NVIDIA GeForce GTX 1080 8GB vs NVIDIA P100 PCIe 16GB

Both GPUs launched in 2016: the GeForce GTX 1080 on Pascal and the P100 on Pascal.

Dense throughput for the GeForce GTX 1080 against the P100: FP32 8.9 vs 9.3 TFLOPS. Sparse figures, where the vendor publishes them, appear under each dense number in the performance table.

Memory is 8 GB against 16 GB, fed at 320 GB/s versus 732 GB/s. More memory per GPU means larger models fit before you have to shard across cards, which often matters more than raw TFLOPS for inference.

Power budgets are 180 W for the GeForce GTX 1080 and 250 W for the P100. At FP32 that works out to 0.05 against 0.04 TFLOPS per watt.

For LLM training and inference, weight the FP8 and FP16 rows and memory capacity most heavily. For scientific and HPC workloads, the FP64 row is the one to read.