While running Spider-Man Miles Morales my Colorful H610M-K M.2 V20 stress test bottleneck quantification feels inaccurate should I export the benchmark tool report to evaluate performance

Stress test starts with decent single core performance but multi-core weaknesses show up fast. At first the benchmark tool curves jumped around and I thought it was cooling related. After exporting the report comparison revealed the main bottleneck in memory bandwidth stabilizing frames from 45-72 range down to 55-65 with much clearer curves. Honestly during multiple verification runs I hit small consistency deviations and only after rerunning scores did results become reliable reaching mid to high level overall performance. Still in high frame scenarios the multi-core shortfall remains obvious and community players rationally analyze similar cases so combining historical curves shows the optimization direction is correct yet fully eliminating bottlenecks still requires hardware upgrades in tandem.