You're comparing GPU and CPU architectures from a deep-learning perspective, and you want to state the most accurate distinction. Which statement best captures the architectural advantage of GPUs over CPUs for AI workloads?

The GPU vs. CPU distinction for AI comes down to architecture, not just performance metrics. CPUs are designed for sequential execution with complex instruction sets. They optimize for single-threaded performance and out-of-order execution. GPUs are designed for parallel execution with simpler instruction sets. They optimize for throughput across thousands of operations simultaneously. Deep learning consists of massive matrix operations that are highly parallel. Matrix multiplication of two 10,000×10,000 matrices involves 10^12 floating-point operations that can happen in parallel. A GPU can split these across its thousands of cores and compute them in parallel. A CPU with 8 cores has to serialize — execute them sequentially, which is orders of magnitude slower. This architectural difference is why GPUs are transformative for AI. The claim about significantly higher bandwidth main memory access, enabling efficient processing of large datasets, is a real secondary advantage: GPUs typically have higher memory bandwidth than CPUs, but that does not capture the primary architectural distinction. The claim about comparable core counts but significantly higher GPU clock speeds is false: GPUs have far more cores, and GPU and CPU clock speeds are both in the GHz range. The claim about identical architectures interchangeable for any workload is false because the instruction-set design, core count, and execution model differ fundamentally. The question drives home the fundamental architectural insight.