions
Will GPT-5 launch before 1 January 2025?
63% chance
Will there be a model that has a 75% win rate against the latest iteration of GPT-4 as of January 1st, 2025?
37% chance
Will an AI solve any important mathematical conjecture before January 1st, 2030?
71% chance
Will a model be trained using at least as much compute as GPT-3 using AMD GPUs before Jan 1 2026?
71% chance
Will a popular forecasting platform accept function definitions as forecasting questions by Jan 1, 2025?
24% chance
Will any TOP500 computer use CPUs or accelerators built off of the RISC-V ISA in 2028?
48% chance
On LMSys, what will be the difference between the top model and GPT-4-1106 on Jan 1 2026?
207
Will a GPT-5-worthy model be able to break 50% on NYT Connections benchmark?
69% chance
On December 31, 2025, will the Hutter prize record be above 8.85?
67% chance
Will the November 2024 top 500 Supercomputer list put us above 10 Exaflops of total computing power?
44% chance