I can almost, with 90% certainty saythat the p2 should use an amd apu!


From my predominantly uninformed and largely uneducated perspective...


The best option I can see for 'portable' x86 gaming at this point in time would be the Microsoft Surface Pro (+ ICP2)


I'm somewhat on the fence trying to decide if it's a worthwhile + practical device for the price... My gut tells me to look the other way and wait for the next generation....
Maybe the RAZR edge pro is a better soln for gaming although it is grossly impractical.
 
http://translate.google.com/translate?sl=auto&tl=en&u=http%3A//www.notebookcheck.com/Im-Test-AMD-A6-1450-APU-Temash.92206.0.html


Power draw is a high under maximum load, fully stressed it goes to 21w - that's GPU turbo and CPU downclock, but idle is a bit lower, at 6.8w balanced. The dual core can be better under max load though.


Power


Idle: (measurements on Windows 8 desktop)


Power saving mode, minimum brightness, wireless off: 5.3 watts


Balanced, maximum brightness, WLAN off: 6.8 watts


High performance, maximum brightness, WLAN on: 10.3 watts


Load: (measurements with maximum performance, maximum brightness and WiFi on)


Cinebench R11.5 single (~ 1.4 GHz CPU ): 13.7 Watt


Cinebench R11.5 multi (~ 1.1 GHz CPU ): 14.9 Watt


Prime95 large FFTs (~ 1.0 GHz CPU ): 16.9 Watt


FurMark (~ 1.2 GHz CPU ): 21.2 Watt


Prime95 + Furmark (~ 0.85 GHz CPU ): 21.9 Watt


3DMark 06: 19.5 watts


From idle we can see that the system zaps a lot of power, WiFi off and display min brightness...is 5.3w, the could even be lower with an ssd and ddr3lrs. Just adjusting the brightness and WiFi that's 5w! It also seems that the GPU is a power hog over 8w draw in turbo. Max soc power draw seems to be around 10-15w
How much of that 5.3W is from the SoC and how much is from the screen and the rest? That's the main question here. It could be nearly nothing for the SoC and 5W for the screen, but it could also be 1 or 2W for the SoC and 3 or 4W for the screen. Can you somehow get the power consumption with the screen off? (and of course no disks spinning or whatever else you have there). There is a huge difference between an idle SoC that uses 0.1W and one that uses 1W, and there is no way to see what is the case from your numbers.

My idle Pandora (screen off, wifi off, just the idle SoC) uses about 0.11W.  With the screen on it uses between 0.65W and 1.08W, depending on the backlight brightness. At 100% usage (NEON stress test), the CPU uses between 0.8W and 1.8W extra (depending on clock speed, the lower number is at 500MHz, the higher at 980MHz) compared to when it is idle, so the total power consumption of my Pandora in use varies between 0.6W (min brightness, cpu idle) and 2.8W (max brightness, full cpu load). This is the ballpark we want to stay in; I wouldn't mind if the upper bound increased, but I would appreciate it if the lower bound also decreased a bit in the Pandora 2; a range of 0.4W - 5W in use (screen on) would be fine (total power consumption), with a similar screen-off idle SoC total power consumption of ~0.1W or better.
 
http://translate.google.com/translate?sl=auto&tl=en&u=http%3A//www.notebookcheck.com/Im-Test-AMD-A6-1450-APU-Temash.92206.0.html


Power draw is a high under maximum load, fully stressed it goes to 21w - that's GPU turbo and CPU downclock, but idle is a bit lower, at 6.8w balanced. The dual core can be better under max load though.


Power


Idle: (measurements on Windows 8 desktop)


Power saving mode, minimum brightness, wireless off: 5.3 watts


Balanced, maximum brightness, WLAN off: 6.8 watts


High performance, maximum brightness, WLAN on: 10.3 watts


Load: (measurements with maximum performance, maximum brightness and WiFi on)


Cinebench R11.5 single (~ 1.4 GHz CPU ): 13.7 Watt


Cinebench R11.5 multi (~ 1.1 GHz CPU ): 14.9 Watt


Prime95 large FFTs (~ 1.0 GHz CPU ): 16.9 Watt


FurMark (~ 1.2 GHz CPU ): 21.2 Watt


Prime95 + Furmark (~ 0.85 GHz CPU ): 21.9 Watt


3DMark 06: 19.5 watts


From idle we can see that the system zaps a lot of power, WiFi off and display min brightness...is 5.3w, the could even be lower with an ssd and ddr3lrs. Just adjusting the brightness and WiFi that's 5w! It also seems that the GPU is a power hog over 8w draw in turbo. Max soc power draw seems to be around 10-15w
 
How much of that 5.3W is from the SoC and how much is from the screen and the rest? That's the main question here. It could be nearly nothing for the SoC and 5W for the screen, but it could also be 1 or 2W for the SoC and 3 or 4W for the screen. Can you somehow get the power consumption with the screen off? (and of course no disks spinning or whatever else you have there). There is a huge difference between an idle SoC that uses 0.1W and one that uses 1W, and there is no way to see what is the case from your numbers.


