Intel Core Ultra 155H Meteor Lake CPU Specs Seemingly Confirmed In Early Benchmarks
A highly detailed Meteor Lake benchmark listing has made an appearance in PassMark's database featuring a new CPU called the Core Ultra 7 155H. The listing features unusually-detailed CPU specifications, suggesting that the chip’s specs could be legitimate. Unfortunately, the chip did not perform particularly well in the benchmark itself, delivering a score that barely outperformed Intel’s old 11th-Gen Tiger Lake CPUs.
As previously stated, the CPU specifications listed are highly specific, more so than usual. According to PassMark the Core Ultra 7 155H sports 16 cores in total, featuring six P-cores with Hyper-Threading along with ten E-cores. The chip features 960KB of L1 cache, 18MB of L2 cache, and a 24MB L3 cache, with a 3.8GHz base clock and 4.9GHz turbo frequency on the P-cores.
Performance-wise, the chip scored 21359 points in PassMark’s test, with a single-threaded rating of 3897. The closest CPU in the PassMark database that approaches the 155H’s score is AMD's Ryzen 7 5800H, a Zen 3 mobile SKU with 8 cores, featuring a score of 21,171 points — 188 points short of the Meteor Lake chip. As a result, the performance margin between the two chips in this synthetic benchmark is basically negligible, with the Core Ultra 7 155H outperforming the Ryzen 7 5800H by a measly 0.9%.
The most important aspects of the PassMark listing are the CPU specifications, which look like they could be legitimate. The use of ten E-cores specifically correlates well with some of the architectural decisions found in Intel’s Meteor Lake. With Meteor Lake, Intel is implementing a couple of low-power E-cores on the SoC tile of Meteor Lake chips to improve power consumption at idle and on very light workloads.
This benchmark result is the opposite of what we saw with the Core Ultra 7 165H. That leak showed strong performance in Geekbench against some of the best mobile CPUs around, like the Ryzen 9 7940HS. However, as with all leaked benchmark runs, we have to take these results with a grain of salt. The performance of pre-production chips can vary drastically compared to their retail counterparts. Plus, summarizing a processor’s performance entirely from a single benchmark is a fool's errand, as chip performance can vary drastically depending on the benchmark in question.
