Geekbench 3 Sandy Bridge v.s. Apple Cyclone IPC comparison

[Nothingness]

It uses ICC with Visual Studio, Visual Studio by itself will give generic vectors for AMD/Intel/VIA. Both AMD and VIA only get scalar instructions though so it is pretty obvious which compiler is being used.

Given how close FP results are, you are again wrong:
http://browser.primatelabs.com/geekbench2/compare/2295521/2300059

 

[ShintaiDK]

Possibly also pricing Intel has published:

http://ark.intel.com/products/76760/Intel-Atom-Processor-Z3770-2M-Cache-up-to-2_39-GHz

$37 isn't that bad but it's still probably a decent bit more expensive than most competitors out right now. Actual big OEMs may be paying less.

Nobody pays listprices when buying from Intel.

I dont think MC got any special deal on 4670Ks, even tho they sell them 40$ cheaper. I am sure there is still a good profit margin left for MC.

http://ark.intel.com/products/75048/Intel-Core-i5-4670K-Processor-6M-Cache-up-to-3_80-GHz
http://www.microcenter.com/product/413251/Core_i5_4670K_34GHz_Socket_LGA_1150_Boxed_Processor

 

[ShintaiDK]

Apple is nowhere near x86 performance with their ARM chips, not even remotely close.

 

[Nothingness]

Apple is nowhere near x86 performance with their ARM chips, not even remotely close.

You know icc is known as the SPEC compiler among benchmarks pros?

As far as Polyhedron goes, icc leads that much because it uses auto-parallelization. If this feature isn't used icc is slower than absoft.
Ref : http://polyhedron.com/pb05-lin64-f90bench_SBhtml

Anyway I don't see the relation with Apple results. Do you have SPEC or Polyhedron scores for A7?

 

[BallaTheFeared]

I agree
also can it play CRYSIS @ 2560 x 1440 max settings or maybe I missed that bench mark.
-it would like a smart car passing you on the highway doing 220 mph on a windy day.er not going to happen.

I agree as well, like the people saying it will take over in the mac book air...

Unless something drastic happens, the mac book air would therefore be running iOS...

Yeah right :|

 

[Nothingness]

I agree as well, like the people saying it will take over in the mac book air...

Unless something drastic happens, the mac book air would therefore be running iOS...

Why would that have to be iOS? I guess Apple has already ported MacOS X to ARM (and in the same way I wouldn't be surprised to learn they have iOS running on x86 ).

 

[jhu]

I agree as well, like the people saying it will take over in the mac book air...

Unless something drastic happens, the mac book air would therefore be running iOS...

Yeah right :|

It's likely that Apple has Mac OS X already compiled for ARM as a "sanity check," just like they had an x86 build during the years they were using PowerPC.

There's also already a rudimentary iOS for x86: the iDevice emulator that's installed when XCode is installed.

 

[Magic Carpet]

Nobody pays listprices when buying from Intel.

I dont think MC got any special deal on 4670Ks, even tho they sell them 40$ cheaper. I am sure there is still a good profit margin left for MC.

http://ark.intel.com/products/75048/Intel-Core-i5-4670K-Processor-6M-Cache-up-to-3_80-GHz
http://www.microcenter.com/product/413251/Core_i5_4670K_34GHz_Socket_LGA_1150_Boxed_Processor

I don't think they make any profit on 4670K at all. It's part of their strategy. They want you to come in store, using 4670K as bait.

 

[shady28]

I agree as well, like the people saying it will take over in the mac book air...

Unless something drastic happens, the mac book air would therefore be running iOS...

Yeah right :|

You do realize that iOS and OS X are both built on top of Darwin, right? At the core they've been the same since the first iPod touch.

 

[Pilum]

I don't think they make any profit on 4670K at all. It's part of their strategy. They want you to come in store, using 4670K as bait.

Yup, you never worked retail and got a chance to look at the buying prices of your store compared to the selling prices.

Read this and then look at the gross margins here. Taking 36% gross margins, average computer/software markup is ~56%. Which leaves plenty room for a ~20% discount while still making a profit. Of course MC "special item" pricing mainly serves to lure customers into the store, but that doesn't mean they don't make a profit if only selling the discounted item. And on average they will sell additional items, which offsets the reduced profits of the special items.

