Skip to content
TheChatArchiveThe Chat Archive
Claudeclaude-3-7-sonnet-20250219· AnthropicHugging Face LMArena Human Preference 140K

how fast do I have to go back in time for my iphone to outperform the best super…

From a Chatbot Arena battleThe same person sent this prompt to two anonymous models at once. Compare it with chatgpt-4o-latest-20250326’s answer.
  1. User

    how fast do I have to go back in time for my iphone to outperform the best super computer of the time at supercomuting workloads (whatever those were at the time)

    don't just look at raw flops numbers, consider the workload

    if my phone can't do something because it lacks the memory and/or memory bandwidth, then consider the old supercomputer to be better

  2. Claude Assistant

    iPhone vs. Historical Supercomputers

    To answer this question properly, I need to consider not just raw FLOPS but also memory capacity, memory bandwidth, and the ability to handle actual supercomputing workloads of their respective eras.

    Modern iPhone Capabilities (iPhone 15 Pro)

    • A17 Pro chip: ~17 TFLOPS (for GPU compute)
    • RAM: 8GB
    • Memory bandwidth: ~50-60 GB/s
    • Storage: Up to 1TB
    • Limited cooling capabilities
    • Not designed for sustained high-performance computing

    Going Back in Time

    Early 2010s (~2010-2013)

    Your iPhone would struggle against top supercomputers like:

    • Tianhe-1A (2010): 2.57 PFLOPS, 262 TB memory
    • K computer (2011): 10.51 PFLOPS, 1.4 PB memory
    • Titan (2012): 17.59 PFLOPS, 710 TB memory

    These systems had vastly more memory and were designed for sustained performance on large-scale simulations.

    Mid-2000s (~2005-2008)

    • BlueGene/L (2005): 136.8 TFLOPS, 32 TB memory
    • Jaguar (2008): 1.64 PFLOPS, 362 TB memory

    Your iPhone has comparable or better raw FLOPS than BlueGene/L but lacks the memory capacity and bandwidth for the actual workloads.

    Early 2000s (~2000-2004)

    • ASCI White (2000): 7.2 TFLOPS, 6 TB memory
    • Earth Simulator (2002): 35.9 TFLOPS, 10 TB memory

    Your iPhone has better raw compute than ASCI White but still couldn't handle the workloads due to memory limitations.

    Late 1990s (~1997-1999)

    • ASCI Red (1997): 1.3 TFLOPS, 1.2 TB memory
    • IBM ASCI Blue (1998): 3.9 TFLOPS, 2.5 TB memory

    Your iPhone exceeds these in raw compute but still lacks sufficient memory for their workloads.

    Mid-1990s (~1993-1996)

    • Numerical Wind Tunnel (1993): 124.5 GFLOPS, 8 GB memory
    • Intel Paragon XP/S (1994): 143.4 GFLOPS, 5 GB memory
    • Hitachi SR2201 (1996): 600 GFLOPS, 256 GB memory

    Here we're getting closer, but memory is still an issue for many workloads.

    Early 1990s (~1990-1992)

    • Thinking Machines CM-2 (1990): ~25 GFLOPS, 4 GB memory
    • Fujitsu Numerical Wind Tunnel (1992): 124 GFLOPS, 2 GB memory

    Conclusion

    You would need to go back to approximately 1991-1993 before your iPhone could genuinely outperform the top supercomputers on their actual workloads. At this point, supercomputers had:

    • Performance: 10-100 GFLOPS
    • Memory: 1-8 GB
    • Memory bandwidth comparable to modern phones

    This is when your iPhone could not only match the raw compute but also have sufficient memory and bandwidth to handle the supercomputing workloads of that era.

Responses were produced in different, uncontrolled contexts (system prompts, settings, prior turns). This is not a benchmark.