Guest ippoP Posted February 10, 2010 Report Share Posted February 10, 2010 (edited) Hi all. I've run the Softweg Benchmark v1.03 on a few Hero custom ROMs. Namely (in no particular order): - Andro-Plus v1.1 - CM Hero v1.0 - AychTeeSee v3.0.1 - MoDaCo v3.1 - Villain v2.1 - Vanillain v1.0 - KaguDroid v1.2.1 For every ROM I've followed this "protocol": 0. attached the USB power chord and wait for full charge 1. downloaded the zip file(s) (for a few I had to download also the Google stuff) 2. wipe + flash + reboot (only 1 reboot, no other recovery image action) 3. put the screen blanking to 10 minutes (no other setting touched) 4. downloaded from the market the benchmark tool 5. run it 3 times but only got the last run results (*) 6. put all data on a spreadsheet to ease the comparison I'm not sure about the FS test, but I am using an 8GB, class4 uSDHC. The results are attached here and I leave the considerations to everyone, as mine are biased by things like personal taste and attitude (maybe as almost every one else). Red is worst score, green is best. I would like to encourage everyone to repeat these benchmarks and, possibly, to enhance the testing protocol (especially point #3, I presume). Have a nice day with your Android! (*) I read the numbers from the application and not from the CSV file. I had some troubles with file caching over USB memory access! Sorry, but can only upload PDFs!heroms_02.pdf Edited February 11, 2010 by ippoP Link to comment Share on other sites More sharing options...
Guest pcs333 Posted February 10, 2010 Report Share Posted February 10, 2010 would be interesting to see a kagudroid result on there as well, because most people seem to get around 270 on the cpu result. Link to comment Share on other sites More sharing options...
Guest ippoP Posted February 11, 2010 Report Share Posted February 11, 2010 (edited) would be interesting to see a kagudroid result on there as well, because most people seem to get around 270 on the cpu result. Kagudroid reboots quite often so I don't even took it into account. You can follow the "protocol" above and post the results. Maybe in the next weekend I'll get some time to do it, despite I think that a score of 270 to the CPU seems to be unlikely. [Update: I got 281 with CPU bencmarking on KaguDroid. I presume it's because of some main change in the kernel and/or libraries] Edited February 11, 2010 by ippoP Link to comment Share on other sites More sharing options...
Guest ippoP Posted February 11, 2010 Report Share Posted February 11, 2010 Added KaguDroid! Go back to post #1. Link to comment Share on other sites More sharing options...
Guest Villarous Posted February 11, 2010 Report Share Posted February 11, 2010 What does benchmark do? Higher benchmark equals less lag or something? Link to comment Share on other sites More sharing options...
Guest ippoP Posted February 11, 2010 Report Share Posted February 11, 2010 What does benchmark do? Higher benchmark equals less lag or something? Very likely, yes: less lags. For the actual benchmark meaning please refer to the Softweg website. Link to comment Share on other sites More sharing options...
Guest pcs333 Posted February 11, 2010 Report Share Posted February 11, 2010 Kagudroid reboots quite often so I don't even took it into account. You can follow the "protocol" above and post the results. Maybe in the next weekend I'll get some time to do it, despite I think that a score of 270 to the CPU seems to be unlikely. [Update: I got 281 with CPU bencmarking on KaguDroid. I presume it's because of some main change in the kernel and/or libraries] i dont get any reboots with kagudroid, i find it just as stable as any other 2.1 release. dont understand why there is such a difference in the benchmarks though because it doesn`t really feel any quicker to me, perhaps someone has a theory on this? Link to comment Share on other sites More sharing options...
Guest ippoP Posted February 11, 2010 Report Share Posted February 11, 2010 Benchmarks are just that: benchmarks, not real life performance measurements. In a very general view, the higher the scores, the better the "user experience". A better compiler can produce faster code. Then maybe there is (much) less debugging code enabled into a ROM than in another. Finally, a newer kernel and hardware drivers can lead to shorter execution times. In the linux world it happens that newer kernels will lead to higher performances, which is not the case for other commercial operating systems ... ;) Link to comment Share on other sites More sharing options...
Guest Villarous Posted February 11, 2010 Report Share Posted February 11, 2010 So from these numbers, it seems like Kagudroid is running smoothest of them? Link to comment Share on other sites More sharing options...
Guest traffic1986 Posted February 11, 2010 Report Share Posted February 11, 2010 I would add to this statement CyanogenMod for HTC Hero Link to comment Share on other sites More sharing options...
Guest ippoP Posted February 11, 2010 Report Share Posted February 11, 2010 I would add to this statement CyanogenMod for HTC Hero CM hero == Cyanogen Mod for Hero Link to comment Share on other sites More sharing options...
Guest ippoP Posted February 11, 2010 Report Share Posted February 11, 2010 So from these numbers, it seems like Kagudroid is running smoothest of them? Not really. To my personal perception. KaguDroid runs smooth, but MoDaCo is smoother. But then maybe the first displays web pages and pictures faster than the second ... it'd not easy to tell from just the scores. What I can say is that Andro-Plus seems to me actually the worse one ... Link to comment Share on other sites More sharing options...
Recommended Posts
Please sign in to comment
You will be able to leave a comment after signing in
Sign In Now