cpu bandwidth测试结果无法对齐

如果按照原先方式：
std::unique_ptr src{new uint8_t[NR_BYTES]{}};
std::unique_ptr dst{new uint8_t[NR_BYTES]{}};
此时得到的带宽偏低，在8295上可能才14 GBps
如果改为：
uint8_t* src = (uint8_t*)malloc(NR_BYTES * sizeof(uint8_t));
uint8_t* dst = (uint8_t*)malloc(NR_BYTES * sizeof(uint8_t));
得到的带宽接近36 GBps。
测试设备 SA8295大核。

The issue names no file or test. Reproduce the CPU bandwidth benchmark on the SA8295 using both std::unique_ptr with new[] and malloc, then inspect how allocation and initialization affect the measurements. Done means the discrepancy has a documented cause and the benchmark results are made consistent or the limitation is clearly recorded.

Written by the indexing model from the issue text.