On Monday I got an email from Microsoft telling me that a $2700 Surface Book was waiting for me at my local FedEx. I hurried over to pick it up and then I immediately went to the Apple Store in town to buy (what I thought was) as comparable MacBook Pro.
Before entering the Apple store I opened up the Surface Book and checked the specs; 512GB SSD Intel i7 2.6GHz, 16GB of RAM. I then looked up the actual value of this laptop on Microsoft's website; $2700, wow....
I went in the Apple store and started looking at 13 and 15 inch MacBook Pros. Every 13 inch MacBook Pro had an i5 processor and was significantly cheaper than my $2700 Surface Book. Luckily I found a 15 inch MBP that appeared to have almost identical specs; 512 SSD, 2.5GHz i7, and 16GB RAM. The price was $2500, $200 cheaper than the Surface Book but it was close. Before I swiped my credit card I showed the Apple employee the exact specs of the Surface Book and asked if the processors were comparable, he said they were, but we both ended up being wrong.
I got back to the studio and started filming speed tests in Photoshop, Lightroom, and Premiere. The tests did not go well for the Surface Book. In Photoshop and Lightroom, the MacBook Pro was able to burn through jpeg and raw files at twice the speed of the Surface Book. In Premiere, the MBP was able to render 4k footage with up to 4 effects on the fly while the Surface Book couldn't even play 4k footage smoothly at 1/8 resolution. As one of the last photographers I know using Windows, I left defeated that night.
Later that night one of our writers pointed out that although both processors are Intel i7 chips, the Apple chip is a quad core while the Microsoft chip is only a dual core. This explained why the Apple was twice as fast, it has a processor that is literally twice as fast.
Once I figured this out I felt better because I knew that Apple didn't have some magic software that could make its hardware twice as fast but I also felt worse because the MacBook Pro was actually $200 cheaper and it was still twice as fast as the Surface Book.
If you're deciding which laptop is better based on power alone, the MBP is the clear winner. If, however, you are looking at all aspects including design quality, included accessories, touch capabilities, and of course the fact that the Surface Book is a laptop and tablet in one, the decision isn't so clear.
As someone who appreciates design and craftsmanship, I've always admired Apple products. Every aspect of an Apple laptop feels completely thought-out while my plastic Windows laptops feel cheap and disposable. In the video below I compare the design and craftsmanship of each laptop.
At the end of the day the Surface Book is an amazing achievement. I don't understand how they made that magnetic hinge, it works so well that it's a pleasure every time I detach the screen from the keyboard but the poor battery life of the tablet portion of the Surface Book makes it pretty useless to use without the base. It's also not the power house that I thought it would be. 4k video footage is quickly becoming the new standard and the laptop was unable to edit it without rendering first, that alone would keep me from spending $2700 on it but I realize that the average Surface Book user, and even the average Fstoppers reader, does not edit 4k video footage. For you, the Surface Book may be powerful "enough" but that's up to you to decide.
Because the Surface Book is so expensive, and the tablet portion has such limited battery life, I personally lean toward the Surface Pro 4 myself. You can read my full write up on the Surface Pro 4 here and you can watch my comparison video between the Surface Pro 4 and Surface Book below.
Here comes the war!!!! Hahahahaha
Fair enough. It is now clear why Microsoft has kept all press mentions and marketing speak to "sixth gen i7 processor" and not simply dual core i7. There's a massive difference between quad and dual core i7 performance, and I was really hoping that MS pulled off some black magic by putting a quad i7 in that thin body. Ah well.
I edit on a Windows Gigabyte P34G V2 - i7 4710HQ, 16GB RAM, 512GB SSD, 2TB HDD etc etc - and I was really looking forward to this Surface Book to perhaps go back to the Vaio Z dimensions of yore. The 2012 Vaio Z had a i7 quad core, 8GB RAM, SSDs in RAID 0, all in a wafer thin frame. Damn I miss those days.
PS: If you really want a Macbook Pro 15 competitor, look at the new XPS 15 with the infinity display. Quad i7, multiple drive systems, up to *32*GB RAM and a really really sexy design. No tablet included though.
