t-series vs. p-series processors BATTERY LIFE

Hey-- I know that the txxxx series is 35w and the pxxxx is 25w, but does that translate into a large real-world difference? I'm looking at a p8600 vs. a t9550.. they'd be the same price.

BTW, the laptop is the Studio XPS 16. Also, there be major heat differences?

thanks

 

If you undervolt, the difference will be negligible, in terms of both heat and battery life.

 

Even if you don't, if you're just doing day-to-day tasks and keeping the CPU idle, the power consumption difference is negligible. Only if you work on your battery under full load constantly, will you notice a significant difference (which most people don't do).

 

Doing basic tasks, I'd peg the difference at 5% (less battery life for the T). Depending on the rest of your configuration, you could see a 15% drop in battery life under load.

 

The P series do have lower maximum idle and sleep power consumptions as well as lower maximum power rating. P series CPUs also undervolt, so the net effect is that they should always have lower power consumption.

How the 1 or 2W lower idle power translates to battery life depends on your other components. My E6400 with a P8600, Intel graphics, LED backlit display and Hitachi 5k500.B can consume as little as 8 or 9W and run for up to 6 hours. with a T series CPU I might lose an hour. Or I might lose half an hour, depending on the actual power consumption of the individual CPU. The 15% already suggested is a fair estimate.

If you want cool operation and longest battery time then the P series is the better option. Or you might strike lucky with a cool running T series.

John

 

If they would undervolt the same percentage that would be true.

I've been hearing that the T series undervolt to the same voltages as the P series. I don't know if it's true though. I can't test it as I have no T series around.

 

Pffttt if you undervolt the T series the difference in power will be negligible I mostly agree but what if you also undervolt the P series? the P series will haver better battery life always...

 

I don't think so. If they're running on the same voltages they will have the same power consumption afaik.

 

Yeah but you could also undervolt the P series and maybe they could use a lower voltage than the T series.

 

Well that would be the question. If they do need less voltage, they would use less power.

But do they need less voltage? I don't know.

And is the whole P and T series just one big marketing scam by Intel? Speculation of course but I would not be surprised.

 

For people who don't know a thing about Uving. It's not a scam ;-) How many people know about UVing or want to deal with that.

 

I don't know man, but according to this datasheet: http://download.intel.com/design/mobile/datashts/32012001.pdf

The 25W P and the 35W T series use same voltage, but the P uses less current(current x voltage = watts).

Seriously, the P series are binned for power. There's no conspiracy...

 

It makes sense there is a difference at normal voltages. But I don't know if the difference is not there when undervolting.

Too bad you were not around when that guy (forgot who it was) was building a pretty good case for the difference not being there when undervolting.

I'd like to see a real world test.

 

My T5800 undervolted was hotter than my P9600 not undervolted. T5800 is 65 nm old generation though.

Besides, as said the voltage isn't the only parameter when it comes to power consumption. Remember your physics class P=UxI. Intensity or current don't know how you call it may be lower in P series.

P9600 lower multiplier also is on 0,875 V without UVing. I don't believe that any T processors has a voltage that low in idle mode.

 

nevertheless don't know if the difference is huge in practice or even significant when you consider the whole rig consumption.

 

From what i've seen going on what people have posted on this forum the T9XX series processors seem to run cooler at idle and load (but not full load) than the P9XX series of processors. This is without any undervolting which i actually don't do and neither do a lot of people. This could possibly have implications for better battery life.

Also, in power saver they do downclock to more or less the same thing.

One would expect may be more a difference.

Of course at full load this is when you may notice a difference but when you're on battery power most people do not run their laptop in this manner.

 

It woud be very interesting if someone is in the position to test all this.

 

You wanted to mean the opposite, right ??? P9xxx are cooler than T9xxx on the same given cooling system, that's a fact. In both idle and full load mode. The stock voltage of a T9xxx in idle mode is 0,950-1V while mine is 0,875V (if i'm right)

 

From what I've seen on the forums, P series processors undervolt to 0.925V (minimum) almost always but I haven't noticed this with the T series.

And I can only guesstimate that P series will always use less power when under the same load as a T series and therefore be cooler. It may not be a big difference either way but it's still there.

But we would need someone willing to test both series processors on the same system to be sure.

 

No, i meant what i wrote.
I'm not the only person to have noticed this.
http://forum.notebookreview.com/showpost.php?p=5033388&postcount=2

I would like to also see more testing done about this.

