Asus Transformer Drops Quad-core In Favor of Dual-core

MrSeb writes with this news from Extreme Tech: "In a move that will shock and disgust bleeding-edge technophiles everywhere, Asus has announced at Mobile World Congress 2012 that its new Transformer Pads — the high-end Infinity Series — will use the recently-announced dual-core Qualcomm S4 SoC. The critically acclaimed Transformer Prime, the Infinity Series' predecessor which was released at the end of 2011, used the quad-core Nvidia Tegra 3. Why the sudden about-face? Well, the fact that quad-core processors don't really have a use case in mobile devices is one reason — but it doesn't hurt that the Krait cores in the S4 are significantly faster than the four Cortex-A9 cores in the Tegra 3, too. The S4 is also the first 28nm SoC, while Tegra 3 is still on 40nm, which means a smaller and cheaper package, and lower power consumption to boot. The S4 is also the first SoC with built-in LTE, which was probably a rather nice sweetener for Asus." The Snapdragon S4 "Krait" CPU is still a bit shrouded in mystery as far as hard specs (Qualcomm has never been one to release docs), but it appears to be similar to the Cortex-A15 in performance; how they stand up to Intel's new Medfield designs remains to be seen.

Comments

Right tool for the job

[Vollernurd]

In other words after carefully considering all their options and went with the one that offered the best overall package, whilst keeping the price point competetive? Not nerd willy-waving, then? Jolly good.

Re:Right tool for the job

[Phoghat]

"Willy-Waving??? performance @ price point has always been my aim.

Re:Right tool for the job

[gsslay]

I have to agree with the article. I am shocked and disgusted beyond measure.

I'm just not sure why. Maybe I need further desensitising against run of the mill tech news.

Re:Right tool for the job

[noh8rz2]

somebody please help me understand this - who cares what's under the hood? quad core, dual core, snappy dragon, spanky man, whatever. Why not judge the phones by performance? I feel like we're stuck in the mindset of the 90s, where my P3-600mhtz is bigger than your P2-400mhtz.

Re:Right tool for the job

[jones_supa]

We certainly need to start using some universal processor benchmark score. Even though it wouldn't always produce completely comparable results, it would be much more useful than MHz (which on x86 desktops at this point tells almost nothing) or throwing around these "snapdragons".

Re:Right tool for the job

[gmhowell]

somebody please help me understand this - who cares what's under the hood?

Every sperg who comes out in an article discussing Apple's A5 or A4 seems to bring up crap like this rather than focusing on the user experience.

Re:Right tool for the job

[noh8rz2]

what's a sperg? Also, what do you think the new iPad will be? a6 or a5x?????????

Re:Right tool for the job

[gmhowell]

Sperg is a perjorative term for someone with Assburger's syndrome, generally of the 'internet diagnosed' variety. As far as that other shit? I don't care. All I care is how well the iPad 3 gets the job done.

Re:Right tool for the job

[noh8rz2]

Oh I see. I thought you meant, omg I'm going to sperg all over your face!

Core count obsession

[Cryacin]

Why is everyone obsessed at the number of cores? The more processors you ahve, the more complex scheduling your apps needs to perform to actually work faster. It's better to hav ea single core that is twice as fast, than two cores running in parallel.

Re:Core count obsession

[snowgirl]

Why is everyone obsessed at the number of cores? The more processors you ahve, the more complex scheduling your apps needs to perform to actually work faster. It's better to hav ea single core that is twice as fast, than two cores running in parallel.

Pfff... actually, the Tegra 3 has five cores, four of them are high-performance, and one is high-efficiency. The CPU is designed to shutdown the four cores for almost nearly everything, and just use the high-efficiency core in order to save on battery life.

So seriously, most of the time, the number of cores doesn't even matter, because unless you're playing a high-end game, the cores won't even be woken up.

Re:Core count obsession

[Theophany]

Everybody knows more cores = more bitches. Nobody wants to be 'that single (core) guy' who is forever alone at the bar.

Re:Core count obsession

[Psiren]

So seriously, most of the time, the number of cores doesn't even matter, because unless you're playing a high-end game, the cores won't even be woken up.

So, unless I was buying a tablet specifically to play high end games on, why would I want to spend money on CPU cores that are going to sit there doing nothing? Surely a dual core CPU is a better move?

Re:Core count obsession

[Mithent]

That's generally true, but it's increasingly difficult to get large performance improvements on a single core, particularly on the desktop (mobile was lagging for a long time, though it's catching up now). It's relatively easy to add extra cores, and they provide a substantial boost in performance - provided applications can make use of them, of course.

Having two cores is immediately pretty beneficial regardless of application multithreading, since it means that any background grumblings can be shunted off to a second core and don't get into contention with the foreground task. There's less case for four on mobile right now because few programs are written to use that many threads, but I'm sure we'll be moving that way eventually.

Re:Core count obsession

[mikael_j]

More cores means better multitasking since threads can run in parallel. Also, even for handheld devices you are unlikely to find, for example, a single-core CPU that is four times faster than each core of a quad-core CPU.

Another major advantage of multi-core systems is if a poorly written piece of software locks up it is highly likely to also be single-threaded and your system will chug along nicely despite the misbehaving program, allowing you to kill the process (by comparison, on a single-core system you're likely to suffer through five minutes of waiting for the system to respond before you are able to kill the process). Sure, in an ideal world this wouldn't happen but when it does happen it's nice to not be locked out of your system because of a single process misbehaving.

Re:Core count obsession

[errandum]

For 99.9% of the games, I'd say yes, if the Dual Core represents better better overall performance and compatibility with other things (like LTE) out of the box. The tablet is not made for that 0.1% that only wants to play that one game that makes use of the 4 cores... It's made for those that use those tablets for what they seem to be designed for (with a dock and everything...): Light work stations and media centers (that give it the 18 hour batteries.

I, for one, applaud this move. Core hype will get you nowhere in the long run.

Re:Core count obsession

[FrootLoops]

The more processors you ahve, the more complex scheduling your apps needs to perform to actually work faster.

Well, sort of. All you need to do is make sure you run each process long enough between breaks so that the ratio of scheduler time to actual processing time is small (fractions of a percent, say). You need a scheduler even on a single core to get multitasking to work anyway. Getting coders to take advantage of multiple cores is the actual problem here since writing bug-free parallel code is often very hard.

It's better to hav ea single core that is twice as fast, than two cores running in parallel.

Again, sort of. For tasks that do not parallelize well, obviously a single 2x-speed core is better than two 1x-speed cores. Power usage increases non-linearly with clock speed, though, so those two slow cores will use significantly less power than the single fast core, all else being equal. On a mobile device that savings might be worth the performance hit.

That said, four cores seems excessive on a tablet. When are you ever running four different CPU-bottlenecked processes/threads? A few games?

Re:Core count obsession

[Elrond, Duke of URL]

The whole "core" obsession on mobile devices seems to be nothing but marketing talk. At least, as far as I have been able to determine.

I have a Droid 3 which has a dual-core CPU and using System Tuner I found that the second core was always shown as "offline". Doing some research online I found that the second core is kept offline to preserve battery life. Supposedly, it only comes online if the load is particularly high.

But, no matter what I did on the phone, I could never get the second core to come online. Using one of the tweaks available in System Tuner, I can apparently force both cores to be online all the time. However, the second core is still shown as offline and I still can't seem to get it to come online via high usage. Also, battery life doesn't seem to have changed.

So, this wonderful second core seems to be entirely useless and nothing but an item for the marketing checklist on the advertisements.

Bah...

Re:Core count obsession

[Anonymous Coward]

Or... System tuner is reporting the status of the second core incorrectly? Surely it's wrong somewhere: it is incorrectly reporting the second core as offline or its force option is not working correctly.

Re:Core count obsession

[pinfall]

Everybody knows more cores = more bitches. Nobody wants to be 'that single (core) guy' who is forever alone at the bar.

Not all cores are coreated equal.
I have a single socket, single upper layer core, which combines the power of 128 microprocs in an 8 core assembly for a whopping 1024 hidden 1mhz cores - almost 1 Ghz of raw power!
# of cores ftw!

