F5F Stay Refreshed Hardware Desktop Comparison of Alder Lake's performance versus efficiency cores (Geekbench 5)

Comparison of Alder Lake's performance versus efficiency cores (Geekbench 5)

Comparison of Alder Lake's performance versus efficiency cores (Geekbench 5)

_
_Dirty_
Member
Find
163
05-27-2017, 09:42 AM
#1
I explored the differences between Intel’s 12th generation efficiency cores and performance cores by running Geekbench 5 tests on each. I picked Geekbench since it reflects real-world usage patterns rather than a generic synthetic workload. Below are the key findings for the i7-12700h in an Asus Vivobook S 14X:

- Core #0 serves as the baseline.
- Cores #1 through #5 show slight variations, with core #5 outperforming the others by about 4%.
- Cores #12 to #19 represent the efficiency cores; their average is notably lower than the performance cores.
- Core #4 and #5 stand out as the fastest within their group, indicating they’re the “favored” options in this chipset.
- The overall improvement from baseline to core #15 is around 3%, while core #18 is very close to baseline at 61.95%.
- The data suggests that tuning or workload selection can significantly impact performance differences between these core types.

For those interested in replicating the test, here’s a link to my GitHub repo with the script and Linux setup: https://github.com/horshack-dpreview/Gee...ationTools
Reply
Reply
_
_Dirty_
05-27-2017, 09:42 AM #1

I explored the differences between Intel’s 12th generation efficiency cores and performance cores by running Geekbench 5 tests on each. I picked Geekbench since it reflects real-world usage patterns rather than a generic synthetic workload. Below are the key findings for the i7-12700h in an Asus Vivobook S 14X:

- Core #0 serves as the baseline.
- Cores #1 through #5 show slight variations, with core #5 outperforming the others by about 4%.
- Cores #12 to #19 represent the efficiency cores; their average is notably lower than the performance cores.
- Core #4 and #5 stand out as the fastest within their group, indicating they’re the “favored” options in this chipset.
- The overall improvement from baseline to core #15 is around 3%, while core #18 is very close to baseline at 61.95%.
- The data suggests that tuning or workload selection can significantly impact performance differences between these core types.

For those interested in replicating the test, here’s a link to my GitHub repo with the script and Linux setup: https://github.com/horshack-dpreview/Gee...ationTools

Reply
Reply