F5F Stay Refreshed Hardware Desktop Comparison of relative IPC and performance of certain Intel processors in Cinebench R15

Comparison of relative IPC and performance of certain Intel processors in Cinebench R15

Comparison of relative IPC and performance of certain Intel processors in Cinebench R15

M
McKinley2
Junior Member
Find
15
07-02-2018, 05:21 AM
#1
This pre-evaluation aimed at a broader IPC assessment. I executed Cinebench R15 across multiple setups on three different CPU models. The 12100F (Alder Lake, Intel 7) at 11700k, the 7980XE (Skylake-X), and the Rocket Lake (14nm) were among the first desktop chips built on 14nm technology, marking a shift beyond Skylake. I chose Cinebench R15 mainly because it is widely recognized and tends to ignore AVX instructions, simplifying the analysis. It generally performs well when memory speed isn't a major factor. I recorded scores, power usage, and clock speeds where feasible. In some low-power tests, clocks fluctuated too much to yield reliable results. The Rocket Lake should closely match Skylake and its successors, though it isn't a perfect replacement. This was a quick check; I didn’t fully control all variables or ensure repeatable outcomes beyond running each test twice. The relative performance (IPC) varied significantly: Skylake-X scored 52.2 points per core per GHz, Rocket Lake 63.3, Alder Lake 72.7, and the newer models improved further. Efficiency metrics showed substantial differences—Rocket Lake and Alder Lake outperformed Rocket Lake across power levels. The data suggests newer nodes deliver better efficiency, especially at lower frequencies. However, the results are limited to Cinebench R15 and a narrow set of configurations. I plan to expand this evaluation to more diverse tasks in the future.
Reply
Reply
M
McKinley2
07-02-2018, 05:21 AM #1

This pre-evaluation aimed at a broader IPC assessment. I executed Cinebench R15 across multiple setups on three different CPU models. The 12100F (Alder Lake, Intel 7) at 11700k, the 7980XE (Skylake-X), and the Rocket Lake (14nm) were among the first desktop chips built on 14nm technology, marking a shift beyond Skylake. I chose Cinebench R15 mainly because it is widely recognized and tends to ignore AVX instructions, simplifying the analysis. It generally performs well when memory speed isn't a major factor. I recorded scores, power usage, and clock speeds where feasible. In some low-power tests, clocks fluctuated too much to yield reliable results. The Rocket Lake should closely match Skylake and its successors, though it isn't a perfect replacement. This was a quick check; I didn’t fully control all variables or ensure repeatable outcomes beyond running each test twice. The relative performance (IPC) varied significantly: Skylake-X scored 52.2 points per core per GHz, Rocket Lake 63.3, Alder Lake 72.7, and the newer models improved further. Efficiency metrics showed substantial differences—Rocket Lake and Alder Lake outperformed Rocket Lake across power levels. The data suggests newer nodes deliver better efficiency, especially at lower frequencies. However, the results are limited to Cinebench R15 and a narrow set of configurations. I plan to expand this evaluation to more diverse tasks in the future.

Reply
Reply