Although it might sound like a long time ago, the Tesla architecture marked a pivotal moment in GPU development, shaping the GPUs we use today.
🔥 The main breakthrough: the introduction of the 𝗚𝗲𝗻𝗲𝗿𝗶𝗰 𝗦𝘁𝗿𝗲𝗮𝗺𝗶𝗻𝗴 𝗠𝘂𝗹𝘁𝗶𝗽𝗿𝗼𝗰𝗲𝘀𝘀𝗼𝗿 (𝗦𝗠), capable of performing any type of work — not just graphics commands, like older GPUs. With the help of their new SM design Nvidia GPUs allowed programmers to use the GPU more as a general-purpose processor, moving beyond the traditional graphics concepts.
🌟 At the same time, NVIDIA introduced CUDA, their programming language designed for NVIDIA GPUs. With the release of CUDA & the Streaming Multiprocessor, NVIDIA aimed to make 𝗚𝗣𝗨 𝗽𝗿𝗼𝗴𝗿𝗮𝗺𝗺𝗶𝗻𝗴 𝗮𝗰𝗰𝗲𝘀𝘀𝗶𝗯𝗹𝗲 𝘁𝗼 𝗮𝗹𝗹 𝗱𝗲𝘃𝗲𝗹𝗼𝗽𝗲𝗿𝘀—not just graphics experts.
⚠️ It’s important to understand Tesla because everything that followed has been an evolution in the same direction: making GPUs increasingly powerful general-purpose processors.