[MrTasselhof]

I have been asked to run some benchmarks on the Aurora at the stock 1 Ghz speed.
Below are the Results - ICS 4.0.1 - Kernel 3.0.8 - Firmware Dated: 20111229

Linpack

Single Thread
MFLOPS: 15.693
Time: 5.34 Seconds
Norm Res: 5.68
Precision: 2.220446049250313e- 16

Multi-Thread
MFLOPS: 11.553
Time: 14.6 Seconds
Norm Res: 3.24
Precision: Inconsistent Result

Quadrant Standard

Your Device: 1478

AnTuTu Benchmark

RAM: 400
CPU integer: 637
CPU float-point: 160
2D graphics: 219
3D graphics: 861
Database IO: 215
SD card write12.2 MB/s) 122
SD card read: (>50 MB/s) 202
CPU frequency: 1008 Mhz
Total score: 2816

AnTuTu Battery Tester

Ran the test with Full Brightness and WiFi on at 100% CPU/System Usage.

Battery Test Score: 603
Start: 14:38 - 100%
Part 1: 15:59 - 73%
Part 2: 17:24 - 46%
Finish: 17:59 - 19%

CF-Bench

Native Score: 3817
Java Score: 1301
Overall Score: 2307

Nenamark1

Landscape: 40.3

Nenamark2

Did not run properly

Softweg

Landscape
Graphic
Total graphics score: 295.92026
CPU
Total CPU score: 1590.811
Memory
Total memory score: 441.79044
Filesystem
Total file system score: 178.83443

Portrait
Graphic
Total graphics score: 342.50156
CPU
Total CPU score: 1699.0762
Memory
Total memory score: 383.8507
Filesystem
Total file system score: 144.42607

 

[Dansplans]

I would love to see detailed quadrant scores. The graph shows competing device scores for comparison.

Correct me if I'm wrong here but isn't the quandrant score (1478) way low? I just ran it on an A10 tablet with 512MB RAM running Gingerbread and got 1850.

I guess ICS still needs a lot of optimization. I would have thought that the 1GB RAM on Aurora would kick a lot more butt.

 

[MrTasselhof]

Just added the AnTuTu Battery Test results.

Tasselhof

 

[pinny716]

Just added the AnTuTu Battery Test results.

Tasselhof

WOW in less than 4 hours the tablet went from 100% to 19%???

 

[MrTasselhof]

WOW in less than 4 hours the tablet went from 100% to 19%???

Mind you this was with it running at full brightness, wifi, and full CPU/System usage for 4 hours solid.

That is standard to better than most of these devices...

It also has only been through a couple charge sequences, which supposedly it takes six or more to fully break in the battery.

Tasselhof

 

[pinny716]

Mind you this was with it running at full brightness, wifi, and full CPU/System usage for 4 hours solid.

That is standard to better than most of these devices...

It also has only been through a couple charge sequences, which supposedly it takes six or more to fully break in the battery.

Tasselhof

whoops take it back missed the 100% CPU usage and didn't think about the break-in time for the battery. I ordered it from pandawill already. I probably will only be using it for 6 hours total a day (and most likely not using 100% CPU). Thanks for everything. Ainol should be paying you for all the business you are getting them!

 

[WAI]

the Build/ROM that's based on this A10 Device is the exact same as the Novo7A but it's just coded for this Device that's all. the Build needs work and from what im looking and reading Ainol seems to be doing a great job on releasing Builds that contain a good amount of fixes and optimization tweaks. expect a better Build release very soon

WAI

 

[sere83]

the Build/ROM that's based on this A10 Device is the exact same as the Novo7A but it's just coded for this Device that's all. the Build needs work and from what im looking and reading Ainol seems to be doing a great job on releasing Builds that contain a good amount of fixes and optimization tweaks. expect a better Build release very soon

WAI

Good news. Where did you see that they are working on optimizing the build? Did they mention anything about improving the lag?

 

[WAI]

Good news. Where did you see that they are working on optimizing the build? Did they mention anything about improving the lag?

this is just all from my point of view from Ainol. no sources or anything. i'm just saying

now the lag can be somewhat fixed with Scripts but choosing the right one is somewhat a hard chocse for new users. recently as of today newtech25 users: lordsbm and seasonscrystal released their script ROM which made my N7A run a little better. can be run on other Allwinnier based Tablets including Aurora of course

DOWNLOAD

and here's my Video of the Boot up Animation

 

[ivanz]

Yeah.. The script is quite great... Am using it right now.. I've just have to remove those free bloatwares apps...

Well, it has an option not to install though...

 

[WAI]

too bad TB doesn't work with this build of ICS sadly. it picks up the Backup Folder but doesn't allow me to restore as they aren't listed. maybe bad backups lol but o well

indeed though this script is well worth it

 

[CarsonUK]

ah does TB noty work with ICS yet? that is annoying!