Re:Core count obsession

[billcopc]

The average user doesn't have the slighest idea how threading works nor why having more cores might be overkill. To them, it's just yet another number that must be increased. They look to us geeks, with our multi-core and multi-socket systems, and figure that's where they want to be once the prices come down. They're like kids emulating adults, and just as stubborn when I try to explain that the average human does NOT need a 12-core workstation with 48 gigs of Ram. It's hard enough convincing them that a Gigabit router won't make their DSL go faster than a 10/100 one, and they go absolutely retarded when they find out I use 10G fibre NICs.

This is what I tried to explain to my not-so-technical friend, who would ask me if the 4-core tablet was better than the 2-core one, and then ignored anything I said. It's a tablet, you don't multitask much on it. You're not running 50 torrents in the background, while your virus scanner eats a whole core protecting you from yourself, and trying to play a Youtube 1080p walkthrough on your second display while you follow along in Starcraft II on the main screen. It's a fuckin' tablet. One app at a time. If that app is smart enough to offload background tasks to a 2nd core, I'm already impressed. It's a very different computing experience from a desktop PC, and even there, most people get by just fine with a dual-core desktop. The mere fact that almost every computing device today has a dedicated GPU, it's like an extra "core" right there, in that it frees up the main CPU to do something else.

Re:Core count obsession

[ozmanjusri]

So seriously, most of the time, the number of cores doesn't even matter, because unless you're playing a high-end game, the cores won't even be woken up.

I'm more interested in video/image editing than games. Those extra cores help a lot with tasks like that.

Re:Core count obsession

[21mhz]

More cores means better multitasking since threads can run in parallel.

In my experience, many of the cases where people thought they need to use threads were really cases of doin' it wrong.
When you have properly designed asynchronous APIs, inter-process communication, a decent kernel scheduler, and a GPU to offload graphics to, you can do non-blocking UI even on a single CPU, and the difference between single and multiple cores only manifests in performance improvements, often fairly marginal, and hitting different bugs (which good software should not have, of course).

Also, even for handheld devices you are unlikely to find, for example, a single-core CPU that is four times faster than each core of a quad-core CPU.

I've been searching for someone to answer to a simple question: what are the tasks at which you feel your high-end smartphone should be faster, that are not attributed to things like network roundtrips where a faster proc is irrelevant? "Four times faster than good enough" just does not equal to "four times as good" to me, especially if it comes with a larger price point.

Another major advantage of multi-core systems is if a poorly written piece of software locks up it is highly likely to also be single-threaded and your system will chug along nicely despite the misbehaving program, allowing you to kill the process (by comparison, on a single-core system you're likely to suffer through five minutes of waiting for the system to respond before you are able to kill the process). Sure, in an ideal world this wouldn't happen but when it does happen it's nice to not be locked out of your system because of a single process misbehaving.

Unless we are talking of processes with privileged scheduling which should be part of the system and therefore should be well-written, which modern smartphone OS can't preempt a process that has locked up?

Not two years ago everything in the smartphone world worked nicely on a single core, and everybody except extreme performance junkies were happy with that (can't vouch for Android users, their pain tolerance threshold appears to be higher), what happened since then?

Re:Core count obsession

[billcopc]

On a tablet, I don't see us needing more than 2 cores for the time being, because of how we use them as toys. It's kind of hard to use a desktop or even laptop while laying on the couch, at least without some type of foldable stand or unnaturally ergonomic beer belly. As a stupid little "I don't wanna get off the couch" internet machine, a tablet kicks ass. For that basic usage, 2 cores is more than enough.

If/when we start using them as true laptop replacements, with a keyboard and stand/dock, that's when we'll start finding uses for more cores. The day I whip out a tablet at the bar, unroll a keyboard and fire up Eclipse for some billable hours, is the day I'll sing praises for a quad or hex core tablet.

Re:Core count obsession

[21mhz]

It's not the number of cores that matters, what matters is how you use them.

Re:Core count obsession

[Daniel Phillips]

It's a tablet, you don't multitask much on it.

Since when?

Re:Core count obsession

[poetmatt]

What are you talking about? Two years ago phones ran like absolute pieces of crap. Even one year ago, for the most part. There was constant stuttering, things weren't that smooth, etc. While apple's UI was smooth, actually running things at first? Not so much. Funny how people can't remember two years ago well.

Only since december or thereabouts have phones actually been getting towards powerful enough to handle everything smoothly.

Four processors on a smartphone is fairly trivial in cost, so no it doesn't include a significantly larger price point. At the end of the day we're talking a $12 chip versus a $25 chip (maximum) for example. Don't think fudsters wouldn't hype that as "100% more expensive processor!" though.

Re:Core count obsession

[James_Duncan8181]

I've been searching for someone to answer to a simple question: what are the tasks at which you feel your high-end smartphone should be faster, that are not attributed to things like network roundtrips where a faster proc is irrelevant? "Four times faster than good enough" just does not equal to "four times as good" to me, especially if it comes with a larger price point.

Gaming at sufficient graphics quality to make full HD gaming possible when I hook up to larger displays. I want PS3 level graphics at the least, and that's coming via a) lots of GPU CUDA cores and b) lots of CPU helping out as with known hardware much of the effects rendering (physics etc) can be pushed back on the CPU cores. It's not like consoles don't use them, and that kind of customisation is the whole benefit of having known SoCs used widely.

Re:Core count obsession

[Zilog]

That's what she tells you, but that's not what she really thought.

Re:Core count obsession

[Anonymous Coward]

More cores theoretically means more performance per watt.

Run the same task instead of one CPU on two CPUs with half the clock.
Half the clock means lower voltage, and you can use slower switching transistors with less leakage.
Not speaking of the advantage you gain, by being able to simply turning of a whole CPU.

Re:Core count obsession

[thereitis]

I've got a Transformer and one thing it is in need of is much longer battery life. If I charge it and let it sit on a table, it lasts 2-3 days. So whenever I reach for it, it's battery is usually dead. Fricken useless.

Re:Core count obsession

[Mithent]

I agree; the Transformer Prime is an interesting example here, though, since it does have a keyboard dock accessory with integrated trackpad, extra keyboard, and USB host port, which holds the tablet itself like the screen on a laptop (hence the name "Transformer" - it's a major feature of that tablet). The hardware has the potential to be something closer to an Android netbook, although the software isn't there yet. LibreOffice is supposed to be coming to tablets, but I don't see a port of Eclipse any time soon. What with Ubuntu for Android coming out, that kind of future might not be so far away though.

Re:Core count obsession

[errandum]

Not sure what you meant "high load", but I guess that pushing a single thread very high won't cut it. It should be multi-threaded (or multi-process, but since it's java I don't really see it) high load.

Either that or the tool you're using has a bug.

Re:Core count obsession

[fph il quozientatore]

You need to update your tablet. But don't worry, I have a 32-core CPU to sell you for a great price. Only, 31 of them will always be "offline"...

Re:Core count obsession

[GreatBunzinni]

The average user isn't the only one who doesn't have the slightest idea of what hardware he really needs to get the job done. If "us geeks" also knew better then any synthetic benchmark would be automatically dismissed as being irrelevant and useless, and the most important property of a computing rig would be its cost/performance ratio, with cost reflecting not only the hardware price, direct and indirect, but also operational cost. After all, it's irrelevant if a certain game runs at 100fps or 10000fps, and for regular use stuff, such as web browsing, office stuff and whatnot, any 6 year old system is overkill.

Yet, geeks salivate with stuff such as cores, MHz, a string of irrelevant benchmark numbers and even statistics on HPC usage, and this for systems which the closest they come to HPC is calculating the n-th digit of pi.

So, cluelessness isn't exclusive of non-geeks. The e-penis factor is always influencing purchasing decisions. The only difference is that some are more knowledgeable about useless numbers and factoids than others.

Re:Core count obsession

[aaaaaaargh!]

Hey, did you know that you can save even more battery life by shutting down all cores---just don't use the device at all!

Re:Core count obsession

[Bill_the_Engineer]

More cores means better multitasking since threads can run in parallel. Also, even for handheld devices you are unlikely to find, for example, a single-core CPU that is four times faster than each core of a quad-core CPU.

According to Amdahl you're looking at 1 / [( 1 - P ) + (P/N)] where N = number of processors and P is the percentage of a program that could run parallel. So if 75% of a program can be made to run in parallel on a quad-core processor we are looking at 1/[(1-0.75)+(0.75/4)] = 2.29, so we are looking at a maximum speed increase of 2.29 times the speed of a single processor not 4 times.

