4 ms·

My question is if this amounts to 1/6 the battery life. I personally don't care for gfx intensive mobile games, i would rather OEM's focus on efficiency rather

My question is if this amounts to 1/6 the battery life. I personally don't care for gfx intensive mobile games, i would rather OEM's focus on efficiency rather then raw power. But thats just my 2 cents.

mseebach 14y ago
More interesting to me in this regard are the "4+1" cores. The +1 is a low-powered standby core. I don't know if it's new, but if not, it could change the battery life profile quite a bit.
- dmbass 14y ago
  The Tegra 3 had this feature and appears to work relatively well according to http://www.anandtech.com/show/5779/htc-one-x-for-att-review/3 http://www.anandtech.com/show/5779/htc-one-x-for-att-review/...
  - fredliu 14y ago
    If i'm not mistaken, that review is for One X for att, which doesn't have Tegra 3 in it. It does have data for Tegra 3 (One X international version), but it doesn't seem to support the claim that Tegra 3 is more power efficient because of the +1 core.
    - mtgx 14y ago
      Maybe not more efficient than everything else out there, but more efficient than if it used 4 cores only - especially at that 40 nm process. I think Samsung's Exynos processor was a bit more efficient, even though it didn't have the 5th core, but it was made at 32nm.
- fredliu 14y ago
  It's not new, it's in tegra3 also. I never saw real profiling data/usage report regarding how effective this "+1" core is in saving power. But I did saw anecdotal report that the "+1" core often times works at the wrong time that hurts performance... (e.g. still used despite of heavy load, etc. )
adrr 14y ago
Tegra does have 5th low power core which is for low CPU tasks. I am curious what percentage of the battery is used for CPU, compared to other parts of a phone. Turning down the screen brightness gives me a significant increase in battery life, also turning off bluetooth and wifi when i am not using them.
- polshaw 14y ago
  >5th low power core As did the Tegra 3. Android breaks down battery use stats for each part of the phone; it really depends on what you are doing. But for the normal phone use; a bit of web browsing, simple apps etc, the screen dominates power consumption (say 70%).
polshaw 14y ago
Raw power and efficiency are not unrelated. The key gain available today, other than process node shrinks, is being faster at the race to idle. If the CPU uses 1W at 100% and 10mw idle, then the quicker you complete your processing task the better. There are some decent options for low power; the new A7 is basically a lower power/multicore A8 (found in SGS1 gen devices). But anything close to 1/6 is not happening, unless you get an E-ink display, possibly.
- jug6ernaut 14y ago
  Very interesting analysis. I am in no part discouting what you say, but i for one would like to see the OEM's push this increase in efficiency. Obviously when it comes to battery life there are a million and one different factors, but considering a mobile devices usefulness is directly related to its ability to run(have battery) you would think this would be a more focused area. I mean i know most people are happy with a full days charge but i dont think it would be outlandish to say that the majority of phones limp in one 1 days use. Battery life has decreased over the years not increased. My comment of "1/6 the battery usage" was more to if it has 6x the power does it use 6x the energy?
  - polshaw 14y ago
    Basically, we are seeing devices with an increased dynamic range WRT power. They are going to idle at similar power, and they will have better performance-per-watt, but at 100%, they will use more. IMO, this isn't so bad. Take it to it's ultimate end, and you get a device that you can plug in and will perform like your laptop[1], but can still last a day, less fast, in your pocket. Would you prefer not to have this option of 'overpowering'? If you give both devices the exact same task, the more powerful one will use less power. The risk of course is developers-- given all this extra power is available-- may be temptated to write lazily, or include superfluous 'sugar' (OTT animations etc), which could lead to worse battery life than the basic last-gen apps. If you take the cutting edge gaming, then it would seem that you will get less playing time on the go with the tegra4 than the tegra3; there is no way that at full-tilt they use close to the same power.. but it can't be 6x either[2] But there seems to be a growing rift between the 'mobile' games and the cutting edge 3d ones.. perhaps we will see more 'low power/graphics' options available for OTG playing? 1. and we are really 'there' with chips like (quad core A15) tegra 4; the dual core A15s in the chromebook are already matching DC atoms. 2. you do have to wonder; how many watts can you pack into a smartphone without burning someone's hands off-- so there is a limit, and it can't be too far from today's devices! Having said that, tegra4 may well be tablet targeted anyway? or feature lower clocks in a phone; we don't really know much about it yet. EDIT: also, i really don't think we will see much from the OEMs on this front, unless it becomes an in thing at some point. It is hard to successfully push anything that doesn't have a well understood metric to it (like screen resolution, cores, clock speed) The issue is really one for the SOC guys (samsung, nvidia, qualcomm etc) and CPU designers (ARM, qualcomm, apple); we may see some movement if apple makes good progress and starts to use the fact heavily in marketing. Just look at the razr max; not efficiency exactly-- but battery life. And that didn't have much of an impact.

