4 ms·
If the CPU test is simply reading 1/6 of the data, then it should be memory that is the bottleneck and not the CPU (even unoptimized). Something smells very wro
by lowbloodsugar 9y ago
If the CPU test is simply reading 1/6 of the data, then it should be memory that is the bottleneck and not the CPU (even unoptimized). Something smells very wrong about his adhoc test. The next piece of data is in the cache-line seven times out of eight. And its reading 1/6 of the data into memory. Should be way faster even without -O3. And if there's something fundamentally broken about the code without -O3, then why post that benchmark at all? Seems dishonest.
Be good to post the code when making such claims.
- zlynx 9y agoI thought it made the point that if you aren't vectorizing your code then it hardly matters if you use an optimized data format. SSE makes a huge difference. With AVX and multiple threads you can actually exceed the memory bandwidth of a CPU socket, which looks funny in performance graphs as adding threads to cores suddenly stops scaling.