I don't believe that such a table exists nor do I think it is possible to generate one. The issue is that, in modern processor designs, with all of the various pipelines supported, the instructions per cycle is dependent upon the mix of instructions currently being executed.
I was trying to calculate the GFLOPs to find a method to evaluate other CPUs in relation to empirical tests done by me with processors of my property. I found a significant relationship with PassMark's benchamarks but not all CPUs are evaluated. Missing the Xeon PHI and the higher score stops at 29315 points for the Intel Core I9 9980 XE.Standing the Passmark scores I should reach a score of 57000 with a single CPU. For this I wanted to transfer everything in the calculation of the GFLOP hoping that the relationships for differences I calculated empirically remain valid . Do you have any suggestions?