I think "throughput" is a better word for what most people think of as IPC these days. Back when CPU's were much, much simpler it was easier to identify IPC and even then it was a theoretical construct as you have to make some assumptions as to what is happening as data is processed.
"Throughput" on the other hand doesn't care about what is going on inside the CPU. Maybe "Iso-Throughput" is a better term. Set a frequency, run an application/benchmark and see how much work gets done, compare that to another CPU running at the same frequency and there is a measure of how well a those CPU's convert input to output at the same frequency. "Efficiency of Computation" could also be used, "EoC." Or perhaps "Computation Work Rate," which would be score divided by frequency, "CWR."
The point is forget about trying to relate this to what is going on inside the CPU these days, both the CPU and applications are too complex to nail that down enough for it to be meaningful.
Or perhaps change "Instructions per Cycle" to "Computation Output per Cycle." "COpC" Nah, too much like COPD.
"Throughput" on the other hand doesn't care about what is going on inside the CPU. Maybe "Iso-Throughput" is a better term. Set a frequency, run an application/benchmark and see how much work gets done, compare that to another CPU running at the same frequency and there is a measure of how well a those CPU's convert input to output at the same frequency. "Efficiency of Computation" could also be used, "EoC." Or perhaps "Computation Work Rate," which would be score divided by frequency, "CWR."
The point is forget about trying to relate this to what is going on inside the CPU these days, both the CPU and applications are too complex to nail that down enough for it to be meaningful.
Or perhaps change "Instructions per Cycle" to "Computation Output per Cycle." "COpC" Nah, too much like COPD.