The AI hardware conversation has changed dramatically over the past year.
中文摘要
AI硬件评估方式过去一年剧变。传统上，GPU比较TFLOPS、内存和跑分，但现在讨论重心已明显不同。
English Summary
AI hardware evaluation has shifted dramatically. Historically, GPUs were assessed by TFLOPS, memory, and benchmarks, but this conversation has significantly evolved.
Original Excerpt
For a long time, evaluating GPUs meant comparing TFLOPS, memory size, and benchmark scores. Continue reading on Stackademic »