 

[ShintaiDK]

Anyway I don't see the relation with Apple results.

Clang is heavily backed by Apple. Its essentially the same as using ICC on Intel.

Instead subpair compilers was used with x86.

While Apple designed a really good chip. its still no x86 competitor in any way.

Apple is just able to control and optimize its entire chain due to its monopoly around the platform.

 

[Magic Carpet]

Yup, you never worked retail and got a chance to look at the buying prices of your store compared to the selling prices.

Read this and then look at the gross margins here. Taking 36% gross margins, average computer/software markup is ~56%. Which leaves plenty room for a ~20% discount while still making a profit. Of course MC "special item" pricing mainly serves to lure customers into the store, but that doesn't mean they don't make a profit if only selling the discounted item. And on average they will sell additional items, which offsets the reduced profits of the special items.

Well, I had a chance to look at the buying prices at Tiger Direct and the % markups weren't anywhere near the numbers you have mentioned above. Especially on the 3770K/4770K parts, of course they make some money, but selling a 4670K for ~$40 less than competition is on the borderline of profitability. There is a good reason, why only MC is selling it at such a big discount on a regular basis. Unless, they have a special deal with Intel (which I doubt), they aren't making much money on it, if at all. Same in Europe, the buying prices for dealers, are quite high (compared to other stuff).

But as a business model, it must be working brilliantly for MC, I know, once you are past the doors, change of plans happen and you end up with a basket full of other goodies. Come on, I didn't come here only for cpu, right? Let's make this moment memorable and mark it with a purchase of other items. And the longer you have to drive to the nearest store, the more business for MC :biggrin:

 

[Enigmoid]

Yup, you never worked retail and got a chance to look at the buying prices of your store compared to the selling prices.

Read this and then look at the gross margins here. Taking 36% gross margins, average computer/software markup is ~56%. Which leaves plenty room for a ~20% discount while still making a profit. Of course MC "special item" pricing mainly serves to lure customers into the store, but that doesn't mean they don't make a profit if only selling the discounted item. And on average they will sell additional items, which offsets the reduced profits of the special items.

They are for sure losing money on that one. Your link shows gross margins of ~36% but PROFIT margins of only 3-4%. Selling that product gets them ahead but when you factor in costs they are behind. MC loses money on the CPUs (the customer pays less than it costs them to buy,process and sell the cpu) but then makes it up on other sales.

 

[Headfoot]

Here are the facts:

Haswell is wider than Cyclone. We don't know how much, but there is no way it can be as wide given the transistor counts. Haswell is the 4th generation of a desktop architecture designed for single threaded performance first, moving into power consumption as they moved through generations. Intel is the largest processor company in the world. Intel's process is the most advanced in the world. Apple's Cyclone is the 2nd custom design ever from their silicon team. It is a smartphone architecture scaled/derived/heavily modified from A9 and whatever changes necessary for ARMv8 (i.e. nowhere near as wide as Haswell). Haswell has vastly more memory bandwidth than Cyclone in these tests. Geekbench 2 was known to be poorly optimized/targeted and gave poor results at some point in its lifecycle.

Now given all of these facts, there are two options:
1. Despite all of the above, Cyclone is faster per clock than Haswell and Geekbench 3 is completely fair cross-architecture
2. Geekbench 3 is poorly implemented on x86 compared to ARM

Unless you're delusional, 1 is simply less likely than 2. Not impossible. Just very unlikely. Are you really betting that Intel is the one who screwed up given the choice of saying either Intel, the multibillion dollar bleeding edge market leader; or Geekbench, the couple guy benchmarking outfit with a bad reputation for accuracy, executed poorly.

 

[Nothingness]

Yes, Intel never screwed up in the past, so it must be that Geekbench is biased against Intel, sure. And Apple is a small company with limited money. And only big companies succeed.

I wonder who is delusional.

 

[ShintaiDK]

Yes, Intel never screwed up in the past, so it must be that Geekbench is biased against Intel, sure. And Apple is a small company with limited money. And only big companies succeed.

I wonder who is delusional.

When something is too good to be true, it usually is.