 

[WAI]

ah does TB noty work with ICS yet? that is annoying!

as i said could be just bad backup

 

[Dr Martin]

I'm getting 1580 on Quandrant in the new 20120301 0.9 firmware. Anyone else getting anything different?

Martin

[quote name='MrTasselhof' timestamp='1327532630' post='305018']
I have been asked to run some benchmarks on the Aurora at the stock 1 Ghz speed.
Below are the Results - ICS 4.0.1 - Kernel 3.0.8 - Firmware Dated: 20111229

 

[Nickos-V]

Please also add in scores for:
Neocore - 3D bench
Vellamo (the final test score) - a comprehensive web browser bench

Some comments on the benchmarks:
1] Linpack. Score is great for Cortex A8. The highest I have ever seen with A8 is around 16.4 MFLOPS.

2] Your Quadrant score is on the low side. Hard to say where Aurora is being dragged down in score without Advanced breakdown.

3] Your Antutu Bench gets great score. I can tell you use class 10 but your read score is super high and only possible if using buffer and not the card. I have 1 class 10 SD card and in Windows 7 reads 15MB/s & writes 11.5MB/s when copying 4 AVIs over. I have never seen any class 10 get higher than 25MB/s.

4] Battery test is great on that tablet. Everyone should note that the 3h & 21m Aurora ran for was @ 100% CPU. ie, this only happens when playing higher end (3D) games. In real use, CPU will average 25-50% and battery last much longer (6+ hours). If someone mostly 3D games, battery will last for about 4 hours only.

5] Nenamark1. Your score is kinda low but probably from ICS and/or higher resolution on Aurora.

Nenamark1 & Neocore are mid level 3D benches. Nenamark2 is very intense 3D bench. Quadrant & Antutu are synthetic benches and can be manipulated to achieve higher scores. That is why looking at many benches is better than just 1 or 2.

For compare to i7 benchmarks; see here:
http://www.slatedroid.com/topic/29865-sdk1-vs-sdk21-firmware/
Going to update those results in 1 or 2 days to the latest.

 

[shivansps]

I took these today

 

[Nickos-V]

@shivansps
Thanks for adding to the benches.

I believe the A10 & maybe AML8726-M chips have dual core Mali-400 GPUs.

I say this because:
#1. Mali-400 runs @ 250-300MHz
#2. Vivante GC800 runs @ 575MHz
#3. GC800 in rk2918 shows 4 texture units and is single core
#4. A10 & AML8726 tablets show 8 texture units ( I assume 4 texture units per GPU? ).
#5. Pretty sure SGS2 also shows 8 texture units and has dual core Mali-400 GPU.
#6. A10 scores about 0-15% higher in regular 3D benches over rk2918 chips and much higher in synthetic benches like Antutu and Quadrant.

This link says the Mali-400 are clocked 50MHz slower in AML8726 compared to A10s. Could explain why there is a performance difference between them. If you consider dual core & both GPU cores clocked 50MHz less, then 8726 GPUs lose 100MHz total. Or about 20% per core in speed which could bring benches down by 20% or so for 8726 chips.
http://blog.slate4sure.com/tablet-cpu-comparison-amlogic-aml8726-rockchip-rk2918-allwinner-a10/
Not sure if he used some app to find out the clock speeds for these chips or somewhere from specs.

 

[lwberndtsonjr]

Thank you,
Lord, there's a lot to learn!

 

[flashatom]

What scores are people seeing in Sunspider Javascript benchmark in default browser?

I'm running it and getting in the region of 10,000 to 11,000ms with the stock ICS 4.0.3 rom...

See below where someone quotes as low as 3773 for Aurora using Essential Clean rom:
Benchmark

Can a few others try it and quote their scores here?

The benchmark can be run from here:
Sunspider javascript benchmark

Thanks!

 

[bash]

I'm finding web browsing on the Aurora can be slow.
So I have been investigating using sunspider bench marks, hence I ended up here.

novo7ess_aurora_rom_v0_3.img
ICS Browser - SunSpider 0.9.1 - 3,754

novo7ess_aurora_rom_v0_7.img
Dolphin HD - SunSpider 0.9.1 - 6,285
ICS Browser - SunSpider 0.9.1 - 7,085

ECR-Aurora-GoodixTS-v1.0a.img - Aniol WebSite
ICS Browser - SunSpider 0.9.1 - 8,304

Wondering why the Clockwork Roms sunspider performance has been getting worse, whats the trade off ?
Is there any easy way to get the performance back to that of 0.3release for the ECR-Aurora-GoodixTS-v1.0a.img.
I have tried copying the 0.3 build.prop file onto v1.0a release, no improvement.

Help would be appreciated, as I'm a newbie,

Thanks !