Mere 3% is hardly unlikely. That's what AMD can gain from increasing memory clock from 2933MHz to 3200MHz and enabling bit tighter timings. Even from macroscopic design point of view, AMD did huge design changes. We do not really know where that lands chip in terms of memory access, cache latencies. But I would say that AMD aims to improve. What AMD did with cores themselves? Spoiler: Following is taken from Techreport. Zen 2 addresses this shortcoming by doubling each core's SIMD register width to 256 bits. The floating-point side of the Zen 2 core has two 256-bit floating-point add units and two floating-point multiply units that can presumably be yoked together to perform two fused multiply-add operations simultaneously. ... To feed those 256-bit-wide execution engines, AMD also widened the load-store unit, load data path, and floating-point register file to support 256-bit chunks of data. Source. I would expect around 10% IPC gain on average (depending on workload). But 15% is not out of question and some workloads may do even better.