Another major advantage of multi-core systems is if a poorly written piece of software locks up it is highly likely to also be single-threaded and your system will chug along nicely despite the misbehaving program, allowing you to kill the process (by comparison, on a single-core system you're likely to suffer through five minutes of waiting for the system to respond before you are able to kill the process).

I haven't seen this. In fairly modern operating systems you'll have multiple services in operation. This means you'll most likely have more threads in execution than there are cores. Context switches between threads within a core of a multi-core processor will still need to be made. I've had a misbehaving program slow down an embedded multi-core processor board because we were "unlucky" that the OS scheduler was running on one of the same cores and other resources on the processor board were being committed by the errant process (eg. Memory, I/O ports, etc.) so the system is not as foolproof as you'd like since memory takes time to to swap, and deadlocks across cores can happen when a computing resource is shared.

Sure, in an ideal world this wouldn't happen but when it does happen it's nice to not be locked out of your system because of a single process misbehaving.

It really is a speed versus power issue. In an embedded environment, where one would hope that the system was well tested prior to being released to the public, such a safety net is really not required.

Re:Core count obsession

[Nikker]

As the old saying goes "Don't throw out the baby with the bath water".

Right now the ULPC category was never fast enough to do much other than Angry Birds, polling for mail in the background and have a few web pages going. Now with more power you have the ability to output via HDMI(mini) and have multiple apps running with a simple dual core. Where will quad + cores come into play? Partly with all the pixels getting crammed into these small displays. Most of the 10" tablets are going to be coming stock with 1900x1200 or higher, while the GPU will handle most of the pixel pushing the CPU still has to coordinate where they should be pushed to. As we get faster processors we will be able to do things like real spreadsheets, some graphics editing, video editing. These devices are converging with Netbooks already and on their way to take over Laptops using the keyboard as just an accessory. So ya I myself was one of those hoping that this time out quad core would become standard before next year and I am a little disappointed I do appreciate 2 faster and more efficient cores but I will likely wait till the faster and more efficient quads come out before I lay down any dollars.

Re:Core count obsession

[snowgirl]

I'm more interested in video/image editing than games. Those extra cores help a lot with tasks like that.

Ah yes, indeed.

Re:Core count obsession

[snowgirl]

so, it has come to this.

Re:Core count obsession

[gl4ss]

run a dualcore benchmark and see if it makes any difference.

pretty easy to write.

Re:Core count obsession

[mikael_j]

...so the system is not as foolproof as you'd like since memory takes time to to swap, and deadlocks across cores can happen when a computing resource is shared.

I didn't say it was foolproof, nor did I think it. I merely pointed out that one advantage of multiple CPU cores is when a runaway process tries to use every available CPU cycle (not any other resource, just CPU, some programs do stupid things like this) and the underlying OS allows it to do this it is a good thing to have more cores so you don't have to sit around and wait for the device to register user input (I had a Nokia smartphone which had a few programs that seemingly did this, they'd begin to process data and lock up the whole phone when they hung).

Re:Core count obsession

[Anonymous Coward]

Whoa, e-penis envy brought out the nerd rage!

Re:Core count obsession

[Anonymous Coward]

Then a tablet is not likely what you need. And just because the tablet *has* extra cores doesn't mean the software is *using* the extra cores. Adobe still has problems getting multi-core and/or CPU/GPU to work together properly all the time on desktops, so I can't imagine the tablet situation is any better.

Re:Core count obsession

[Bill_the_Engineer]

What you are describing is more of a function of the OS (Preemptive vs. Cooperative) multitasking.

The OS being used on the ASUS transformer is fairly modern and preemptive. By being preemptive the scheduler is a high priority process that uses an interrupt signal to trigger and perform context switches (switching between process threads). This is not as foolproof as you'd think because I've had Linux slow down to a crawl because of an errant spin-lock without a yield (sleep) which caused the scheduler to wait the full 10ms between context switches and of course the ready/wait issue on the process list which caused the other processes to starve due to the errant process always being ready and therefore selected again. When this happened the response time was slooooow but I was still able to kill the process it just took a while. I do not believe the number of cores would make much difference, since I had this happen on a quad-core processor board.

Preemptive multitasking on an OS that enforces real-time constraints or a much simpler embedded OS that simply gives each process a time slice without checking status are more resistant to this error. However, the likelihood for needing this foolproof of an OS is low which allows Linux and Linux based OS that are soft-realtime to be an acceptable choice in 99.9% of the other cases.

Re:Core count obsession

[FunkyELF]

So a quad-core tablet isn't for you, this doesn't mean nobody has a use for it.
This isn't Apple where you have a one size fits all product.
There are plenty of existing dual core Android tablets out there; and single core ones; and ones with small screens; and ones with big screens; etc.

Most iSheep under-use their $600 iPads and they have no choice because there is only one model (with various amounts of storage)
The good thing about Android tablets is the variety. Most people could save $400 and get a Kindle Fire, but if they want a big boy tablet, its nice that things like the Transformer Prime exist.

I would love to have one of these quad-cores with the new Ubuntu for Android running on it.

Re:Core count obsession

[ozmanjusri]

What makes you say that? I'm using it already, and it's fine for the purpose. The touchscreen is tactile and HDMI out even makes it simple to preflight the videos.

Re:Core count obsession

[Haxagon]

Okay, you made a whole hubbub about choice, but still haven't explained why quad-core would be a good choice. I think the debate is over whether or not quad-core is worth it, not if there should be a right to choose.

Re:Core count obsession

[Anonymous Coward]

The average user doesn't have the slighest idea how threading works nor why having more cores might be overkill. To them, it's just yet another number that must be increased. They look to us geeks, with our multi-core and multi-socket systems, and figure that's where they want to be once the prices come down.

They're not looking to the geeks, they're looking to the marketers, and the marketers know that numbers are an easy way to distinguish their product. First it was Gigahertz. We haven't really had any meaningful changes in GHz in five years (although look at the overclockers...). Now it's cores, even though they're almost completely useless outside of a few special applications. Even games - I was struck that Deus Ex Revolutions uses only one core for 80% of the game, only occasionally (by all appearances) ever adding a second. Video is better handled on separate, dedicated decoder that's essentially standard anymore.

Marketers have found a number that resonates with consumers, so expect it to keep going up. Just like the number of transistors in a 1960s radio, where they'd sell you a 12 transistor radio that only had 4-6 actually wired in.

Nice to see a company forgo the marketing hype for practical function

Re:Core count obsession

[noh8rz2]

dude, you got owned by the AC. harsh! also, you sell computers? can I buy one? do you sell macs?

Re:Core count obsession

[21mhz]

Gaming at sufficient graphics quality to make full HD gaming possible when I hook up to larger displays. I want PS3 level graphics at the least, and that's coming via a) lots of GPU CUDA cores and b) lots of CPU helping out as with known hardware much of the effects rendering (physics etc) can be pushed back on the CPU cores. It's not like consoles don't use them, and that kind of customisation is the whole benefit of having known SoCs used widely.

So you don't want to buy a console and tap its mature and focused market of games, instead you expect game vendors to have come forward to meet your use case and not just provide a shrunken down phone gameplay for mobile use. How many of them already did so? You plug in your phone whenever you need to play, and carry the extra hardware in your phone, putting up with some overhead in size, battery life, etc., even if you don't use it most of the time?

Re:Core count obsession

[noh8rz2]

then get a laptop. or an iPad - iPad photoshop was just released, and by all accounts it's awesome.

Re:Core count obsession

[noh8rz2]

iphone 4 came out two years ago, and it ran flawlessly. still does under iOS 5. iPad came out 2 years ago, and millions upon millions are still in use. FUD much? although I grant you that androids sucked (and still do).

Re:Core count obsession

[garyebickford]

And then someone will come out with the ultimate answer to more cores: "New, Coreless Computing(TM) - no need to wonder how many cores is right for you! With our new Coreless Computing technology, you can beat all those pathetic multicore junkies!" - picture of pathetic nerd looking glumly at his now-obsolete shiny.

Of course, I have no idea what 'Coreless Computing' might be - maybe processor-per-cell memory? Neural network processors?