 

[jhu]

Here are the facts:

Haswell is wider than Cyclone. We don't know how much, but there is no way it can be as wide given the transistor counts. Haswell is the 4th generation of a desktop architecture designed for single threaded performance first, moving into power consumption as they moved through generations. Intel is the largest processor company in the world. Intel's process is the most advanced in the world. Apple's Cyclone is the 2nd custom design ever from their silicon team. It is a smartphone architecture scaled/derived/heavily modified from A9 and whatever changes necessary for ARMv8 (i.e. nowhere near as wide as Haswell). Haswell has vastly more memory bandwidth than Cyclone in these tests. Geekbench 2 was known to be poorly optimized/targeted and gave poor results at some point in its lifecycle.

Now given all of these facts, there are two options:
1. Despite all of the above, Cyclone is faster per clock than Haswell and Geekbench 3 is completely fair cross-architecture
2. Geekbench 3 is poorly implemented on x86 compared to ARM

Unless you're delusional, 1 is simply less likely than 2. Not impossible. Just very unlikely. Are you really betting that Intel is the one who screwed up given the choice of saying either Intel, the multibillion dollar bleeding edge market leader; or Geekbench, the couple guy benchmarking outfit with a bad reputation for accuracy, executed poorly.

Should be easy enough to find out. Someone should get the binaries and examine the code emitted and try to see what it's doing. Would be easier if geek bench was open source, but that's what happens when you use non-open source programs.

 

[Nothingness]

When something is too good to be true, it usually is.

You mean like Bay Trail being that great as many describe it?

 

[Nothingness]

Should be easy enough to find out. Someone should get the binaries and examine the code emitted and try to see what it's doing. Would be easier if geek bench was open source, but that's what happens when you use non-open source programs.

I have been doing that, and I have yet to see anything obvious. With AnTuTu it was pretty clear icc was cheating (I know people in denial will come and cry, but I don't care, I know this is right).

 

[Eug]

From a n00b perspective...

Regardless of who is the IPC WINNAR! here, you can bet Intel isn't just sitting back and saying "We have the best chips, so we can just ignore Apple". They're probably dissecting every last bit of the A7 in their labs as we speak.

If someone had told me 5 years ago that there would be serious threads in 2013 about Apple's chip designs vs. Intel's, I would have laughed. In this short period of time Apple has oved from being a chip customer to arguably a leading ARM chip designer and supplier... even though their supply only goes to themselves.

 

[ShintaiDK]

I have been doing that, and I have yet to see anything obvious. With AnTuTu it was pretty clear icc was cheating (I know people in denial will come and cry, but I don't care, I know this is right).

If ICC is cheat. What is clang then on IOS?

 

[Nothingness]

If ICC is cheat. What is clang then on IOS?

Has it been proven that clang is cheating? It might, but I have seen no evidence yet.

 

[jvroig]

If someone had told me 5 years ago that there would be serious threads in 2013 about Apple's chip designs vs. Intel's, I would have laughed. In this short period of time Apple has oved from being a chip customer to arguably a leading ARM chip designer and supplier... even though their supply only goes to themselves.

All true. Although remember that everything here (industry expertise, manufacturing, design, R&D) depends pretty much just on one thing - money to burn.

Disregarding Apple's current state for the meantime (whatever state it is):
1.) Why does Intel have the best fabs available? They have the most money. They spend a lot more on fabs, even investing in tool vendors for tooling development.
2.) Why can they afford to have better design teams and more advanced tech in their processors? They spend a lot more money than AMD - pay people higher salaries for all the industry expertise they need, they have much more of these people, and these engineers get a lot more in the way of tools to help facilitate their specific jobs.

It's how the "R&D magic" happens -> The dollar investment is generally directly proportional to resulting tech. Bigger investment, cooler tech.

And Apple is sitting on a truckload of cash (how much is it now? 100-150B?), and we've all heard reports from way back about how they were scooping up (pirating) the needed expertise to design processors. Again, nothing magical, it's not because they are Apple. It's simply because the effort/initiative needed money, and Apple had money and were willing to use it. And so we have the current iteration of their processor design effort.