It would be really cool to test those XPS 15 - form factor vs. performance seems to beat MBPs
FYI, when it comes to processors, the "speed" listed (ie how many GHz) doesn't actually tell you how fast the processor is. We've been at roughly 3 GHz maximum speed for a very long time now. I've found the linked website below to be dead on when it comes to relative speeds of processors (engineer by trade - I run a lot of very computationally heavy processes that, at times, took days to run). When it says one processor scores 8000 and another scores 4000, you can be sure that the 8000 is twice as fast as the 4000. Cheers.
From a marketing standpoint, why would Intel allow this? Wouldn't they want the consumer knowing exactly what they are buying?
Most people don't know that's the point. And they don't care.
Those companies cater to the larger uninformed mass. Not the computer savvy minority.
No, they want the customers buying stuff. And given that Apple is huge compared to Microsoft in terms of hardware sales, Intel is likely making more off of every Surface processor they sell than they make on every MBP processor they sell, so there might actually be an economic advantage to the fuzz for them.
Allow what? Benchmarking? Of course they'll allow that - they dominate the charts over their competitors.
Or do you mean allowing people to be misguided by clock speed? Clock speed is an actual measurement of the processor's capabilities - its not a marketing bit unless someone wants to use it as one. Its like horsepower in a car. More is better, sure. But a 350 HP car can easily be outrun by a 300 HP car if the 300 HP car is significantly lighter in weight. More HP also doesn't tell you much about fuel economy except that more is likely less efficient, but even there, a 400 HP car that weighs 2500 lbs can likely get better fuel economy than a 350 HP car that weighs 5000 lbs. With cars, things like the power curve, gearing, aerodynamics and a host of other factors will affect the performance of the car. Same is true for processors - clock speed is but one of many factors affecting performance.
You can make the same argument with cameras...higher number of pixels doesn't necessarily equate to better images...its but one of many factors you must consider.
No, a simple naming system that reflects the power of their products.
When intel released its Core i3, i5, and i7. The numbers indicated the quantity of threads that a processor can handle:
-A core i3 handles two threads which is only expressed through a dual core
-A core i5 handled four threads which can either be expressed through a hyper-threaded dual core or through a native quad core.
-A core i7 handlesd eight theads which is expressed by a hypet threaded quad core.
Apple's purchasing power influenced intel to altar the i3, i5 and i7 into a good/better/best moniker. That trend has spread so that it now impossible to compare products between brands without performing research.
Their naming system is as simple as it can be while still maintaining accuracy as to what it is you're buying. Like Andreas Werner mentioned before me, the core chip's number denotes the number of threads a processor can use simultaneously.
Take a look at this full name example: Intel Core i7-6700K 8M Skylake Quad-Core 4.0 GHz LGA 1151
i7-#### is the iteration of that specific chip and it may be followed by a "k" which denotes whether the chip can be overclocked. The following number denotes the CPU cache (8MB), and then a processor code name which reflects the process the chip was created with (usually stepping down in size [nm]), this affects power consumption and heat generation. Next we have number of cores, followed by the clock speed (4 GHz), and, finally, the CPU socket. Each one of these numbers plays a very real and noticeable role in the processor's capabilities.
As photographers and videographers (particularly video people) we can leverage the use of i7s because multithreading is highly sought after for rendering videos because it directly impacts the speed at which rendering happens. For others though, business users (and even gaming), an i5 is plenty and they would gain more real world effect from a faster clock speed and more cpu cache, than they would stepping up to an i7. It's relatively complicated, but not needlessly complicated...
Should also consider the TDP of the processors used. Macbook Pros typically use 40W cpus. The MS laptops are using ULV cpus with 15W TDP.
Steer clear from that site.
Its tests ( done with a mediocre software called Passmark Performance Test ) are synthetic and non-real-life valid.
You can have a much better image, yet by using a similar site/tool, called Bench from AnandTech.com which uses actual application tests not synthetic tests which even in "lab coded" conditions aren't consistent.
Last but not least, when you are after let's say the fastest processor for a certain or list of applications, see what they do in those, not in a bunch of similar or not similar at all applications ( or worse, synthetic benchmarks such as Passmark, PC Mark, SuperPi, wPrime, Prime95, SiSoft Sandra, AIDA64, etc ).