Re:Core count obsession

[Anonymous Coward]

I tried System Tuner on the Galaxy S2 and the second core really IS usually offline. However, after running multicore-aware apps (such as MX Video player, FPSe, probably many games), you can see the second core stays active for a few seconds after switching back to System Tuner.

Of course, what really matters is seeing real life gains. What I find impressive is seeing MX Video player smoothly decode fullHD video in software (e.g. no more transcoding - ever).

Re:Core count obsession

[James_Duncan8181]

Uh, presumably they'd be doing that because tablets are at the same resolution. And yes, thanks to Moore's law phones are approaching PS3 level graphics. Does this surprise you? Even on a 4-5 inch display that makes a difference, so it becomes acceptable for both. Tegra 3 is pretty much ps3 level visuals, and fits in phones that are this size. That's small enough, light enough and can do it the job. Given that it drops down to one lower powered core for the majority of the time and has good batter life, why wouldn't I want the extra high quality game use case and general ability to ramp up processor performance if required? It also means stuff like Google Goggles can do real time image analysis and identification and runs smoothly. And as for relevant games? Sure, there are some already coming. And what it gives me? Same games wherever I am, ability to maintain game state, ability to not buy a console for more money, access to my mobile casual style games on the big screen, an open platform I can develop for and games that last more than one hardware generation. YMMV on these things, I guess.

Re:Core count obsession

[21mhz]

What are you talking about? Two years ago phones ran like absolute pieces of crap. Even one year ago, for the most part. There was constant stuttering, things weren't that smooth, etc.

Sounds like a typical Android UI experience :) I suspect that part of the drive for multi-core devices is to compensate for poorly designed software that couldn't respond to the user and wait for something else in the same thread.

Only since december or thereabouts have phones actually been getting towards powerful enough to handle everything smoothly.

I have a Nokia Lumia 800, that was released in November and was right then panned as hopelessly inadequate in the Self-Evidently-Important Core Count Smackdown charts. As it happens, I fail to find a case where I would want it to work much faster. Few applications take a little time to load into a useful state, but nothing blocks or causes annoying disruption. I would hazard an observation that one important thing often overlooked is how cleverly the software uses the CPUs that it is given to run on.

Four processors on a smartphone is fairly trivial in cost, so no it doesn't include a significantly larger price point. At the end of the day we're talking a $12 chip versus a $25 chip (maximum) for example.

Let's see how it looks like in retail.

Re:Core count obsession

[jo_ham]

My internet connection is faster than a 10/100 switch, so it really did get faster when I put a gig switch into one of the branches where a couple of machines were stuck far from the router with only a single cable run to share and I didn't have a gig switch handy.

Still, I agree with the spirit of your post. ;)

Re:Core count obsession

[Anonymous Coward]

Navel gazing nerd wannabe fascist motherfuckers. Did all those wedgies in high school turn you into some kind of no-power having power struggler? Listen, fuckwit, I have a Xoom and my friend has a Transformer Prime. The Xoom is clocked significantly higher than the TF due to root/overclocking yet the TF is a faster machine. The relevant difference? Twice as many cores, stupid. So stick your core nazi bullshit up your tight little asshole.

Re:Core count obsession

[21mhz]

Uh, presumably they'd be doing that because tablets are at the same resolution. And yes, thanks to Moore's law phones are approaching PS3 level graphics. Does this surprise you? Even on a 4-5 inch display that makes a difference, so it becomes acceptable for both.

I can halfway understand the case for tablets, but are you saying that the same game UI can be played equally well on a big TV screen and on a phone, that furthermore has little in the way of input usable for typical 3D games? I guess I'm just not into gaming that much, and certainly not to a compulsory degree that would make me want to play big screen games on a phone.

Re:Core count obsession

[oakgrove]

Dude, we get it, you like your iPad. I like mine too. Would you stop with the obnoxious platform zealotry? Please?

Re:Core count obsession

[oakgrove]

I don't see a port of Eclipse any time soon

If you want Eclipse so you can write Android apps on your Android device, you can always use this. With that you can write java cli apps or regular Android apps all day long on your TF.

Re:Core count obsession

[symbolset]

We don't chare if the scheduler gets stressed out. We didn't have to design him.

Re:Core count obsession

[Anonymous Coward]

I've been searching for someone to answer to a simple question: what are the tasks at which you feel your high-end smartphone should be faster, that are not attributed to things like network roundtrips where a faster proc is irrelevant? "Four times faster than good enough" just does not equal to "four times as good" to me, especially if it comes with a larger price point.

Web browsing. Until I got my first smartphone recently, I assumed that modern phone web browsers were slow just due to network latency. Then I teathered my laptop to my phone and it definitely felt a slower than normally browsing over a cable modem, but not by all that much. Certainly nowhere near how slow I expect a smartphone web browser to be (which, to be fair, is fast enough to be usable unlike, say, the web browser on the N810 which was painfully slow since the device just wasn't powerful enough for a web browser). Of course, multiple cores wouldn't help here: both were running Firefox which is single-threaded.

Re:Core count obsession

[biraneto2]

Are you aware that the android OS runs on a modified Linux kernel? Android multitasks all the time. The fact that it simplifies the UI and gives the impression of a single task being executed does not mean it does not take advantage of multiple cores. Do you really think that the app needs to be smart to use a second core? It's as simple as creating a thread or drawing a graphic component.

Also, video is a great example of a situation where multiple cores are very welcome and would immediately prove 4 cores is better. It's funny you mentioned since tablet users watch video all the time.

I don't get people saying that core count is an obsession... stop saying that tablets and smart phones don't need more than 512k of ram.

Re:Core count obsession

[BlackCreek]

I've been searching for someone to answer to a simple question: what are the tasks at which you feel your high-end smartphone should be faster, that are not attributed to things like network roundtrips where a faster proc is irrelevant? "Four times faster than good enough" just does not equal to "four times as good" to me, especially if it comes with a larger price point.

I use "paper camera" (a image processing app that post-process photos adding effects). It is slow'ish when taking the pictures and post-processing. The time it takes to be able to take the next shoot is much larger than for "normal" photos (i.e. those that don't require special post processing). The last time I tried (there was an update since) a video with 'paper camera' also had lower resolution because of processing constraints. (I own a Galaxy Nexus.)

I can just take the shots and post-process them later, but that is less fun than just shooting directly.

Re:Core count obsession

[21mhz]

Web browsing. Until I got my first smartphone recently, I assumed that modern phone web browsers were slow just due to network latency. Then I teathered my laptop to my phone and it definitely felt a slower than normally browsing over a cable modem, but not by all that much. Certainly nowhere near how slow I expect a smartphone web browser to be (which, to be fair, is fast enough to be usable unlike, say, the web browser on the N810 which was painfully slow since the device just wasn't powerful enough for a web browser). Of course, multiple cores wouldn't help here: both were running Firefox which is single-threaded.

Right, if the browser engine can do heavy lifting in a separate thread/process, another core could be helpful. But it's easier said than done, as illustrated by Firefox: the whole thing about flowing and rendering a webpage, running attached scripts, while being able to handle scripted I/O events etc. does not sound like an awfully parallelizable lot of things to do.

For what it's worth, IE on Windows Phone 7.5 works smoothly enough, as does WebKit on the N9. Maybe I never gave them a properly gross webpage. Most sites I use with my phone provide "iPhone-optimized" versions which work like a charm. Ignoring gobs of gratuitous flash probably helps as well.

Re:Core count obsession

[Anonymous Coward]

GGP didn't say "I want PS3 games", he said "I want PS3 level graphics at the least". So control schemes and UIs aren't relevant, if you can conceive of any phone game that has enough stuff going on to make PS3-level graphics sensible (e.g. Galaxy on Fire).

Re:Core count obsession

[dfghjk]

"The more processors you ahve, the more complex scheduling your apps needs to perform to actually work faster."

No you don't.

"It's better to hav ea single core that is twice as fast, than two cores running in parallel."

Sure, but those cores that are twice as fast come with significant downsides.

Re:Core count obsession

[noh8rz2]

frown face :(. Just trying to introduce some sense into the conversation. why did adobe release photoshop for iPad 2 yesterday when iPad 3 comes out next week? I would have waited a week then called it "TEH AWSUMZ PHOTOSHOP 4 IPAD 3 ZOMG!"