mseebach 14y ago

More interesting to me in this regard are the "4+1" cores. The +1 is a low-powered standby core. I don't know if it's new, but if not, it could change the battery life profile quite a bit.

dmbass 14y ago
The Tegra 3 had this feature and appears to work relatively well according to http://www.anandtech.com/show/5779/htc-one-x-for-att-review/3 http://www.anandtech.com/show/5779/htc-one-x-for-att-review/...
- fredliu 14y ago
  If i'm not mistaken, that review is for One X for att, which doesn't have Tegra 3 in it. It does have data for Tegra 3 (One X international version), but it doesn't seem to support the claim that Tegra 3 is more power efficient because of the +1 core.
  - mtgx 14y ago
    Maybe not more efficient than everything else out there, but more efficient than if it used 4 cores only - especially at that 40 nm process. I think Samsung's Exynos processor was a bit more efficient, even though it didn't have the 5th core, but it was made at 32nm.
fredliu 14y ago
It's not new, it's in tegra3 also. I never saw real profiling data/usage report regarding how effective this "+1" core is in saving power. But I did saw anecdotal report that the "+1" core often times works at the wrong time that hurts performance... (e.g. still used despite of heavy load, etc. )

adrr 14y ago

Tegra does have 5th low power core which is for low CPU tasks. I am curious what percentage of the battery is used for CPU, compared to other parts of a phone. Turning down the screen brightness gives me a significant increase in battery life, also turning off bluetooth and wifi when i am not using them.

polshaw 14y ago
>5th low power core As did the Tegra 3. Android breaks down battery use stats for each part of the phone; it really depends on what you are doing. But for the normal phone use; a bit of web browsing, simple apps etc, the screen dominates power consumption (say 70%).

polshaw 14y ago

Raw power and efficiency are not unrelated. The key gain available today, other than process node shrinks, is being faster at the race to idle. If the CPU uses 1W at 100% and 10mw idle, then the quicker you complete your processing task the better. There are some decent options for low power; the new A7 is basically a lower power/multicore A8 (found in SGS1 gen devices). But anything close to 1/6 is not happening, unless you get an E-ink display, possibly.

jug6ernaut 14y ago
Very interesting analysis. I am in no part discouting what you say, but i for one would like to see the OEM's push this increase in efficiency. Obviously when it comes to battery life there are a million and one different factors, but considering a mobile devices usefulness is directly related to its ability to run(have battery) you would think this would be a more focused area. I mean i know most people are happy with a full days charge but i dont think it would be outlandish to say that the majority of phones limp in one 1 days use. Battery life has decreased over the years not increased. My comment of "1/6 the battery usage" was more to if it has 6x the power does it use 6x the energy?
- polshaw 14y ago
  Basically, we are seeing devices with an increased dynamic range WRT power. They are going to idle at similar power, and they will have better performance-per-watt, but at 100%, they will use more. IMO, this isn't so bad. Take it to it's ultimate end, and you get a device that you can plug in and will perform like your laptop[1], but can still last a day, less fast, in your pocket. Would you prefer not to have this option of 'overpowering'? If you give both devices the exact same task, the more powerful one will use less power. The risk of course is developers-- given all this extra power is available-- may be temptated to write lazily, or include superfluous 'sugar' (OTT animations etc), which could lead to worse battery life than the basic last-gen apps. If you take the cutting edge gaming, then it would seem that you will get less playing time on the go with the tegra4 than the tegra3; there is no way that at full-tilt they use close to the same power.. but it can't be 6x either[2] But there seems to be a growing rift between the 'mobile' games and the cutting edge 3d ones.. perhaps we will see more 'low power/graphics' options available for OTG playing? 1. and we are really 'there' with chips like (quad core A15) tegra 4; the dual core A15s in the chromebook are already matching DC atoms. 2. you do have to wonder; how many watts can you pack into a smartphone without burning someone's hands off-- so there is a limit, and it can't be too far from today's devices! Having said that, tegra4 may well be tablet targeted anyway? or feature lower clocks in a phone; we don't really know much about it yet. EDIT: also, i really don't think we will see much from the OEMs on this front, unless it becomes an in thing at some point. It is hard to successfully push anything that doesn't have a well understood metric to it (like screen resolution, cores, clock speed) The issue is really one for the SOC guys (samsung, nvidia, qualcomm etc) and CPU designers (ARM, qualcomm, apple); we may see some movement if apple makes good progress and starts to use the fact heavily in marketing. Just look at the razr max; not efficiency exactly-- but battery life. And that didn't have much of an impact.