As I said, my experience running CFD code matched up with their results with surprising accuracy. If they said processor X is twice as fast as processor Y, sure enough, processor X would take 24 hours to run the code while processor Y would take 48. Same can be said for smaller differences. Synthetic or not, their results matched my real world, computationally intense application quite well.
But every application is different; if you've found another site that works for yours, by all means have at it.
From the very basics, there's no linear scaling in processor performance.
A 3GHz Skylake 4cores/8threads won't take twice as long as a 3GHz Skylake 8cores/16threads chip.
I just warned people, if you feel like arguing for the sake of it, you're welcome, just find who I am and what I do for a living first ;)
You've either drastically misinterpreted my statements, or you're pulling a straw man argument. I never once said that specification X on processor A is twice as much as processor B, therefore processor A is twice as fast. Not once. I said if that website scores processor B at 4000 and processor A at 8000, in my direct and objective experience, processor A will complete the simulation twice as fast as processor B.
My original post was spelling out that specifications do not necessarily scale...so why are you redirecting this conversation to specification sheets?
So you are saying, assuming that we take a processor rated by them at 4000 and one at 8000, we put them through 30 application tests ( Photoshop, Lightroom, AfterEffects, Premiere, Visual C, SunSpider, SAP® Benchmark, Maya 3D, AutoCAD, etc ) the 8000 rated processor will be twice as fast as the 4000 rated CPU ?
I am saying that for a $70,000 CFD program that utilizes a discretization method that is heavily dependent on CPU speed (not RAM, not graphics, not disk speed), that yes, it scales exactly as they report - the processor that they rated as being twice as fast was literally twice as fast. We're talking about crunching numbers and only crunching numbers - not moving a mouse around, not multi-tasking, not graphics rendering - pure CPU calculations. This is also done over the course of days/weeks - not seconds. So I can be quite accurate in that assessment.
I also specifically said: "But every application is different; if you've found another site that works for yours, by all means have at it." Meaning that no, it may not scale as accurately for your particular application and another site may be a better resource. This is also a discussion about pure CPU speed - many applications make heavier use of RAM, graphics, and disk speed - so the CPU may not be the bottleneck.
So no, I'm not saying anything of the sort. I'm saying that the CPU Passmark site scales as they state for a computationally intense application like CFD computations and that your application may vary.
Of course "real-world" tests like sporadic Photoshop and Chrome don't scale well, but some applications do scale somewhat linearly. CFD seems like something that would scale well.
Sounds like the intent was to "disprove" than to evaluate.
They genuinely had no idea whatsoever going into this what the performance difference would be. Lee is a Windows user, too.
"As one of the last Photographer's using Windows that I know, I left defeated that night."
You're not alone, Lee, you're not alone... ;)
Lee...yes you will see a drastic performance difference from dual core vs quad core processors. So, the evaluate is not really matching apples to apples.
Although it may not be apples to apples, for $200 less, why wouldn't you buy the larger, faster processing macbook? Personally, I can't justify a somewhat gimmicky removable touchscreen when there's a larger, quicker laptop in roughly the same pricepoint.
I'll take this one. You are paying a lot of money for all of the extra features the form factor of a hybrid device provides. For some it's a gimick. For others, including me, it's a huge and unequaled productivity boost to have a computer like the SB.
You're not put off by the loss of processing speed compared to a traditional laptop at a lower price point? I've only used a macbook for editing photos (currently a 13" air because I'm a broke college student and this is what is included in tuition), no wacom or anything, just the track pad. Is the use of a touchscreen worth the $200 and lesser performance?
Well said. For me to whip the tablet out of the camera bag, plug into USB 3 and rip across a few hundred images and start editing on the spot - priceless! In fact I use the SP3 so much as a tablet with Lightroom that I almost forget how to use it with a mouse! ;-)
C'mon, they're not even classed as the same type of machine. The MS Surface is a full OS tablet, not a laptop, and as such a lot of it's architecture is dedicated to things other that running Photoshop. There is also no mention of optimization on either machine. We really should have better technical maturity these days.
Obviously the Surface Book is a different beast but it was MS that compared it to the 13inch MBP in their press event. They claimed it was 50% faster than a 13 inch MBP. They also failed to mention how much cheaper a 13 inch MBP is.