Re:Core count obsession

[Anonymous Coward]

Marketing. Back in the '40s and '50s the buzz in radio was the number of tubes. Companies would brag about how many more tubes their radio's had over their competitors. It got to the point that some of the tubes weren't even connected, but that didn't matter - consumers associated more tubes with higher class radios.

Re:Core count obsession

[Bigbiff]

Unless you are running a custom kernel you probably don't have the sysfs interface to turn on the 2nd core reliably. I think most manufacturers are using a userspace daemon to control when the 2nd core will come online.

Re:Core count obsession

[Mithent]

Photoshop Touch has already been out for a few months on Android tablets actually, and the iPad version is "as close to identical to the Android version as is possible, given the platform differences". So that's no reason to get an iPad over an Android tablet.

Re:Core count obsession

[TrancePhreak]

One core for the OS to function, one core for the phone parts to update, one core for the game logic, one core for the game graphics to update. It's not that hard to see why more cores could easily be better.

Re:Core count obsession

[Cederic]

Personally I want my phone to be successfully streaming media from the 'net, playing it through my bluetooth headset, waiting for inbound calls and rendering web pages correctly all at the same time.

Given the web page rendering can max out a single core all by itself, how were you planning to avoid stuttering audio, interrupted audio feeds or missed calls?

Sure, it _can_ be done - but it can be done better and quicker with multiple hardware threads.

Re:Core count obsession

[Niomosy]

Given that a single-core phone can handle all of that at the same time, it's still not a compelling reason as to why four cores are needed.

Re:Core count obsession

[m50d]

On my original (dual-core) transformer, I find system performance suffers noticeably if I'm torrenting in the background. Niche use case maybe, but I do it fairly often, and it seems like more cores should be able to help with that.

Re:Core count obsession

[21mhz]

Personally I want my phone to be successfully streaming media from the 'net, playing it through my bluetooth headset,

Unless HD video is involved, all this should not take a large share of scheduling slots on a reasonably good smartphone CPU. It worked on Nokia N800 with plenty of cycles to spare.

waiting for inbound calls

This is what the modem unit is supposed to busy itself with.

Given the web page rendering can max out a single core all by itself, how were you planning to avoid stuttering audio, interrupted audio feeds or missed calls?

By boosting the audio service process' priority, the way it's done in all modern operating systems? Note that it does not take much processing time to feed the audio sink with buffers, it just has to be done on time.

An incoming call should suspend all media playback and put the active application to the background, naturally. If your phone software does not do it, you got bigger problems than a lack of cores.

Sure, it _can_ be done - but it can be done better and quicker with multiple hardware threads.

I'd put it as "it can be done with more slack, and the warts will still show up at times". Certain popular devices seem to follow this design philosophy, indeed.

Re:Core count obsession

[21mhz]

GGP didn't say "I want PS3 games", he said "I want PS3 level graphics at the least".

He did mention hooking up to a large screen, so I assume he meant console-like gameplay as well.

So control schemes and UIs aren't relevant, if you can conceive of any phone game that has enough stuff going on to make PS3-level graphics sensible (e.g. Galaxy on Fire).

I can't conceive of a game that would make me strain to resolve so much detail on the phone screen. For phone games I did play, moderately busy OpenGL ES scenes on a 800x480 screen worked fairly well. The GPU is there to help, too.

Re:Core count obsession

[adisakp]

One app at a time. If that app is smart enough to offload background tasks to a 2nd core, I'm already impressed.

Tablet programmers should know a little multithreading. Even if you have only a single core, offloading slower processing tasks to a second thread in the background allows you to have a snappy responsive main / UI thread.

Re:Core count obsession

[unixisc]

Why is everyone obsessed at the number of cores? The more processors you ahve, the more complex scheduling your apps needs to perform to actually work faster. It's better to hav ea single core that is twice as fast, than two cores running in parallel.

Precisely!!! And for tablets, there ain't much SMP that one is doing - one typically works on just 1 app @ a time, and is not likely to have another app actively running (except for say, a download, but that would probably require better Wi-Fi chips - say 802.11n in it). Maximize the performance of a single core given the most commonly used apps and benchmarks, and then do what's possible to minimize the power consumption and maximize the battery life. The dual core here is enough, if not more. Instead of the extra core, pack in more cache memory, and maybe even have a burst-mode flash memory that the CPU can access, where the OS resides but which is different from the flash used to store user data, communicate w/ external wireless devices and so on. More cache would reduce memory swapping, prolonging battery life, while having a separate flash for just the OS allows a vendor to make it extra secure, and allow it to be updated only by the vendor and reduce the risk of the user damaging it due to some hostile downloads and making it unusable.

Re:Core count obsession

[Anonymous Coward]

Why is everyone obsessed at the number of cores? The more processors you ahve, the more complex scheduling your apps needs to perform to actually work faster. It's better to hav ea single core that is twice as fast, than two cores running in parallel.

Well, one advantage is in power consumption. When you have a single core running twice as fast, it really sucks when your machine is not really doing a lot (which is like 99% of the time.) With more cores, you can shut them down when you are not using them to save power.

Re:Core count obsession

[JazzLad]

He has an iPad.



/there goes my karma ...

Re:Core count obsession

[JazzLad]

The guy sitting next to 'that single (core) guy' is the one with the pocket protector talking about how his "single socket, single upper layer core, which combines the power of 128 microprocs in an 8 core assembly for a whopping 1024 hidden 1mhz cores - almost 1 Ghz of raw power"


/grin

Re:Core count obsession

[ozmanjusri]

I keep my files on external hard drives and plug them into the Transformer's USB ports. How do I do that with an iPad?

Re:Core count obsession

[noh8rz2]

get an iPad dock connector with the SD accessory! duh!

Re:Core count obsession

[rerogo]

Yes, but a 4-core phone can do this at a quarter the clock speed the single-core phone would need. Power consumption is quadratic with frequency, so adding additional cores winds up giving a net power savings (at the expense of die area)

Re:Core count obsession

[mjwx]

Why is everyone obsessed at the number of cores? The more processors you ahve, the more complex scheduling your apps needs to perform to actually work faster. It's better to hav ea single core that is twice as fast, than two cores running in parallel.

Pfff... actually, the Tegra 3 has five cores, four of them are high-performance, and one is high-efficiency. The CPU is designed to shutdown the four cores for almost nearly everything, and just use the high-efficiency core in order to save on battery life.

Introducing the 16 core processor.

The first core walks by the process assessing it's potential time consumption.
The second core types this out into a report and forwards this report to the other cores.
The third core skims the report before filing and ignoring it.
The fourth core empties the inbox of the third core, failing to note the process.
The fifth core is focusing on it's career and promotion through middle management.
The sixth core notices that there is a process and tries to point this out the the third, fourth and fifth core.
The seventh core is having a nervous breakdown.
The eighth cor0xDEADBEEF.
The ninth core is dealing with the problems from the malfunctioning 8th core.
The tenth core distracts the process by acting as a door to door salesmen.
The eleventh and twelfth core hold the process down whilst the thirteenth core goes through it's wallet.
The fourteenth core takes the process's statement.
The fifteenth core actually runs the process.
and the sixteenth core is just along for the ride.

Re:Core count obsession

[Anonymous Coward]

BS.

Its called a CPU cache. And CPU stalls while waiting on devices.

Multi core allows one processor to handle the IO for a single process, for example, which would trigger a system context on a single core
alternately, two programs can run in parallel without disrupting each others cache (and therefore waiting on slower system memory, etc)

Re:Core count obsession

[Elrond, Duke of URL]

Sadly, the Droid 3 has a locked bootloader, so even though I have rooted the phone, I cannot change the kernel. However, if it is using a userspace daemon, or even if it is a custom kernel interface, then surely System Tuner could use the same method to control the second core?

But, as others have pointed out, perhaps there is an error or bug with how System Tuner reports and/or controls the second core.

It's hard to say. But I have experienced periods where my phone becomes very sluggish and it *seems* to be caused by a single app becoming very busy for a period. I should think that in these instances the second core would become active and assist the system. Fortunately, I didn't buy the device for the dual-core feature, but it would be nice if it appeared to do more.

I'll have to do some more investigating, including trying the multi-core aware video decoder that was suggested.

