i5 & i7 surface 3 , also new surface 3

[megatronium]

Good point. If Intel designed the i5 and i7 with the same TDP then they should have accounted for performance difference in the design. Which from an engineering perspective your given a TDP target and performance metric you'd hit them with what ever techniques were needed to make both.

In another company in another time with much bigger hardware if the entry level system design over performed it was slowed so that it didn't cannibalize the mid range models and on up the line. there's give and take there also... often techniques developed for one can be used in others.

There's an article about boosting performance by 20% using undervolting somewhere in these threads. The reasoning is that the fan kicks on too early and throttles the CPU at 85 degrees when the CPU is rated up to 100. By undervolting the CPU, it stays at an acceptable temperature which allows the CPU longer sustained peak performance. I'm not an expert on this by anymeans but have dabbled with over/under volting and overclocking with good results on my desktop machines I've built. Who knows? There are benchmarks out that say the i7 is outperforming the i5 and i3 anyway and I think it'd be foolish of Microsoft to put out a product for more money that is outpaced by cheaper configurations.

 

[GreyFox7]

There's an article about boosting performance by 20% using undervolting somewhere in these threads. The reasoning is that the fan kicks on too early and throttles the CPU at 85 degrees when the CPU is rated up to 100. By undervolting the CPU, it stays at an acceptable temperature which allows the CPU longer sustained peak performance. I'm not an expert on this by anymeans but have dabbled with over/under volting and overclocking with good results on my desktop machines I've built. Who knows? There are benchmarks out that say the i7 is outperforming the i5 and i3 anyway and I think it'd be foolish of Microsoft to put out a product for more money that is outpaced by cheaper configurations.

Interesting... on a crude test while running a graphics benchmark I measured 96.1 degrees on the back of the case with a temperature probe while the fan was running. the test ran about 30-40minutes and the fan kicked in a few minutes into the test and ran continuously for the duration. I had warmed it up prior to starting the long test with a few iterations of a shorter test which yielded fairly consistent results across each iteration.

 

[Kif]

There's an article about boosting performance by 20% using undervolting somewhere in these threads. The reasoning is that the fan kicks on too early and throttles the CPU at 85 degrees when the CPU is rated up to 100. By undervolting the CPU, it stays at an acceptable temperature which allows the CPU longer sustained peak performance. I'm not an expert on this by anymeans but have dabbled with over/under volting and overclocking with good results on my desktop machines I've built. Who knows? There are benchmarks out that say the i7 is outperforming the i5 and i3 anyway and I think it'd be foolish of Microsoft to put out a product for more money that is outpaced by cheaper configurations.

I'm not an expert either but if you closely read Anandtech review he says in his opinion Microsoft made a better compromise between form factor and performance. Microsoft gave up some performance for a superior form factor. I remember reading that Microsoft worked with Intel to engineer the cooling solution so it's probably pretty well optimized at this point. Overall performance metrics are complicated but I would expect that the i7 will out perform for the most part however it's gains will be muted because of the thermal constants. This result is really expected in a ultra book\tablet form factor. I think people should temper their performance expectations to match the reality that this is a tablet. Personally, I'm completely happy with the performance of my i5 SP3.

 

[kristalsoldier]

I'm not an expert either but if you closely read Anandtech review he says in his opinion Microsoft made a better compromise between form factor and performance. Microsoft gave up some performance for a superior form factor. I remember reading that Microsoft worked with Intel to engineer the cooling solution so it's probably pretty well optimized at this point. Overall performance metrics are complicated but I would expect that the i7 will out perform for the most part however it's gains will be muted because of the thermal constants. This result is really expected in a ultra book\tablet form factor. I think people should temper their performance expectations to match the reality that this is a tablet. Personally, I'm completely happy with the performance of my i5 SP3.

Very true! And, I sometimes wonder why such expectations arise - particularly with the Surface devices!

 

[megatronium]

What expectations? That the i7 would perform better than the i5???

 

[grumpy]

Or that the SP3 would perform better than the SP2.

 

[Moonsurface]

Very true! And, I sometimes wonder why such expectations arise - particularly with the Surface devices!

Me too... it is a compromise of having such a thin and light device, if you're really wanting to hammer the performance then perhaps a Surface is not the right device for you.

