Good question. I had to modify my code to run more efficiently, since not throttling implies that the copper block reaches a steady state with very little temperature changes over time.

But, with the changes, I can say that there is no copper block length that would prevent throttling with a 120 W CPU. It seems the heat transfer within the block is slow enough over such long lengths that you get diminishing returns with longer and longer copper blocks. Here's a graph I made summarizing the different block lengths that I tested

With a 65 W CPU, a 32 cm (double the original length) copper block is sufficient to prevent throttling, but it'll reach steady state at 97 C