Skip to question
asic.fyi
DesignVerificationSystemVerilogFirmwareArchitectureInterviews
asic.fyi/Interview questions/Place a dense neural-network kernel

Hardware interview practice

Place a dense neural-network kernel

MediumComputer ArchitectureSystemVerilog

A mobile SoC has a CPU, GPU, and NPU. The workload is a supported dense neural-network inference with fixed tensor shapes and strict energy limits. Which engine is the best first placement choice?

Try it in the question bankReason first. Then compare.

Keep this exact question selected while you check your answer and review the full solution.

Practice this question →
Choose one

Answer choices

  1. A. The sensing hub, because it owns all system memory and replaces the other processors.
  2. B. The GPU, because every dense neural kernel is graphics and cannot run on an NPU.
  3. C. The NPU, because it is specialized for supported neural tensor operations at high performance per watt.
  4. D. The CPU, because general-purpose control always gives the lowest inference energy.
Continue practicing

Related questions

Computer ArchitectureForward bytes from a small pending-write queue→Computer ArchitectureMatch hit-under-miss responses by transaction ID→Computer ArchitecturePause headroom at 400 Gb/s→
asic.fyi · Learn silicon end to end.info@asic.fyi