[mntmoss]

For most programming tasks I would focus on the base of the pyramid: Make the storage low-latency for small files, then up the memory size and bandwidth, and then use a CPU appropriate to the workload(ideally, it can go wide and parallelize - otherwise you're back to single-threaded perf). This mitigates the worst case of poorly optimized builds that need to frequently return to storage, and it improves all factors of the operating system when it chooses to swap to virtual memory(Windows has become a very aggressive swapper since Win10 launched and they added a new page file system on top of the old virtual memory).

You can spend time gazing at the benchmarks for each CPU, but I would not pinpoint it as the bottleneck for a responsive and pleasant programming environment. The new AMD chips are good all around. The new Intels are still OK, but the top end probably too hot and loud to reccommend.

For gaming, high clocks/high single-threaded IPC remain the primary factors. Games are mostly designed towards a certain number of cores and speed of I/O. Fast disk and memory will reduce sources of stutter but this is dependent on how often the game tries to load something.