Here is an interesting comparison: i5 vs i7

[ctitanic]

The title explains it all

http://tabtec.com/windows/surface-pro-3-intel-core-i5-vs-core-i7-performance-battery-life/

 

[leeshor]

Interesting. If you break those differences down to percentages it really is a bigger difference that it would appear to be on the surface. More in line with what I would expect, despite reports here to the contrary.

 

[Moonsurface]

Interesting the comparison in the link between the i7 running on battery and running on power. I feel happy now as I generally don't run it when charging - but I guess for those with a docking station this may be an interesting point - it may throttle more when charging.

I don't feel the need to justify my purchase of the i7 to anyone but this has made me feel happier that the performance upgrade is not just placebo effect!

 

[Liam2349]

10% battery loss in 4 hours of sleep? That's a lot.

Last I checked, mine used 1 or 2%.

 

[Moonsurface]

10% battery loss in 4 hours of sleep? That's a lot.

Last I checked, mine used 1 or 2%.

True actually this was from my battery report from my i7 ...

 

[GreyFox7]

10% battery loss in 4 hours of sleep? That's a lot.

Last I checked, mine used 1 or 2%.

Some of this may be attributed to the methods they used to gather performance metrics. I know from experience you will get lots of detail and lower performance as well as lower battery life running XTU.
Still All things being equal the results are valid. YMMV.

 

[kristalsoldier]

True actually this was from my battery report from my i7 ...View attachment 3743

I see the same in my battery report. The lowest I have 1% and the max is 3%. And - based on my usage - I get 8.32 hrs of battery life.

 

[grumpy]

Of course, the cherry picked stats presented here stats hardly tell the whole story. The 3DMarks are entirely misleading, since there was likely little to no throttling occurring. Of course, people don't normally play games a few minutes at a time. Look at the 3DMarks scores during charging and the true story begins to reveal itself.

The WinRAR test starts to tell the story. The i7 performed faster initially, but was quickly throttled resulting in a 10% gain over the i5. A longer test would likely show an even lower performance gap as the time throttled significantly outweighed the time unthrottled. For applications requiring sustained performance, the i7 is only marginally faster than the i5. The moral of the story - save your pennies and don't be duped by Microsoft (and their shills).

 

[GreyFox7]

Of course, the cherry picked stats presented here stats hardly tell the whole story. The 3DMarks are entirely misleading, since there was likely little to no throttling occurring. Of course, people don't normally play games a few minutes at a time. Look at the 3DMarks scores during charging and the true story begins to reveal itself.
View attachment 3746
The WinRAR test starts to tell the story. The i7 performed faster initially, but was quickly throttled resulting in a 10% gain over the i5. A longer test would likely show an even lower performance gap as the time throttled significantly outweighed the time unthrottled. For applications requiring sustained performance, the i7 is only marginally faster than the i5. The moral of the story - save your pennies and don't be duped by Microsoft (and their shills).

The i5 also throttles.
Sustained performance tests with an i5 showed it jumped out of the starting blocks, throttled back sharply then gradually throttled a little less reaching its max sustained performance point.

Given that the Surface design will dissipate x amount of heat whether it has an i5 or i7 and both packages have a 15w Thermal Design Point with different performance there will be a performance difference regardless of the value of x. Watts translates to heat. Its analogous to two 15w light bulbs that produce different amounts of light.

Your benchmark of sustained performance is no more valid as a single measure of total performance than any other. The reality is most users usage profiles are not sustained usage rather they tend to be intermittent and bursty. YMMV, if you fall into the grumpy category you are not most users.

 

[grumpy]

The i5 also throttles.

*sigh*
Yes, yes... As does the SP2, the MBP, etc, etc...
The SP3 throttling just makes the i7 "upgrade" pretty pointless.

Given that the Surface design will dissipate x amount of heat whether it has an i5 or i7 and both packages have a 15w Thermal Design Point with different performance there will be a performance difference regardless of the value of x. Watts translates to heat. Its analogous to two 15w light bulbs that produce different amounts of light.

I suggest that you learn about TDP ratings. In short, under load, the i7 is capable of generating more heat than the i5. Why do you think that the i7s are more prone to thermal shutdowns....

Your benchmark of sustained performance is no more valid as a single measure of total performance than any other.

I'd wager that many considering the i7 over the i5 are concerned with sustained performance. After all, I don't think those users want an i7 so that their web pages render a few milliseconds faster...

The reality is most users usage profiles are not sustained usage rather they tend to be intermittent and bursty. YMMV, if you fall into the grumpy category you are not most users.

And most user's needs could be met with a WinRT device. The truth is that the i7 is marketed to power users and those users likely expect superior sustained performance - something that the i7 SP3 fails to deliver.

 

[Geek.Verve]

*sigh*
Yes, yes... As does the SP2, the MBP, etc, etc...
The SP3 throttling just makes the i7 "upgrade" pretty pointless.

Only if 4MB of L2 cache vs 3MB and a higher boost speed is not important to you. I don't encode video or render 3D objects on a tablet, but I *do* compile code - a process that generally takes less than 30-seconds with my projects. That is where the i7 will shine.

I'd wager that many considering the i7 over the i5 are concerned with sustained performance. After all, I don't think those users want an i7 so that their web pages render a few milliseconds faster...

Try multiple full seconds faster. Again, exactly the sort of thing where the i7 will be beneficial and why it's worth every penny of the $250 price difference.

Which SP3 do you have, if I may ask?

 

[megatronium]

Even rendering 3D objects or encoding video (for my needs) uses OPEN CL (Premiere, Photoshop, etc) which is not using CPU SUSTAINED PERFORMANCE. It is using the HD5000. If we're talking sustained CPU usage with regards to video, we're likely talking about Handbrake which isn't a concern of mine. I am a power user - the i7 is better for MY USES. YMMV.