My idle Pandora (screen off, wifi off, just the idle SoC) uses about 0.11W.  With the screen on it uses between 0.65W and 1.08W, depending on the backlight brightness. At 100% usage (NEON stress test), the CPU uses between 0.8W and 1.8W extra (depending on clock speed, the lower number is at 500MHz, the higher at 980MHz) compared to when it is idle, so the total power consumption of my Pandora in use varies between 0.6W (min brightness, cpu idle) and 2.8W (max brightness, full cpu load). This is the ballpark we want to stay in; I wouldn't mind if the upper bound increased, but I would appreciate it if the lower bound also decreased a bit in the Pandora 2; a range of 0.4W - 5W in use (screen on) would be fine (total power consumption), with a similar screen-off idle SoC total power consumption of ~0.1W or better.
The lower bounds will be difficult for the apu to compete with a single core but

Power projections based on calculations carried out by AMD Performance Labs measuring total system and individual component power at Windows Idle and under various system loads while web browsing and/or viewing a 9:57 minute online video in h.264 format, viewed at 1080P setting at 100 nits. The AMD “Larne” reference platform is projected to measure APU power at 1.2 W at idle, 1.40 W during web browsing, 2.35 W during video playback and .02 W during a system S3 “sleep” state. Total system power for the reference platform is projected at 2.8 W at idle, 3.7 W during web browsing, 5.3 W during video playback and .07 W during a system S3 “sleep” state. Battery life calculations were derived using a 35Whr battery pack at 98% utilization. The power projections are based on the “Larne” reference system with a configuration including the A4-1200 Dual Core 1.0GHz APU, AMD Radeon™ HD 8180 series graphics, 2GB DDR3-1066 system memory and Microsoft Windows 8.
And claimed the soc draws 1.2w idle and has 0.03w s3 deep sleep, the upper bounds though would be stretched to around 5w under load
 
its all a numbers game.
The very fact that it isn't a numbers game is what makes this community special.
If it were, then ED would have left all the preorders to Craig and vanished to find something profitable.


If it were a numbers game, the P1 would be considered a prehistoric piece of junk.
I am not sure of eds back story but he gotta be making something, this can't all be passion.
ED and Craig haven't made a cent on the Pandora. They both have a monstrous amount of debt.
Yep. Not every person on this planet is a capitalist. Nuff said.
 
Last edited by a moderator:
its all a numbers game.
The very fact that it isn't a numbers game is what makes this community special.
If it were, then ED would have left all the preorders to Craig and vanished to find something profitable.


If it were a numbers game, the P1 would be considered a prehistoric piece of junk.
I am not sure of eds back story but he gotta be making something, this can't all be passion.
ED and Craig haven't made a cent on the Pandora. They both have a monstrous amount of debt.
 
Yep. Not every person on this planet is a capitalist. Nuff said.
Yet he owns a business...look at Linux, greatest piece of software ever, people actually get paid to develop it via capitalist means, strategy and philosophy can be sand boxed yet complement each other
 
Last edited by a moderator:
How much of that 5.3W is from the SoC and how much is from the screen and the rest? That's the main question here. It could be nearly nothing for the SoC and 5W for the screen, but it could also be 1 or 2W for the SoC and 3 or 4W for the screen. Can you somehow get the power consumption with the screen off? (and of course no disks spinning or whatever else you have there). There is a huge difference between an idle SoC that uses 0.1W and one that uses 1W, and there is no way to see what is the case from your numbers.
It's probably 1-2W. Good for a laptop chip, but unacceptably high for a phone or small tablet chip.

The bigger question in my mind is, what accounts for the 3.5W difference between balanced w/wifi off and maximum performance w/wifi on? Unless the tester was running bittorrent in the background while the unit was supposed to be idle it can't be the wifi. The dark ages of wifi chips using a bunch of power just by being on are long over, they have much better power management than that now. The usage should be almost nothing, meaning that most of the difference is probably on the SoC. I'm guessing maximum performance mode is preventing the lower idle states and maybe even DVFS and core power gating entirely, which would leave the power savings only to clock gating. This would be done in order to maximize response time, possibly over-aggressively. Whatever the case I expect that ~3W of that is from the SoC, probably from the CPU cores.

Let's say there was 1.5W in the SoC upfront when in "balanced" mode. That puts the single core @ 1.4GHz number at ~8W and the quad core @ 1.1GHz (where it's probably being power or thermally throttled) at ~9W. I'll ignore the Prime 95 numbers for now, those are probably pretty pathological, although Cinebench may not represent real core pegging.

