This comment thread is ridiculous. Can’t we have open discussion around the article rather than back and forth discussion obsessing over the minutiae of what’s a better deal for x, y or z.
the first few posts were clearly trolls - I think we all agree gaming results are not a mess on windows, and has lots of room to grow on Linux.
we should also acknowledge that AM4 boards with zen 5000 support are generally cheaper and so that does somewhat offset the cheaper intel prices. Exact trade off is incredibly use case specific YMMV. but don’t kid yourself cross comparing pricing for bargain bin MB for AM4 against flagship AL boards without acknowledging the significant platform benefits of the later. Surely better comparisons will be possible in a few weeks once the cheaper platform options are launched.
Can we also agree that all has nothing to do with this article, and move on?
——
back on topic
power efficiency for 8P [+ HT] + 8E was really impressive vs 8P [HT]. Seems to validate intel's choice to go heterogeneous and that is despite the kernel lacking heterogeneous aware scheduling for AL. Should only improve going forward.
it’s too bad you can’t completely disable the P cores. Cam CPU pinning/affinity be set such that the one P core is never used by the task under test? Or was that done and I missed it?
It would also be nice to see efficiency and total power to complete tasks in addition to the perf and power plots.
seconding the request to see an AMD part added for comparison sake. But keep in mind comparing CPU consumption across brands/generations does not capture the full picture if your concern is anything other than bragging rights. You need to measure power at the wall to do that properly (ex: RAM and platform controller are significant power consumers)
the first few posts were clearly trolls - I think we all agree gaming results are not a mess on windows, and has lots of room to grow on Linux.
we should also acknowledge that AM4 boards with zen 5000 support are generally cheaper and so that does somewhat offset the cheaper intel prices. Exact trade off is incredibly use case specific YMMV. but don’t kid yourself cross comparing pricing for bargain bin MB for AM4 against flagship AL boards without acknowledging the significant platform benefits of the later. Surely better comparisons will be possible in a few weeks once the cheaper platform options are launched.
Can we also agree that all has nothing to do with this article, and move on?
——
back on topic
power efficiency for 8P [+ HT] + 8E was really impressive vs 8P [HT]. Seems to validate intel's choice to go heterogeneous and that is despite the kernel lacking heterogeneous aware scheduling for AL. Should only improve going forward.
it’s too bad you can’t completely disable the P cores. Cam CPU pinning/affinity be set such that the one P core is never used by the task under test? Or was that done and I missed it?
It would also be nice to see efficiency and total power to complete tasks in addition to the perf and power plots.
seconding the request to see an AMD part added for comparison sake. But keep in mind comparing CPU consumption across brands/generations does not capture the full picture if your concern is anything other than bragging rights. You need to measure power at the wall to do that properly (ex: RAM and platform controller are significant power consumers)
Comment