Whether this A7 is really as good as the most positive reviews of it seem to make it, or it's just being blown way out of proportion due to excitement, I'm glad Apple decided to play this game. Making a chip that won't immediately be embarrassed by anything from Intel requires a lot of money and dedication. Apple is one of the few companies that have both in spades.

Realistically, AMD has no hope of catching up in any material way to Intel (again, it's really just a matter of R&D budgets), so they'll always be relegated to "cheaper second-source" or gain a few niche markets using their GPU tech as leverage. But that's that, and no more than that. A real Xeon competitor? Nope. Desktop competitor? Nope. Laptop / mobile? Still nothing. There is nothing they are doing (other than their integrated GPU tech) that Intel cannot do much better. And because of the reality of the available R&D budgets for both companies, there is simply no way they will magically catch up.

For Apple, the cards are completely different. They are literally sitting on over a hundred billion dollars, and have a very loyal, very dedicated user base. AMD's problem of limited cash is not going to be a problem for Apple in the next 6 years even if they decided to stop selling products now (based on a ~$20B 12-month total operating expenses; I sort of remember an ~$18B figure some time ago, just rounded up from there).

I don't even own any Apple product (I do have 3 Android devices of wildly different quality and sizes), so it's kind of funny to me that I'm writing what seems like a very pro-Apple piece. I'm not really so much as supporting Apple, I'm just more excited at the possibility that a really serious competitor in the CPU area might be getting born before our very eyes. I love Intel, they have awesome products, but it won't break my heart if another cash-rich corp decided they want to play the game too for whatever purpose, instead of just having "competitors" who can't afford anything close to what Intel spends annually for R&D.

I don't have data at hand, so I can't comment on how profitable the CPU designing venture is for Apple (honestly, I don't even understand right now why they bothered, but then again I don't really follow Apple). I hope it's very profitable, so they'd have the inclination to continue and maybe even expand their efforts.

Do I still have time for an analogy? I promise it won't be a car analogy (I know even less about cars than I do about Apple products). I'm going to use boxing.

Intel vs AMD is like a boxing match between Mayweather (Intel) vs an amateur boxer (AMD). I simply see no outcome that isn't an event that Mayweather dictated himself (be it a quick finish or a drawn out fight for show).

But Intel vs Apple? Now that's more like Mayweather vs Pacquiao. Fair fight, big names, no doubt a decent and exciting match - if only Pacman decided he wanted to play.

 

[Eug]

I don't have data at hand, so I can't comment on how profitable the CPU designing venture is for Apple (honestly, I don't even understand right now why they bothered, but then again I don't really follow Apple).

Well, you can't deny the huge advantage of being able to tailor your chips to your products, instead of the other way around.

If they hadn't done this, for example the iPhone 5S would likely be sporting a less efficient (for mobile applications) 32-bit quad-core CPU without built-in cryptography support and a different image processor, which could mean a less elegant fingerprint sensor solution and either a different set of features for the camera, or else a set of features that taxed the general purpose CPU core much more and wasted more battery life.

P.S. Money isn't everything. I don't know how much money MS has spent on mobile, but look at where they are currently. And Blackberry just took a write down of a billion dollars last quarter.

 

[jvroig]

P.S. Money isn't everything. I don't know how much money MS has spent on mobile, but look at where they are currently. And Blackberry just took a write down of a billion dollars last quarter.

For non-engineering initiatives (like aesthetics, graphic design, brand management, predicting the taste/preferences of consumers, etc) it takes more than money. Apple has Jonny Ive (and Steve Jobs before he died) to manage their design / aesthetics, and Steve Jobs had great influence managing the brand.

But for engineering initiatives (and I should have made that more clear, I just assumed I was already clearly talking of engineering, but reading it now, I guess I wasn't), it's that simple. More money, better result. If an an engineering outfit outspends their rival in R&D by double, triple, quadruple or more, they will consistently deliver a superior product.

No matter how much of a visionary and a MacGyver the David team is, the Goliath team will always produce a better, faster, more efficient product if they have at their disposal a surplus of hundreds of millions of dollars in R&D compared to the scrawny budget of their David-like competitor. It's simply the way of science and engineering in our modern world. And that's the reality that has gripped the Intel vs AMD "fight", and one that won't be the case for Apple.