RTX 6000 is significantly faster than a P100 outside of FP64, and is the fastest or 2nd fastest GPU outside of FP64 work [1] (the GV100 is sometimes faster, sometimes slower than the RTX 6000 but costs more). For FP64, GV100 based GPUS are quite a bit faster than P100s.
Also, you should really ignore pretty much all of the comparison sites that show up when you search for computer component comparisons as they're nearly all awful. The one you posted doesn't show a single benchmark comparison between them, and compares numbers like clock speed which isn't comparable between architectures, or memory clock speed instead of memory bandwidth leading to the laughable conclusion that the RTX 6000 has "9.9x more memory clock speed: 14000 MHz vs 1408 MHz" vs the P100 when the P100 uses HBM2 and has 732.2 GB/s vs 672.0 GB/s of actual memory bandwidth.
Also seems to be a lot cheaper than AWS counterpart.