I am the farthest thing from an Apple Fanboy. I have literally never owned a Mac and I own 6 Windows machines personally. I was rooting for the Surface Book in this test but it obviously didn't work out. I tried to be completely clear about my "error" but being that these products are similarly priced, I think this is still a fair test because MS marketed the Surface Book as an ultra powerful laptop.
Exactly. Microsoft is the one that opened the door to the comparison. They used it as part of their marketing.
Wait if you knew that they compared it to the 13ich one why did you not do that too... lolol and you asked to get the close as possible specs.. guess what.. they don't have a 3k display laptop, with a pen and basically a changeable GPU
Are you actually faulting MS because they offer a model with a discrete graphics processor that costs a lot more than the best model of Apple without one?
(this completely ignores of course the other differences in hardware)
Actually, that MacBook has an AMD R9 M370X 2GB, which benchmarks at twice what the Nvidia GTX 940M in the SurfaceBook does.
Even the Intel HD 5200 Iris Pro integrated graphics on the MacBook Pro 15 benchmarks 20% higher than the SurfaceBook's GPU.
For the price difference, you can then go buy Windows, and bootcamp that Mac... have a machine twice as fast, a better warranty and service, and still have some money left over.
we can't compare full laptops vs ultrabooks/convertiables thats like comparing phones vs desktops.
R9 M370X just doesn't go into a 13"
Lee, you can't compare a 13" to a 15" because they are different class in both weight and the equipment you can put in them. 15" use H processors and will fit a dedicated graphic cards in almost all cases. while almost all 13" ultra thin have U processors and no dedicated graphics. so we can only compare the surface book to the mac book pro 13"
So compare let us compare 13" mac and surface book
On the website after configuring the 13" mac pro to have i7, 16GB memory, 512 SSD the price comes out to $2199 while the surface book is at $2699. But at this point the mac is still missing a touch screen, stylus, ability to convert to tablet mode, and a dedicated graphic card. with all these additional feature that should cover the additional $500. therefore the price make sense even if the surface book is not faster. the fact that its faster should be a bonus.
Now all these features doesn't always apply to everyone. It sounds like your work load demands more of a full laptop rather than an ultrabook. In this case apple's 15" makes more sense for your workload while MS still have not come out with a device in that area yet. so if you go windows then you probably have to do with one of those "not so nice looking" third party vendors.
I understand but from what I've seen the mbp 13 is still faster in the majority of tests because gpu isn't used often. I'll do my best to test it though.
I haven't gotten my hands on the surface book yet so I am interested in seeing a comparison with other comparable devices. some poeple's work require them to travel often so these ultra thin convertibles is what is need. funny how on planes they let you use your tablet (surface) but not a laptop (surface with the keyboard on)
When you have the 13" comparison I think you may want to take down this comparison as the title will mislead normal readers to think macbooks in general are twice as fast. which I doubt the 13" mac will be twice as fast as the surface.
ok interesting. So that means a 6 core cpu would be 50% faster than a quad core at the same tasks you put them through here?
It depends, now you are getting a bit murky. Part of it is how well the software handles the cores. If the software only runs four threads for example, then it doesn't matter how many cores you have you won't get a huge improvement after four. You might get some (modern OS's and processors are designed to spread a thread across processors as much as possible, but they can only do so much.)
see the first video in the article, where adobe apps runs twice as fast with double the cores. hence my question.
If they are the exact same generation, clock speed and equivalent TDP; roughly yes*.
*as long as the software benefits from multi-threading in the first place.
This was a strange comparison, I love my macbook for photos and video but when I need raw power nothing beat customizable pc. http://www.boxxtech.com/
Strange enough that Microsoft themselves compared the SurfaceBook to the MacBook in their marketing and presentations?
They compared it to the 13ich one... you compared with the 15ich one which can pack more power... lololool getting your fuckin facts correct
This awkwardly unreasonable test must be extremely embarrassing for Lee and other Apple fanatics who think that funnily wrong and biased comparison is anything close to fairness, Shouldn't be used to recommend people anything but how not to seriously test anything.
I am not an Apple user! I have never owned a single Apple computer.