• After 15+ years, we've made a big change: Android Forums is now Early Bird Club. Learn more here.

Exynos or Snapdragon ?

Nobody knows. They are both quad core processors as far as benchmarks are concerned (the "Octa" is 4 A15 cores plus 4 slower A7 cores, with only one or other set used at any time). And unless it's a pure CPU benchmark you need to consider other elements of the system - we've seen often enough that different phones with the same processor and clock speed don't always produce the same benchmarks.

Most importantly, benchmarks have only a loose correlation with real world performance, so actually aren't really important. People geek on them because they are numbers which can be compared and look more "objective" than assessments of responsiveness, lag etc, but that doesn't mean that they are necessarily meaningful. Wait until we have samples of both and see whether it actually makes a difference, and if so to what aspect of performance.
 
Upvote 0
OK, I'm thoroughly confused. ARM Cortex-A7 vs. ARM Cortex-A8 vs. ARM Cortex-A9 vs. ARM Cortex-A15. Wow, so many different ARM architectures.

So what is the Snapdragon 600? Is it A7 or A15? What about the Octa-Core CPU that the US isn't going to get? Is that A15 or is it A7? Is there really that much of a difference between all of those architectures?

And I thought Intel was confusing with Nehalem, Sandy Bridge, Ivy Bridge, Haswell, etc.
 
Upvote 0
The "octa core" isn't octa core. It's dual quad-core: Quad A15 for heavy lifting, plus quad A7 to save power when doing light work. At any time it will be using either the A15s or the A7s, but never uses more than 4 at a time.

Snapdragon 600 is quad Krait 300 - another arm-compatible architecture, not identical to the A15 but with a number of similarities.

Other components (e.g. graphics) will also make a difference to performance.

For all of the bloggers who wrote yesterday "Samsung weren't saying, but I reckon it was a snapdragon because it wasn't as blazingly fast as I'd expect the Octa to be", I'll personally be surprised if it actually makes a real world difference to the user (at least in performance - it will be interesting to see how power use compares). And I suspect a lot of those who wrote that type of stuff didn't understand what the "octa" actually is...
 
Upvote 0
ARM licenses their architecture designs and others build them. So does Qualcomm. But unlike Qualcomm, there are many ARM implementations, but each Snapdragon is its own model number.

So, if I say, here's an ARM Cortex from Tegra and here's one from Samsung, and they're the same class (A9 as was used in 2012 for example), they'll be mostly the same cpu silicon because they were built under license - but they'll perform just a little differently because Tegra and Sammy roll a little differently.

The Snapdragons use their proprietary Krait CPU cores, not an ARM Cortex design, but they operate on the ARMv7 instruction set (not to be confused with the ARM A7 chip).

Confused? Don't be, you've seen this elsewhere for years - Intel vs. AMD. The AMD will run great, run Windows and your favorite apps because it uses the Intel instruction set, but it's not an Intel design.

The Krait cpu core is an A15-class processor in terms of speed and performance. It's not an A15, just like an AMD chip isn't a Pentium.

The A15 is generally considered to be almost double the performer of the A9 in many circles.

Because the S4 dual core is an A15 class, Samsung thought nothing of providing the dual core S4 in the US while providing the A9 quad core elsewhere on the SGS3.

Samsung somehow managed to break from ARM in one important aspect last year - independent CPU clocking.

The Snapdragon multicores have always run their cpu clocks independently. The A9 dual and quad cores from others ran all cpus at the same speed, whether they needed it or not - not power efficient. Samsung, beginning in 2012, started running their cpu cores independently, like the Snapdragon.

For that reason, some write-ups will call the Sammy chips "like A-whatever (9 or 15)" while others will simply call them "A9" or "A15."

Here's your ARM comparison chart from the horse's mouth -

Cortex-A Series - ARM

Likewise for the Snapdragon -

Snapdragon 800, 600, 400, 200 Processor Specs | Qualcomm

The Snapdragon 600 ought not be compared so strongly to the Snapdragon S4 Pro.

The 600 cores are model Krait 300, the S4 Pro is model Krait 200 cores. The Krait 300 provides 15% faster processing than the Krait 200 when running at the same clock speed.

ARM vs. Krait involves only the cpu cores.

It doesn't involve the GPU, DSP, ISP, media management or other cores - just the CPUs.

Both the Snapdragon 600 and S4 Pro use the Adreno 320 GPU, but it's clocked faster in the 600.

And I agree - benchmarks are pretty worthless for most uses.
 
Upvote 0
On that note, you notice how in many ways mobile OS's are becoming more demanding than their desktop counterpart? I played with a Tegra 3 on my friends's touchscreen laptop and it was super fast, yet a Tegra 3 running Android can be a bit laggy at times, like the HTC One X.

That was more likely caused by differences in the kernel, memory and io logic than the processor.
 
Upvote 0
OK... let's see if I can wrap my brain around all of the info you provided.

The Samsung Exynos Octa-Core CPU really is nothing but a 4 + 4 design; a quad-core ARM Cortex-A7 Class CPU and a quad-core ARM Cortex-A15 Class CPU smashed together onto the same silicon.

