These days AMD next-generation processor architecture Zen3 news a lot, especially IPC performance caused players’ concern, optimistic news that IPC performance improved by 17%, AMD CTO revealed is not less than 10%. In addition to the architecture upgrades, the Zen3 also uses the 7nm plus (N7P) process of On-Stage Cladding, which also improves performance.
The current TSMC to AMD’s Zen2 architecture contract is 7nm HPC process (N7), or the first generation 7nm, and Zen3 an important upgrade on the process, evolved to the second generation of N7P, that is, 7nm plus EUV process, the current Huawei’s Kirin 990 The 5G processor uses tSMC 7nm EUV, but this is for low-power products, and AMD should use the HPC version.
For the second generation 7nm process, TSMC officially giveless information, previously indicated that the core area of the 7nm-EUV process will be reduced by 10-20%, the same performance of power consumption reduced by 10%, but the performance improvement is not mentioned – considering that the EUV photolithography machine is only the production process, the transistor structure has not changed, There should be no change in process performance.
However, a 10% reduction in power consumption at the same performance means that AMD is still likely to increase the frequency of the Zen3 processor under the same TDP, and previous reports have shown that the Zen3 frequency can be increased by 100-300MHz, and if it is 300MHz, then the acceleration frequency should be promising to reach 5GHz. After all, the Ryzen 9 3950X has reached 4.7GHz.
For AMD’s performance improvement, Intel’s side is not likely to be, the 10th generation Core Comet Lake-S processor is expected to be launched in the coming week, Core i9-10900K with the new Turbo Acceleration mode, the maximum frequency up to 5.3GHz, is expected to maintain the frequency advantage of AMD. But given Intel’s same architecture, amd’s side should be able to achieve a single-core advantage reversal by even 10% increase on the AMD side.