UPDATE readme with section about OEM "optimizations" and a guide to do a sanity check - #2
UPDATE readme with section about OEM "optimizations" and a guide to do a sanity check#2tfuerholzer wants to merge 5 commits into
Conversation
mbestavros
left a comment
mbestavros
left a comment
There was a problem hiding this comment.
Thanks for the contribution @tfuerholzer! It's much appreciated!
I have two thoughts here:
-
While this may be useful information, I think it would fit better as part of a new "troubleshooting" section rather than as part of the standard setup docs. E.g. "if you see abnormally low Speedometer scores, consider checking whether throttling is occurring". Would you mind reformatting it as such?
-
Even with that said, the point of the
browser-power-hourbenchmark is to figure out relative power consumption of different browsers. If they're both tested under throttling conditions, that may still be a meaningful result. Though, again, I agree we should document the possibility of throttling and how to fix it.
Besides those comments, some minor structure/formatting suggestions.
|
|
||
| | Soc | Primary CPU | OctaneJS Chromium | OctaneJS Firefox | | ||
| |:--------------:|:------------------------------:|:-----------------:|:----------------:| | ||
| | Snapdragon 810 | Cortex-A57 @ 2 GHz | ~ 9500 | ~ 5400 | | ||
| | Snapdragon 820 | Kryo @ 2.15 GHz | ~ 11000 | ~ 6800 | | ||
| | Snapdragon 835 | Modified Cortex-A73 @ 2.45 GHz | ~ 14500 | ~ 8700 | | ||
| | Snapdragon 845 | Modified Cortex-A75 @ 2.8 GHz | ~ 23000 | ~ 13000 | | ||
| | Snapdragon 855 | Modified Cortex-A76 @ 2.84 Ghz | ~ 32000 | ~ 19000 | | ||
| | Snapdragon 865 | Cortex-A77 @ 2.84 GHz | ~ 41000 | ~ 22000 | | ||
| | Exynos 7420 | Cortex-A57 @ 2.1 GHz | ~ 10800 | ~ 5300 | | ||
| | Exynos 8890 | Samsung M1 @ 2.6 Ghz | ~ 13000 | ~ 8500 | | ||
| | Exynos 8895 | Samsung M2 @ 2.3 Ghz | ~ 17000 | ~ 9300 | | ||
| | Exynos 9810 | Samsung M3 @ 2.9 Ghz | ~ 24000 | ~ 12000 | | ||
| | Exynos 9820 | Samsung M4 @ 2.7 Ghz | ~ 33000 | ~ 18000 | | ||
| | Kirin 955 | Cortex A-72 @ 2.55 Ghz | ~ 14500 | ~ 8000 | | ||
| | Kirin 960 | Cortex A-73 @ 2.3GHz | ~ 13500 | ~ 8000 | | ||
| | Kirin 970 | Cortex A-73 @ 2.3GHz | ~ 14000 | ~ 8500 | | ||
| | Kirin 980 | Cortex A-76 @ 2.6 Ghz | ~ 31000 | ~ 17000 | | ||
| | MTK Helio X20 | Cortex A-72 @ 2.3 Ghz | ~ 11000 | ~ 6500 | | ||
|
|
There was a problem hiding this comment.
Consider wrapping this in spoiler text to avoid an overlong README.
Co-authored-by: Mark Bestavros <markbest@bu.edu>
Co-authored-by: Mark Bestavros <markbest@bu.edu>
Co-authored-by: Mark Bestavros <markbest@bu.edu>
|
i added the spoiler to the readme, sorry it took some time i had a very busy week. |
2 participants
Adds a section to the readme informing readers about past shenanigans especially when it comes to browsers and how to check for potential artificial limiting of browser performance.
The results in the table are all from testing on my devices (I have a collection of about ~ 120 devices for testing devices, mostly older flagship phones). For all of the SoCs I have at least two or more phones that produce these results. In my case I ran brave (Chromium 125 and Firefox 127), however the results are the same for all other Firefox or chromium browsers, given that none of the JS and WASM related flags are touched.
The data for the SoCs is sourced from phonedb.net.