Re:Core count obsession

[ozmanjusri]

But the dock connector doesn't work in host mode, nor does it power portable hard drives. With the tools I have now, I can plug my ordinary 1Tb portable drive straight into the tablet.

Why would I want to make it more complicated?

Re:Core count obsession

[bingoUV]

it *seems* to be caused by a single app becoming very busy for a period

My dual core Atrix too does it, but most of the times it is due to high I/O load. Second core can't do much here. Is that the case with your phone as wel?

Re:Core count obsession

[noh8rz2]

A better question is, why do you think you need that much storage? iPads have already been designed with consumer needs in mind.

Re:Core count obsession

[billcopc]

They should, but often times they don't. I think one of the reasons they don't is because IOS makes it too damned easy since Objective C effectively spams the kernel with threads for every little thing. Android does it a tad less but it's still easy to ignore threading, with the notable exception of progress indicators (spinners), and still end up with a perfectly snappy app. The only platform that actually requires you to perform explicit threading is the Blackberry (Java), where if you get it wrong, your entire app slows down to a crawl, network I/O times out, and eventually the whole damned system crashes.

I hate to admit it, but I'm rather fond of the Objective C model because it forces the developer to work in a thread-friendly way, letting the OS figure out parallelism at runtime (or so it would seem). Even after a year of writing mobile and desktop apps, the only times I've actually written thread management code are when I was doing two non-UI tasks in parallel, typically some sort of batch processing job. I really don't see myself doing that kind of number crunching in the great majority of tablet apps, except maybe the odd game or media editing tool, so the built-in UI threading is sufficient for most iPad and Android apps.

Re:Core count obsession

[billcopc]

Video seems like something best handled by the GPU, no ? And even if not, it's still a single-thread job due to the very nature of video compression. I don't see why 4 cores would be better for video, unless you're watching 4 videos at once.... okay, 3 videos and one soundtrack, assuming each video requires 100% CPU time.

And I don't think smartphones need zillions of gigabytes of Ram. My old 3GS gets by with 128mb, and while that number often plummets, it is largely due to the so-called "multitasking" that keeps old suspended tasks in memory. I am a developer by trade, and even I cannot find uses for that much memory on a mobile device. I'm not using it for number crunching or high-resolution graphics manupulation. For a tablet, 512mb is comfortable, with 1024mb being "luxurious". Just don't expect anything from the Mozilla project to run on it *ducks*

Re:Core count obsession

[biraneto2]

Video can be handled by the gpu if the gpu has video player acceleration (nvidia chipsets usually do have that) and it has usually a lot of cores too (tegra 2 has 8 cores).
It's been a while since video decompression can be split into many cores... a lot of desktop players do that already (using the appropriate decoders).
Smartphones are computers with mobile network ability. I can't wait for them to evolve and give me the ability to plug it into a monitor and keyboard and do the work I currently still need my desktop for.

Re:Core count obsession

[ozmanjusri]

Why do you think that's a better question?

Re:Core count obsession

[noh8rz2]

Lau Hsu said, to look inside for the truth rather than asking it of others.

Re:Core count obsession

[Elrond, Duke of URL]

My dual core Atrix too does it, but most of the times it is due to high I/O load. Second core can't do much here. Is that the case with your phone as wel?

That could very well be, and I had even considered this... but I'm not sure how to check what the I/O load is on the phone or for a particular app.

System Tuner allows one to change the cache size for the SD card. On the Droid 3 it defaulted to 128 before I touched anything, and I have read that this is particularly low. So I set it to 2048 using the app. This is for the external SD card. For the internal SD card, the app shows the cache size at 1000000, and that seems quite high. Also, the value for the internal SD card always seems to reset to one million after a reboot, while the value I set for the external SD persists (there is an option in System Tuner to set this on boot).

Regardless, my playing with these values seems to have had little effect on the phone, good or bad.

In an update to my earlier post, another screen in System Tuner lists the amount of time the CPU has spent at each frequency level. Even though the bar graph still shows the second core as always offline, according to the frequency time breakdown, almost no time has been spent in the "offline" speed. This would seem to indicate that my forcing the second core to be online has had an effect. But, as I haven't noticed any difference in performance, perhaps I will put it back the way it was.

I guess if I really want some more speed, performance, or free memory on this device, I'll need to ditch the stock ROM and install a 3rd party ROM. I've got nothing against them, but I've been planning to wait under CyanogenMod 9 (ICS) is ready for Droid 3 use. That way I only need to change the ROM once.

Re:Core count obsession

[ozmanjusri]

Manifest plainness,
Embrace simplicity,
Spell Lao Tzu correctly.

Re:Core count obsession

[noh8rz2]

you must be thinking of somebody else.

http://www.meetup.com/RAW-Foundation/members/24218882/

What happened with odd-core configurations?

[OliWarner]

I thought 3 and 5 core tablets were supposed to be coming out, where the "odd" core is so underpowered it can be left on when the screen and other cores are off, using practically no battery but still letting the tablet run its background processes.

I'm surprised more emphasis isn't being put on improving "standby" battery time because that seems to be the real killer in so many mobile applications these days (like my 14h SGS2 battery of doom).

Re:What happened with odd-core configurations?

[Mithent]

Tegra 3 has 5 Cortex A9 cores, of which one is low power and used for light loads. The other 4 make up the full-power quad-core, and each core is power-gated so they're only used when required. This means that despite it having 5 cores, battery life is as good or better than the dual-core Tegra 2.

Re:What happened with odd-core configurations?

[errandum]

The tegra 3 does that... But the prime never had that much battery problems, since it got up to 18 (or more like 15) hours of battery if you had the dock. On the other hand, since there are very few devices with 4 cores, almost no application make use them, turning them useless.

Re:What happened with odd-core configurations?

[bemymonkey]

Your SGS2 is configured wrong. You should be getting a standby drain of about 1%/hour (or less) with sync enabled.

Two things are at fault here, of course:

1. The awful apps that keep the phone awake and active during standby - for instance: Facebook
2. Android, for not telling the user THIS APP IS KEEPING YOUR PHONE AWAKE, KICK THAT CRAP!

In your specific case: Check your battery usage (in your SGS2's settings), and find out which process is keeping your phone awake, either with the old battery history (Gingerbread and earlier, accessible via Spare Parts, apps like BatteryMonitorWidget or a dialer code that varies from handset to handset) or (ICS only - because someone at Google decided to remove the battery history) with an app like BetterBatteryStats.

The interesting part is usually partial wake usage. Eliminate the apps causing the most partial wake usage, and you'll have a power draw of next to nothing. Standby battery life with Google sync, a few IM clients (I run Skype and imo.im), Whatsapp, Viber and so on, should be around 4-5 days.

Re:What happened with odd-core configurations?

[Anonymous Coward]

I want mod points! If someone had told me this about my SGS (not 2) ages ago it would have saved me a lot of hassle.

Just like the the OP I couldn't get through a day, even on standby. Now (properly configured with the suckware banished) I get 2-3 days of normal usage. The key is to really pay attention when you notice a battery drop. In addition to poor configuration, I had several buggy apps that would just occasionally go nuts. Replacing them with equivalents made a huge difference.

Hmmm... just realized... I would probably have better luck getting mod points if I logged in...

Re:What happened with odd-core configurations?

[gbjbaanb]

3/5 cores - this is a Tegra thing, the snapdragon does it differently.

Tegra has a 'companion' core that is low, low power for standby tasks, then it switches the main cores on individually as they are needed. Note that the Tegra main cores are on/off designs.

The Snapdragon doesn't do this, it varies the power to each core individually, so they are all running in a low-power mode all the time until they need more. This means it doesn't need a 3rd core. Its debatable which is better, but the snapdragon can run its cores at half-power unlike the Tegra which is probably more efficient overall (when you use limited resources doing boring tasks like reading your messages).

Here's a nice article on the subject

Re:What happened with odd-core configurations?

[lexman098]

I have this problem, but even spare parts doesn't seem to give me any useful info. The battery starts dropping like crazy after I turn my wifi off. Something is staying on, and spare parts says that the ADW launcher is using the most cycles and has the highest start count. I've been turning off random widgets to no avail.

Re:What happened with odd-core configurations?