This also includes RAM and power loss in regulators. The new review establishes only 2GB of DDR3L-1600, exactly like the Exynos Chromebook. I estimated for that it uses 2W. I was going to go with a full estimation here, but honestly this is way too haphazard; the difference between 1W vs 1.5W @ 1GHz is a very big deal. Instead it's best to just wait for a site like Anandtech to actually probe the SoC power rails. But I'm pretty skeptical that it'll be competitive.

My idle Pandora (screen off, wifi off, just the idle SoC) uses about 0.11W.  With the screen on it uses between 0.65W and 1.08W, depending on the backlight brightness. At 100% usage (NEON stress test), the CPU uses between 0.8W and 1.8W extra (depending on clock speed, the lower number is at 500MHz, the higher at 980MHz) compared to when it is idle, so the total power consumption of my Pandora in use varies between 0.6W (min brightness, cpu idle) and 2.8W (max brightness, full cpu load). This is the ballpark we want to stay in; I wouldn't mind if the upper bound increased, but I would appreciate it if the lower bound also decreased a bit in the Pandora 2; a range of 0.4W - 5W in use (screen on) would be fine (total power consumption), with a similar screen-off idle SoC total power consumption of ~0.1W or better.
That's not just the CPU, but RAM + loss due to regulators. Pandora uses MDDR, 1.5V stuff which is behind LPDDR2 in perf/W.

Also, for anyone reading and thinking my "well below 1W" figure was off, note that this is an original CC Pandora and 980MHz is really high. And NEON stress test is probably not a normal full CPU time load. 1GHz Pandoras will use a lot less power at 1GHz by virtue of being on a better process.

If we use a multicore Cortex-A15 SoC it'll use a few W at full load. But it should only need about 500-700MHz with one processor to match 1GHz Pandora performance, where it should use < 500mW. This will be enough to at least match Pandora's battery life when running the same sorts of things. This is also why I think it matters that it does run the same binaries, to further discourage people from letting the software bloat to use up the extra clock speed.
 
And claimed the soc draws 1.2w idle and has 0.03w s3 deep sleep, the upper bounds though would be stretched to around 5w under load
I just did a quick test with my P1:

  • BT on, inactive
  • wifi on, downloading stuff
  • backlight at full brightness
  • CPU stress test running
I reached 2.8W for the entire device

With:

  • BT off
  • wifi on (inactive)
  • screen dim
  • CPU at 15%
I get 0.8W for the entire device
 
emot-words.gif
 
No not every handheld needs to be mainstream but it needs a good business model...in the end its all a numbers game.

Also about the just because, the reasons given to stay arm, save for power, are all weak IMO.


The legacy argument doesn't really hold up, 900 packages versus 1000s more and commercial software.


The prestige-image argument doesn't make sense either, who buys a Pandora for an arm part processor?


The dev argument is also a chicken and egg issue, with a more flexible platform with larger audience could bring in more devs.


As for the power argument, we were debating the quad core version that uses more power but in my original post I was making a case form its lower powered dual core variant, which he said would lose out on performance.
Ah, my misunderstanding. 

The legacy argument goes along with the power consumption argument, if ARM were outclassed then I would agree we wouldn't have to stay. But 0.11W idle vs maybe 1.2W... that's not quite the same ballpark. Our software is optimized against the former, so it makes more sense to stay.

I must have missed the posts about prestige and Open Pandora, I get more weird than envious looks tbph.
 
I am kinda disappointed in the a6-1450, although it could just be implementation. The phantom 3.5W wifi radio, 1W keyboard backlight and all the other inefficiencies, not to mention the relatively crazy power utilization under max load and my high expectations makes my excitement shortlived. Maybe the a4-1250 will fair better.
 
monstercameron wins the prize for most deadly sig ever, with 10 youtube videos!

guaranteed to kill firefox on your Pandora or netbook

my ego wants me to believe that some people like or are interested in my apu related threads...
Yeah I don't mind, I'm just stunned that you can get 150 posts on this topic in a couple days.

If the future is now and x86 smartphones are now competitive to ARM on performance and power, you can go ahead and build an x86 Pandora. I'll buy one.
 
Last edited by a moderator:
I guess you were trying to say "oh sorry about that" :/
 
if your sig is changed, it changes for all old posts too. so it's ok now
 
Last edited by a moderator:
Add caanoo support, so I can dust it off...
I'd like to, but sadly I don't have a Caanoo.  I'm not going to attempt to port it without the device.  How many users does Caanoo have compared to Pandora?  Maybe I'll get one, pretty cool that Notaz ported PCSX re-armed.
 
Last edited by a moderator:
I say go AMD. I had an Atom N450 netbook and it was the worst $400 I had ever spent. If the P2 has the word atom anywhere on it, I'm out. So, back to the Power vs battery life debate....

How about we have a dual-processor system? AMD for the POWAH and ARM for Android and the "old" OP1 OS.
 
Back
Top