Whereas the Snapdragon 600 CPU that's going to be featured in the US version of the Samsung Galaxy S4 utilizes a Krait 300 CPU which is an ARM Cortex-A15 Class CPU. Right?

So if I'm right so far, then tell me this. Since the ARM Cortex-A7 Class CPU is supposed to be a power-saving CPU that's used a majority of the time and the ARM Cortex-A15 Class CPU is only supposed to be used when the uber CPU power is needed, how is this going to effect battery life in the US version when the US version only has the higher-end ARM Cortex-A15 Class CPU?
 
  • Like
Reactions: EarlyMon
Upvote 0
OK... let's see if I can wrap my brain around all of the info you provided.

The Samsung Exynos Octa-Core CPU really is nothing but a 4 + 4 design; a quad-core ARM Cortex-A7 Class CPU and a quad-core ARM Cortex-A15 Class CPU smashed together onto the same silicon.

Whereas the Snapdragon 600 CPU that's going to be featured in the US version of the Samsung Galaxy S4 utilizes a Krait 300 CPU which is an ARM Cortex-A15 Class CPU. Right?

Perfect-a-mundo! :)

So if I'm right so far, then tell me this. Since the ARM Cortex-A7 Class CPU is supposed to be a power-saving CPU that's used a majority of the time and the ARM Cortex-A15 Class CPU is only supposed to be used when the uber CPU power is needed, how is this going to effect battery life in the US version when the US version only has the higher-end ARM Cortex-A15 Class CPU?

The answer to that is no one really knows yet.

And we don't know if the A7 quad will run the majority of the time (referring to when the phone is in actual use, not sleeping, for this part).

And because of task scheduling, we don't know when it will switch from all four A7 cores to just two of the A15 cores (not all cores need to run, so that's potentially a valid situation).

Did Samsung tune it for power savings, running the small quad most of the time? If they did, will that make people happy? (And that's like - way slow compared to either SGS3 last year, ok.)

Or - is it going to be tuned for performance?

~~~~~~~~~~

Sleeping is another matter - ideally, sleeping ought to run for what feels like forever, if it can run just one of A7 cores at its minimum clock speed.

Some people's phone use is for sleeping the majority of the time, others, not so much.

~~~~~~~~~~

If Samsung hits it out of the park, then you'll see very little or no difference as a user in performance between the two models, and the dual quad will get better battery life.

If they goof, then the Snapdragon 600 model will be better.

Anyone who says they know today is really guessing on that. ;)
 
Upvote 0
I was wondering why the A7 consumes less power than the A15 in the octo-core? Does it have less transistors, and as a simpler design use less power? I thought that each iteration of CPU technology was more efficient than the last, or is that because they use more miniaturized silicon? And so fewer transistors on miniaturized silicon uses less power than the newer cpu design? Did I just answer my own question?
 
  • Like
Reactions: EarlyMon
Upvote 0
I was wondering why the A7 consumes less power than the A15 in the octo-core? Does it have less transistors, and as a simpler design use less power? I thought that each iteration of CPU technology was more efficient than the last, or is that because they use more miniaturized silicon? And so fewer transistors on miniaturized silicon uses less power than the newer cpu design? Did I just answer my own question?

Toss in lower clock speed on top of that and yep, you did. :)
 
  • Like
Reactions: davoid
Upvote 0
but... then is it only miniaturization that makes the newer cpu designs more efficient, and not improved memory controllers etc? I ask because it's heard said that newer processors can do more work for the same clock speed. But if that is the case, why not use newer processors at lower clock speeds rather than the older processors? Oh dear, I started off well...
 
  • Like
Reactions: EarlyMon
Upvote 0
but... then is it only miniaturization that makes the newer cpu designs more efficient, and not improved memory controllers etc? I ask because it's heard said that newer processors can do more work for the same clock speed. But if that is the case, why not use newer processors at lower clock speeds rather than the older processors? Oh dear, I started off well...

No no - great questions! :)

Smaller structure geometry in silicon does indeed equate exactly to higher performance and lower power consumption.

And - we know a lot of what to expect by examining the manufacturing processes.

Let's use some numbers to sort it all out.

Both the S4 Pro and the Snapdragon 600, for example, are made with the same 28 nanometer manufacturing process.

The 600 is 15% faster because of its design, not the transistor size. Let's see the possible outcomes to know what that means, and assuming the same number of transistors (close enough for this discussion's purposes) -

I could do the same job on the 600 as the S4 Pro but run the clock 15% slower, and I'd get a power savings with no performance difference.

Or, I could run my 600 and S4 Pro at the same rates, but get 15% more processing power out of the 600 without using any extra battery.

See how that works?

Ok, let's use that, turn it around and see how that explains the Samsung chip -

Both the A7 and A15 quads are made on the same silicon, probably no bigger than the result of a 32 nanometer manufacturing process.

The A15 will have minimum and maximum possible clock speeds - ditto for the A7, but they will both be lower than for the A15.