Btw, i have a T9300 processor and mine underclocks on power saver to about 8.xxV something. I need to check it again for the actual figures.

 

Voltage is only half the story. Leakage current (= power = heat) is the other half.

My P8600 goes down to only 0.925V at SLF (800MHz) and is undervolted to 1.05V at 2.4HGz. However, core 0 is at 28C while core 1 is 32C (on idle) which I see as a strong indicator that one core leaks more power than the other (the temperature differential increases slightly under load). If both cores were like core 0 then I would have a delightfully efficient CPU. But it probably would be a P series any more, but one of CPUs with an even lower power rating (and higher price tag).

John

 

Well, It seems to vary from one rig to another...There seem to be no strict rules at all.

I have one core hotter than the other too (28-32°C too) It appears to be a P core particularity .

 

Fact? You mean based on the specifications right?

 

Mine is like that as well.
http://usera.ImageCave.com/rachuk/temps2.JPG
I tested it again when it was not in idle and my temps were like 30c and 32c.

Kallogan you're right results will vary from computer to computer.I still would like to see more testing though.

Some Z users computer idles at 40c using integrated graphics and that is a bit higher than i expected. John's though appears to idle much lower.
http://forum.notebookreview.com/showpost.php?p=4888556&postcount=8107
http://forum.notebookreview.com/showpost.php?p=4600556&postcount=6961

P9XX should have lower temps i agree.

 

I would not compare Z temps with E6400. E6400 is a much larger notebook.

 

The cooling system is another factor here. The E6400 has a big heat sink and an overall bigger chassis than the Sony Z.

John

 

Why not? You are always so keen on battery life.

 

Yeah if you take the strict "lower" voltage/amperage specifications. Of course there are always exceptions just to annoy me

 

The Z is not the only example that i have seen and could have given.
If i have time i will link to some more posts later.


Phil , i don't undervolt because i don't have any need to. Plus it is funny but i don't like any extra applications etc running when i'm on battery.

 

May i Add that your T9300 is the lower clocked cpu of the T9xxx serie and has only a 800 fsb and is not representative at all ???? Bad faith me ??? No...

 

You can ignore my results but i what i observed wasn't based on my results.

Btw, what are your temps Kallogan without undervolting?

It would be helpful if people could share this information and what their setup is.

I still think that Z's temps are a bit too high.

As we have seen though so far the temps of the T9XX at idle are may be lower than what some people may
have expected.

Edited to add:

The fairest example would be comparing say laptops that could be configured with either a TXXX or a PXXX series processor and compare the results.

 

I'm just teasing you. There seems to be no difference or so when it comes to idle mode.
0,875V is the original voltage for the lowest multiplier. No change. (28/32°C)
At full load and without Uving (stock 1,05V), the max i've seen after playing a demanding game is 70°C on the hotter core and 64°C on the other according to my memories.

Edit: That's not totally exact, when my gpu isn't UVed, cpu goes into the 70s i guess.

But an accurate test would be to just take the cpu separately in Orthos test.

 

Those are good results Kallogan.

I actually bought a P9500 processor to try and replace my T9300 procesor in my laptop but of course i discovered it did not work.

I started reading some results later that people were getting with the PXXX series processor and I was not impressed. I read a few people getting temps in the 40s at idle that owned different makes of laptops.

However, as this thread has shown there are some people that are hitting low temps also as i would have expected in the beginning which is good to see.

It makes sense to go with a PXXX series processor, battery life there might be nothing in but at load you should have a cooler running laptop.

 

The lowest my modest T8100 can go is 0.950V for all multipliers.

I'm pretty sure a P series processor is a cool processor that does not consume lots of power and give great battery life. Also, other components of your machine may contribute to increasing/lowering the overall battery life, for example a single platter vs dual platter (the 250GB Hitachi 5K500.B model comes to my mind right now) because it have ridiculously low power requirements I may get one of those in this month to do some testing, relating to my stock 80GB versus a new HDD with low power requirements.

 

My T5550 is at the min voltage for the bottom half of the multipliers. It runs super cool.

 

A user on a dutch forum collected these maximum undervolting results:

t5270 0.862
su9400 0.875
t6500 0.925
p8600 0.925
t7300 0.937
t8100 0.950
t2300 0.950
t9300 0.950
p9600 0.975
t7300 0.987
t9300 1.000
t7300 1.025
T9300 1.050
t7700 1.062
t8300 1.150