I really appreciate you coming here to discuss your addon and to clarify certain points. I’m so happy you are updating it and look forward to be able to test more platforms to add to the ever growing list. Will be keeping an eye out for when i’s released
so the update (if things go to plan isn’ far away)
From my understanding this test is encouraged to run on different platforms, since it’s doing the same thing you are measuring against different platforms, so you can run on all of them really and see any differences, I doubt we’ll have differences between NG NE NO, but who knows
Look forward to any testing and reporting you can provide, anyone else also welcome to join with the testing.
Update Lunatixz released an update to the CPU Benchmark Addon, it now does several passes (8 on mine, 4 on lower Speced hardware) then gives you a final score.
So far it’s working on all platforms so Lunatixz will be updating the one in official kodi repo. I have 57 datapoints now, with notable mentions
Dune HD Pro 8K Plus (Amlogic S928X-KJ) - 190,242
Ugoos SK1 - 190,200
N2+ non-oc - 102,908
Ugoos AM6B Plus (Amlogic S922XJ) - 98,548
Odroid N2+ - 96,254
Homatics Box R 4K Plus - 57,778
Homatics Box R 4k Plus - 56,380
Banana Pi M5 (CoreELEC) - 43,484
Homatics Box R 4K Plus (Android TV 12) - 38,034
still many boxes I’d like to have data on, If anyone wants to help out, happy to take results. or if you just wanted to see you get the same results you can do so, the addon is in the kodi repo under programmes.
I plan on making a google sheet and giving write access to certain members that can carry this on if another member stopped, the list can keep growing and we can have some kind of measure to measure the python single core speed for addons
thanks
I’ve exported the data to google sheets, kept it simple
I’ve made the AM6B+ as 100% because it’s the most popular hardware from this survey
and it plays the most from any device out their.
So if someone got an N100, yes it’s faster and it’ll show you potentially how much for python single threaded addons but it’s missing ALOT , I’m waiting for
@ Lunatixz updated addon to get accepted in kodi repo so I’ll ask the reddit community to list their boxes, hoping to get 100 datapoint, I’ve got 67 so far. allot of the most popular boxes are their
the addon is much improved now and shows you more data (thanks Lunatixz)
these results are very consistent, it’s almost idenical to the Homatics R 4k Plus (which basically has the same SOC)
very happy with this and it should serve well to give some idea of how a performance of majority of addons that rely on single threaded python performace will behave in relation to other boxes. My list is on post above. more results are welcome thanks
it’s updated in official repo freddy
once you are able to run the tests, let me know your gmail so I can add you in the editor lists , you can add any results you get now and in the future to the table.
all it took was 4 days. lots of varied data. whose going to be the 100th? how many boxes left? people can see data from TV’s , high end smartphones, low end smartphone , all sorts of OS.
It even helped a vendor figure out that the python version of 3.8 is leaving 50% of python performance compared to boxes with the same SOC that has python 3.11
this is more than I could have hoped for. maybe it will push other vendors who are still on older version of python
You’re running a slightly higher python version in kodi 22 alpha, you got slightly higher results than another user whose running python 3.11
but they are in line.
The TOX3 having the same SOC as the homatics is nice, so you pay $40 less (half price) without DV
If you have any other boxes you’d like to test, I’d be happy to add . thanks
Back in North America
I allready had this addon installed, and it updated itself to lastest version when CoreELEC updated to latest Nightly.
Did some quick tests on X96X10 Pro 8/64
Changing skins varied results
I also changed a setting in CoreELEC Hardware
CPU performance from Performance to On Demand ,which also varied results.
Both were not dramatic differences, but differences none the less.
I think i like the older version percentage output that was given exp: Top % 135 ,as opposed to score ranking
I appreciate all the work you have put into this.
However i use very small amount of addon on my CoreELEC boxes, which probably cant list.Media Playback is number 1 factor why a box is chosen.
In my opinion this is more Kodi user across all platforms benefit. I like to set up Boxes for Friends/Family, after it leaves my place many users install lots of addons ,some even install Custom Builds on top of CoreELEC.
I will let them know about this Excellent work/ addon matrix that may be a factor in there next purchase.
If you still want results from my boxes to add Data let me know
welcome back, thanks for doing the testing. ht2tweak has done so much work on that spreadsheet, he even has his YouTube where he reviews stuff. I don’t think their is a better place to get recommendations on hardware.
The Top %% was not what people thought it was. it’s just comparing to the SOC that the addon developer used at the time, which was very powerful.
the i5-4670K, since no one is going to use that going forward, it made no sense. I can easily know what the top%% is for any hardware by using the ratio.
It’s a funny coincidence that one of the best boxes for Kodi, (AM6B+) measure almost exact 100k
so if someone had a score of 190k , I know it’s “TOP 188%” without doing anything (if we compared to the AM6B+, and we should since it’s the best)
their is so many nice patterns, I think if you don’t care about DV or HDR10+, CEC and wake from Lan , the N100 platform is the best, the scores are ridiculous 331k, and on sales this can be had for $89 (to UK).
The best value seems to be something like Vontar X4, I have results for Tox3, waiting for someone who can provide one for Vontar X4, it can be had for like $43
I would love to have any results you have. please list them here and I’m happy it will help others in the decision making of new boxes.
I’m asking for which skin someone used, which OS and the paste log if possible THX
I will probably fill in some holes for some Rockchip devices if you don’t get any responses over time but no point in testing when the relative order is pretty much known. As you know I consider this unscientific benchmark a waste of effort and there haven’t even been any surprises. As an example It won’t make Rockchip devices more attractive as options to Amlogic.