For the A7, I don't care a lot about design efficiency - that affects performance. I don't care about performance there - I care about power efficiency. If I want performance, I'll go to the A15 quad. If I want power saving, I go with fewer transistors, running slower than my A15, all made the same size per transistor.

If I have two chips and run one at half the speed of the other, I'll save a lot of power. If I cut the number of transistors in half, I'll double my power savings (not saying that those are the exact numbers between the A7/A15, just laying it out to clarify).

And that's what the A7 is all about for power efficiency.

This approach actually appeared last year in the Tegra 3. That used an A9 quad plus a single slow, small core for the power savings.

Like the new Sammy, that too was a big/little design.

Some complained that the transition on the Tegra 3 from the single tiny core was kind of great when jumping to the A9 or vice versa.

And that would have gotten worse with an A15 quad.

So, Samsung evidently decided to combat that with a quad of smaller cores. I don't know, but I could see that it might scale up from 1 to all 4 cores on the A7 and then switch over to the A15 - just as one design example.

I'll stop there - does this help any??
 
  • Like
Reactions: davoid
Upvote 0
So really this Octa - Core stuff is only marketing BS.

Yep.

An octacore would be 8 identical processors all running at once, same rules as a dual core or a quad core.

Both Intel and AMD make octacores. Pretty neat stuff.

And here we have a dual-quadcore where you don't run all 8 at once, so it's not really an octacore.

But according to Samsung marketing, 2x4=8 and 8=8 and who is going to argue with Samsung?

The press? Not likely! :D

They're getting a lot of mileage out of this. People are excited - OMG, an OCTAcore! Who can blame them? That would be exciting.

But - just as many people are saying - What the heck do I need an octacore for?? :thinking: I want better battery life. :(

And if people are lucky enough to find half-way decent articles, and then read up, and then click through to the Samsung links explaining what I just did, I'd bet they'd agree with me that Dual-Quad is a really great thing to market because if you like Quad, and you like power, maybe you'll like Dual-Quad.

But no.

That's all hard to believe when Samsung says up front, it's an octacore, in an age where people know what octacore means.

And here we are. :eek:
 
Upvote 0
I guess my next question is what causes these improvements in chip technology? Obviously technology progresses, but if someone had an idea in 2009 of how the Snapdragon 600 should run on a 28nm process, what would actually prevent them from mass producing such a chip? Even if you give 1 year to get assembly lines ready, what is preventing (and how) the manufacturing of such a chip until a few years later?
 
  • Like
Reactions: EarlyMon
Upvote 0
The Nexus One was built using 65 nm technology, 45 nm appeared on the HTC Thunderbolt, 32 on the Sammy Exynos, 28 on the Snapdragon S4, and finally 22 18 for Samsung memory last year.

Each technology reduction is incredibly difficult and insanely expensive.

It requires changes in manufacturing tooling, processes, materials, testing (failure modes change as size is reduced), vendors and personnel.

The initial wafers from which chips are cut costs millions of dollars, and then you get to spend more refining everything until the yields (ratio of good vs attempted parts) are acceptable.

The stages of reduction over the years is laid out in a roadmap followed by semiconductor manufacturing industry, and if any steps could have been cut out, they would have been.

Then on top of that comes design layout of the processors. At GHz speeds, the internal data paths are screaming along like high frequency antennas. So speed increases can't just happen without a whole regime of engineering just for that one effect.

Then comes the design improvements for the other cores and glue logic (last year's chips weren't dual or quad core, that was just the cpu count, anywhere from 6 to 9 cores was reality).

And I've only scratched the surface.

You could have imagined an HDTV in 1960 but you couldn't have built then what we have today.

Comparing to SoC processor technology of 4 years ago, it's actually a good analogy.

Hope this answers a little. :eek:
 
  • Like
Reactions: Xxmen and davoid
Upvote 0
Interesting. So the size keeps shrinking. I heard that eventually we'll reach a point where the chips can't be shrunk any further because the electrons will act erratically between transistors or something like that. How small would that be, are we close?

We're already operating by exploiting quantum effects.

We'll never fly, we'll never get to the moon. ;) :)
 
Upvote 0
Frankly, I would be happy with A6X performance. MAMEReloaded 1.39 on my iPad 4 destroys my S3, Excite 7.7 and TF300. Not to mention others recent tablets tested, in spite of the iPad pushing a LOT more pixels. MAME is a very good app to test how powerful the actual CPU is.

If the Snap 600 is faster than the A6X, I will be happy.

Added: This of course assumes the new radio improves the weak 3G performance and battery sucking demon ways with low signal. Worse of any device I have owned or tested. Still, the S3 is my favorite. Call range is great, in spite of the data radio issue and I am mainly in decent 4G areas- but not always. I turn the data off in weaker areas to avoid battery drain when needed.
 
Upvote 0

BEST TECH IN 2023

We've been tracking upcoming products and ranking the best tech since 2007. Thanks for trusting our opinion: we get rewarded through affiliate links that earn us a commission and we invite you to learn more about us.

Smartphones