https://news.ycombinator.com/item?id=39518293 :
> [PQ] SSL/TLS Accelerators and ML Accelerators / AI Accelerators
An opportunity for a new RISC-V arch, perhaps.
> With some newer architectures, the GPU(s) are directly connected to [HBM] RAM; which somewhat eliminates the CPU and Bus performance bottlenecks that ASICs and FPGAs are used to accelerate beyond.
The HBM3E Wikipedia article says 1.2TB/s.
Latest PCIe 7 x16 says 512 GB/s: https://news.ycombinator.com/item?id=39933959