I asked the question in another thread but don't think anyone really answered. The i7 seems to do quite well in the 3Dmark tests, and there was discussion about whether running the test multiple times would get the Surface to exhibit the same throttling as the i5. My question was, what sort of application would you be running (not games) that would cause a prolonged high activity similar to that of the 3D mark benchmark test? What would you be doing in the real world to hit that ceiling and for it to be a major problem with the activity you're doing?

I might hit the throttle if I was to try what I'm currently doing on my desktop PC which is exporting 200 RAWfiles from a 36megapixel DSLR to 1080p sized jpegs in Lightroom. I can tell it's stressing the PC a bit as I can hear the fan on the GPU kicking in, but would that sort of activity get close to limit of a Surface. I expect it might.. and it might slow down.. but heck, I've done the same task on my old laptop which would take a long time to do the same activity, the laptop ramps up the fan just when sat doing nothing... sometimes even while the lid is closed... I suspect the Surface would do better than that...and I'd never have even dreamed I'd be able to do the same on a tablet sized device.

 

[leeshor]

This may be a little off topic but this thread does remind me of a story. A long time ago a law office hired me to take care of their systems and network. I did check out everything the first day and didn't see a problem but the next day the "network" went down. When I came to their office I saw that the problem was the server, it was off.

I tuned it on and it appeared to run just fine but I decided to stick around for a while. Eventually it went off again When I took a look at it I noticed it was extremely hot on top of the case. It turns out that the network going down was one of their primary complaints but they never told me the server was turning itself off only that the network was having problems. I quickly realized that the fan in the power supply died. Replacing the power supply solved ALL their network problems.

(Kicker) after discussing the server issue with the office manager I found that it was only a few months old. The previous support company, after installing the, (new), server, set it in the middle of the floor with a fan blowing on the side with the cover off. The office manager asked what they were doing and they claimed they were seasoning the system. The other company knew all along there was a problem.

So what does this all mean? Environment counts. For any of the temp & throttle tests people are taking, for them to be able to compare with anything else, you would need to describe the environment. That includes the surface the Surface was on , the ambient air temp, background processes and more. It's unlikely that any 2 people will get the same result. But that's only my opinion.

 

[Moonsurface]

I'm going to wait to get mine use it and if I'm happy with how its running XY and Z I don't t really care.

In answer to Grumpy not all upgrades are to performance some are to usability. For me the SP3 is a viable and attractive option because of the changes Microsoft have made to the size and screen making it vastly more usable and attractive to me. I liked the SP2 but it wasn't something I'd considering buying, but as soon as I heard about the specs of the 3 it was much more attractive to me. I'm pretty sure I won't be disappointed but we shall see in time.

 

[Kif]

Or that the SP3 would perform better than the SP2.

I'm sure if a person tried hard enough they can find the negative in anything. For me I see the positives of having a larger better screen in a thinner lighter form factor. When I'm using my SP3 I don't ever think to myself I wish this thing had better performance. It quite literally blows the doors of my iPad Mini Retina. It's plenty fast in my opinion.

Regarding the cost versus performance ratio for aI7 processor upgrade it's always been a negative return of investments buying a machine with a top end processor. This has been true since the original Pentium processor came out. Buying more ram or larger HD is almost always a better investment. That's totally cool if people want to spend the extra $$$ to buy a top end processor but they should know it's a luxury item and not a necessity. If you really need the extra performance do yourself a favor and jump up to a machine that has M series instead of the Intel Y series.

 

[mcsenerd]

While I agree in theory with Kif, the reality is that the situation is somewhat likely worse than average with the SP3. With the limited thermal overhead of the machine design as is, that limited improvement for the I7 is even more limited. I'm not saying folks should or shouldn't buy it, but I wouldn't go into the I7 purchase thinking it's going to be some super clear performance winner in all cases against the i5 or the i3 for that matter. I'd say the biggest win would be just getting the bigger SSD personally.

 

[grumpy]

I'm sure if a person tried hard enough they can find the negative in anything.

You have seen my screen name, haven't you?

Seriously, my negativity is really a result of disappointment. I really like the idea of the Surface Pro. I loved my SP1. Sure, it had lackluster battery life, but it exceeded my other expectations and was problem free. I bought the 256GB SP2 with the expectation that it would be better than the SP1 - and initially it was, until the problems started. Then I was beaten down by Microsoft's abhorrent support. I certainly wasn't about to be an early adopter of the SP3 given my history, but I was hoping that MS would surprise me. Unfortunately, I think that they sacrificed too much chasing the iPad form factor.