[Svartalf]

Considering that the S4 eval board clocked at 100MHz faster than the Asus Prime did half-again better in a multithreaded computational benchmark thrown at it, I'd say that you're probably looking at the differences between an A9 and an A15- and you might have found a CPU that's considerably faster single-core than Tegra 3's single cores at clock. If so, there's an explanation for the move. The S4's cheaper. It consumes less power doing what it does at peak. And...if it's faster doing most of the things users do while doing the other two...win.

Re:What happened with odd-core configurations?

[bemymonkey]

Spare Parts => Battery History => Partial Wake Usage.

If possible, post a screenshot and we can work from there :)

Re:What happened with odd-core configurations?

[bemymonkey]

Howdy Tex,

Wrong thread! :)

Re:What happened with odd-core configurations?

[noh8rz2]

you should get an iPad. at least a week of standby, probably more but I never push it that far. if your goal is to leave your tablet tethered, then just get a laptop.

Re:What happened with odd-core configurations?

[Anonymous Coward]

Shut the fuck up you fucking iPad fanboy fuck. Jeez the fanbots are fucking this site up.

Re:What happened with odd-core configurations?

[mjwx]

2. Android, for not telling the user THIS APP IS KEEPING YOUR PHONE AWAKE, KICK THAT CRAP!

This isn't Android's fault, at least not completely. If the App is working correctly, Android should ignore it.

It's the fault of this school of thought known as "User Experience". The user is to be coddled and treated like an idiot whilst any errors must be ignored. User Experience has to be the most bollocks marketing term, it never considers how people actually use a device (the physical side of it, learning curves, etc..) but rather focuses on how the user feels. This allows designer to ignore the wisdom gained from years of research into HCI and HMI in order to push their crap as the latest and greatest. IOS is the biggest offender, like most of the human race, I'm right handed, so why put the back button in the upper left corner of the screen, where my right thumb cant reach it if I'm holding the phone comfortably. Requiring two hands to go back to a list of messages seems stupid to me, but that's "User Experience" for you.

Back on topic, error messages have become dirty words, problems need to be ignored so the user doesn't have to deal with it rather then presenting the information in an easy to digest format then suggesting fixes. Android 2.2 and on have had a really good battery metering tool built in that displays the amount of battery consumed by each process/device (Cell Idle, Screen, Android OS, Maps, et al.), but most users dont know about it. As of 2.3 ish it even recorded signal strength along with a graph of the battery level. Here's an example, powerful tools hidden from the user because designers insist on keeping users idiots rather then trying to help them.

Re:What happened with odd-core configurations?

[bemymonkey]

It's not really a matter of error messages... but rather the fact that it's too easy to make one little mistake (flip a sync switch here, install a new app) that causes you to end up with an empty phone after just a few hours of standby. This is caused by exactly this school of thought:

"If the App is working correctly, Android should ignore it."

If the app drains 40% of the phone's battery over the course of three hours in standby, it's not working correctly - this should never happen. Yes, it should be possible to manually force these scenarios (because they are realistic for some particularly heavy users), but having apps be able to run like this by default is insane. That's the price we currently pay for full multitasking and a very open "you can do pretty much anything with it" system, but there are better ways to use it... say a simple whitelist for allowing apps to run during standby.

"Android 2.2 and on have had a really good battery metering tool built in that displays the amount of battery consumed by each process/device (Cell Idle, Screen, Android OS, Maps, et al.), but most users dont know about it. As of 2.3 ish it even recorded signal strength along with a graph of the battery level. Here's an example [tapatalk.com], powerful tools hidden from the user because designers insist on keeping users idiots rather then trying to help them."

Unfortunately, this battery metering tool is blunt and inaccurate at best. Not only is the system very simple and not particularly easy to read accurately, it doesn't show WHICH processes are keeping the phone awake very accurately. See the "awake" section on the screenshot you linked to? Problematic phones have solid blue strips of that (even though the screen is off at the time), and the only way to diagnose which app is causing it is battery history or third party tools like BetterBatteryStats (or the command line equivalent, but that's a little finnicky on a handset).

The built in battery monitoring is great for a quick overview, and for checking if everything's OK, but when I see a handset with bad battery life, the culprit only shows up in there about 50% of the time...

Re:What happened with odd-core configurations?

[Anonymous Coward]

I think you misunderstood - this is useful information. His iPad is something he typically doesn't use for 3-4 days at a time before picking it up to plug it in. If I ever bought a tablet that couldn't multitask, couldn't side-load any app I wanted on & despite having a large 10" screen was the same resolution as really cheap 7" tablets, I'd really want to know how long I could just let the bloody thing just sit there unused. Posting anon as this will get someone's panties in a bunch & she may have mod points.

Re:Who gives a damn about cores, does it work?

[errandum]

In a website like this, yes, people care about that. You seem to be lost, the appstore is that way --->.

Summary wrong again?

[Anonymous Coward]

From what I articles I saw yesterday I gathered that there would be two levels of the new Asus pads. One with the Tegra and the other with the new Krait. Here is one article that talks about it: http://www.anandtech.com/show/5586/the-asus-transformer-pad-infinity-1920-x-1200-display-krait-optional

Of course we won't know anything for sure until Asus releases the product details.

Re:Summary wrong again?

[Monoman]

Mod parent up. Another incorrect summary on /.

Re:Summary wrong again?

[Nimey]

Not to mention the breathless sensationalism that has a better place on low-class sites like the National Inquirer or Fox News.

Seriously, "shock and disgust" over what is probably a good technical decision? Slashdot has found another new low.

Re:Summary wrong again?

[Anonymous Coward]

I read the same thing on Engadget. The wifi-only Infinity will have the Tegra 3, while the wifi/3G/4G/LTE/eieio will have the Krait. C'mon, /.ers, how about some research?

Re:Summary wrong again?

[Anonymous Coward]

Of course we won't know anything for sure until Asus releases the product details.

Here
With powerful and higher standard processors using either the Qualcomm Snapdragon S4 Krait Dual-core(4G/3G) or the NVIDIA® Tegra® 3 T33 Quad-core (Wi-Fi) CPU with 1GB ram,

It would seem the Wifi version gets the Tegra3, and the 3G/LTE version gets the S4 Krait.

Re:Summary wrong again?

[Cederic]

Of course we won't know anything for sure until Asus releases the product details.

All of which will be utterly irrelevant, as the Transformer Prime experience has demonstrate that Asus will be wholly incapable of meeting market demand or indeed of fixing the initial design flaws that prevent it being a truly great piece of hardware.

I speak from a position of disappointment. It's taken so long to acquire a working Prime that I've given up trying and I'll wait for the next tablet+keyboard combo instead.

Interestingly,

[MrHanky]

the one outstanding product to come out of the MWC, Nokia's 808 "Fuck everything, we're doing 41 megapixels" PureView, is ignored by Slashdot for whatever reason, while tiny product differentiations that don't warrant attention at all are posted.

Re:Interestingly,

[errandum]

No, that is the dumbest of them all. Symbian is dead for smartphones (anyone can see that), why on earth would they not add that to their windows mobile line? It's like saying "look, we have this 500bhp engine, lets fit it into a horse carriage".

Re:Interestingly,

[MrHanky]

Sure, it's so dead that it sold more than six times as much as WP7 in Q4 2011 and almost half as much as the iPhones. Any blind and gullible person can see what they're told they should see, and the idiots will repeat it ad nauseam to show how perceptive they are.

Fact of the matter is, Nokia just released some interesting technology. This site is supposedly about "news for nerds", not news for gadget consumers who feel they need to prove themselves by buying whatever "experts" in the social media tell them is the next big thing.

Re:Interestingly,

[Mithent]

I agree, it was a strange decision. 41 megapixels is a huge headline-grabber, and they use that to promote the platform they're killing off rather than the one they're trying to build up?

Re:Interestingly,

[errandum]

Ok, then tell me how many of those were smartphones. Symbian sold in low end normal phones (where it actually fit), the symbian adaptation for touch is a fiasco that even Nokia has seen already (why on earth would they bow down to windows phone when they had, at least, 2 other OS choices).

I don't know if you're trolling me or actually being serious, but there is a reason Nokia (even though they sell that many phones) has been downsizing left and right.

Re:Interestingly,

[Anonymous Coward]

No, that is the dumbest of them all. Symbian is dead for smartphones (anyone can see that), why on earth would they not add that to their windows mobile line? It's like saying "look, we have this 500bhp engine, lets fit it into a horse carriage".

