https://blog.cloudflare.com/on-the-dangers-of-intels-frequency-scaling/ https://blog.cloudflare.com/on-the-dangers-of-intels-frequen... It turns out AVX512 doe

https://blog.cloudflare.com/on-the-dangers-of-intels-frequency-scaling/ https://blog.cloudflare.com/on-the-dangers-of-intels-frequen... It turns out AVX512 doesn't give you the kind of performance speedouts you think. On those benchmarks, AVX512 is only around 8% faster than AVX2. The reality is that AMD is killing Intel at every price point. My 3950x scores 12444 (no OC) and only costs around $400. Looking up on Newegg, your 10980XE costs... wait for it.. $3000. Even with your crazy overclocking, that intel cpu is only 35% faster and 650% more expensive. How can you seriously claim that Intel is at all competitive with AMD? And speaking of ML, it's rare to find a ML workload that is expensive that doesn't involve neural networks. And for neural networks, you should be running on a GPU anyways so I'm not sure how useful this DL Boost tech is.