What is the documented speed boost range for MMX and SSE2?

• A. 5% to 50%
• B. 10% to 400%
• C. 50% to 200%
• D. 0% to 10%

Answer: (B)

SIMD acceleration is what MMX and SSE2 enable: processing multiple data points in parallel within a single instruction to boost throughput. The documented speed-up range reflects how much you can gain depending on how well your task maps to vector operations and how memory access behaves. For workloads that are highly vectorizable and keep data in fast memory, you can approach a fourfold improvement (about 400%). In many real scenarios, gains are smaller because only a portion of the code is vectorizable and overhead or memory bandwidth limits performance, so you might see around 10% or more improvement. So the documented range is roughly 10% to 400%, depending on the workload and implementation.