The carbon footprint of GPT-4

Let’s first estimate GPT-4's energy consumption. According to unverified information leaks, GPT-4 was trained on about 25,000 Nvidia A100 GPUs for 90–100 days [2]. Let’s assume the GPUs were installed in Nvidia HGX servers which can host 8 GPUs each, meaning 25,000 / 8 = 3,125 servers were needed. One way to estimate the electricity consumption from this information, is to consider the thermal design power (TDP) of an Nvidia HGX server. TDP, denoted in watts, expresses the power consumption of a piece of hardware under maximum theoretical load [11], ie the actual power consumption may differ. <a href="https://towardsdatascience.com/the-carbon-footprint-of-gpt-4-d6c676eb21ae">Visit Now</a>