As posted elsewhere (in a different site), this was 5 years in development. But now that it's up and running (sort of a proof of concept I guess), they are rolling it out on their future (WP) phones.

Re:Interestingly,

[JanneM]

They're binning the samples, for an actual resolution of 5mp. And they have to; a lens that size is unable to create an image of sufficient resolution for anything like 40mp being useful. You go above 8mp or so and you'll only get better pictures of the lens blur.

And it's not at all clear that binning several individual photosites is better than simply having larger sites in the first place either. Of course, being able to write "41mp!! *woot* *Munchkin FTW!* " in your promotional material is a likely sales improvement even if the technical improvement is nil.

Re:Interestingly,

[21mhz]

Spot on. Also, this is a more ingenious use of the resolution than just throwing in megapixels for the spec sheet value with no difference to actual image quality for the tiny lens and the respectively constrained matrix size. Without multisampling or other processing, higher resolution matrices may actually produce worse results because sensitivity of individual elements gets lower as they get smaller.

Re:Interestingly,

[MrHanky]

Untrue. They've released a bunch of 38 MP images with staggering detail. It is, of course, not its primary function: they take up far too much space for a phone, and in the end you'd prefer noise free 5 MP images for phone usage. But of course, you're free to dismiss things out of hand, based on no evidence whatsoever.

Re:Interestingly,

[Xerolooper]

Nokia's 808...PureView, is ignored by Slashdot for whatever reason...

But does it run Linux...?

Re:Interestingly,

[Anonymous Coward]

Symbian phones are, by definition, smartphones. They do proper mulit-tasking, for instance. The Nokia phones you are talking about use a completely different operating system. You are a fool.

Shock and disgust?

[GPF(BSOD)]

What in the hell? How can anyone be shocked or even disgusted at something this trivial?

Exciting!

[scotjam]

I'm not shocked or disgusted, but I am rather excited. I have the Transformer, and its one big shortcoming for me is the lack of 3G. A high-end LTE version sounds awesome. Number of cores be damned.

Next on my wishlist: better support for Office documents and shared calendars

Does this mean I don't get to be classified as a "bleeding-edge technophile"?

In the market for a 4 core Android tablet

[Anonymous Coward]

Pity, I wanted a multi-core Android tablet to do some decent interactive number crunching on. The more pixels the better, the more cores the better.

Who says the only use for a powerful interactive tablet is games?

So what's the option for me, what's the fastest tablet, most cores, runs Android, as many pixels as I can get, needs minimum 1Gb of ram (more if possible), price not a problem, battery life isn't important.

Re:In the market for a 4 core Android tablet

[errandum]

I'd say that the new Samsung phone with the projector has 6GB of RAM, so I'd go for it. It's not a tablet, but since you can use it as a 50" screen it should be enough for your needs?

And you have to understand, you're the 0.01%. Most people with your requirements will get a computer (any i7 notebook should fill your needs and give you 10 times the power of any tablet).

What's your specific problem? Maybe something else is more suited.

Re:In the market for a 4 core Android tablet

[noh8rz2]

So what's the option for me, what's the fastest tablet, most cores, runs Android, as many pixels as I can get, needs minimum 1Gb of ram (more if possible), price not a problem, battery life isn't important.

it's called a desktop computer, mofo.

"Bleeding edge technophiles" are a bit sensitive!

[sirwired]

If all it takes to "shock and disgust" "bleeding edge technophiles" is a technical decision to pick a CPU with faster cores instead of more of them, then these "bleeding edge technophiles" must not get out of Mom's basement very often and are in need of some serious therapy.

It's got the standard Qualcomm hypervisor

[tlambert]

They run it to be able to run the radio on the same hardware. But that means that it won't necessarily run anything in bounded time, hence the need for more than one core. Oh, it also comes in 2 and 4 core packages.

-- Terry

Only the LTE model

[jgfenix]

The wifi model will still use the Tegra 3. The LTE model will use the Snapdragon http://www.theverge.com/2012/2/27/2827588/asus-transformer-pad-infinity-series-and-transformer-pad-300-series

The reason is LTE

[Sollord]

The Tegra3 isn't compatible with any LTE modems and won't be for several months so ASUS opted to use the S4 for all 3g/4g transformers so they could have something for carriers to sell nowish. The Wifi only models will keep using the Tegra3. Either way this isn't really something ASUS can fix itself since Nvidia never bothered getting its product to support any LTE modems.

Re:The reason is LTE

[Sollord]

Also... Based on Anandtechs review of the S4 vs the Tegra3 Prime there's not much benefit to having a quad-core vs dual-core as the Tegrra3 lost almost every test save for some of the GPU tests and the linpack test was just brutal to the Tegra3 compared to the S4

Re:The reason is LTE

[Sollord]

WTF where did my links go... lets try that again and /. needs a damn edit button.

http://www.anandtech.com/show/5559/qualcomm-snapdragon-s4-krait-performance-preview-msm8960-adreno-225-benchmarks

http://www.anandtech.com/show/5563/qualcomms-snapdragon-s4-krait-vs-nvidias-tegra-3

Re:The reason is LTE

[YoopDaDum]

Yes, and the summary is sensationalist (par for the course I guess) and misleading (more annoying).
The story is just that Asus preferred the Tegra3 when having to chose a platform without LTE. And when taking into account LTE, it preferred the new Qualcomm S4 combined AP and LTE solution.

Now when you say:

The Tegra3 isn't compatible with any LTE modems and won't be for several months

It can be misunderstood as an issue on the Tegra3 side, which would be unfair. There's nothing special to do to hook a LTE baseband to a T3. It's all Android and very common interfaces. The issue is more on the LTE modem offering side: the current chips are either not available (internal developments at Samsung, Moto, ...) or not attractive enough. We'll have to see a new generation of LTE modems arrive to see a nice combo based on an NVidia chip, and this is a problem for NVidia for sure. But all the work will be on the modem provider side, and it's not much work to support a T3. All modem vendors nowadays are targeting Android and from the modem support software point of view there's not much difference between one AP running Android and another.

Re:The reason is LTE

[Svartalf]

That and the S4's got a higher performance profile with a lower power consumption for actual ICS usage. It's as much about battery life and perceived performance as it is LTE support out of box.

First HTC drops its quad core chip for a dual core

[AlienIntelligence]

First HTC suddenly drops its quad core chip for a dual in a phone that
was supposed to have a quad core chip since it was leaked back in July.

And days later, Asus drops a quad in favor of a dual core.

Same chip was dropped.

Someone... is keeping a secret. There is a problem with the quad core
chip and 'something' new(er) that is appearing in the phones. I read that
an LTE chip appeared in the "One X", while the quad core disappeared.

Is LTE and quad core not playing nice? Are there production shortages?
Overheating issues, battery issues?

The whole story isn't out. I'm curious what it is. I've been waiting
and salivating at the promised "Quad" core offerings for smartphones.
The Samsung SIII is supposedly going to have one, but from a different
company, their own Exynos chip. So, we won't see that quad be cut in half.

Hopefully.

Regardless of what the non-power users say about not needing more cores,
I see my dual cores maxed out regularly. I need the extras, I was willing to
sell my life, I mean soul, I mean sign a new 2 year contract for it.

-AI

Tegra is a flop.

[mykos]

Maybe not in sales, but the Tegra continuously lags behind many manufacturers both in performance and power consumption every generation. Nvidia's adventures in SOC Land might be profitable, but they're always the bottom of the barrel when it comes to performance.

Re:Tegra is a flop.

[Fnkmaster]

Ummm Tegra 2 was the fastest platform for Android for quite some time. The G Tablets are still pretty blazingly fast. The issue is just that Tegra 2 was released for such a short time before Tegra 3 came out that it never got much saturation, and then Tegra 3 came out with a bunch of faster options close on its heels.

NVidia has great hardware engineers, but awful software driver people on their mobile platform. They have done a terrible job supporting their chipsets after release with Android, or getting good manufacturers to adopt them.

Re:Tegra is a flop.

[gl4ss]

Ummm Tegra 2 was the fastest platform for Android for quite some time. The G Tablets are still pretty blazingly fast. The issue is just that Tegra 2 was released for such a short time before Tegra 3 came out that it never got much saturation, and then Tegra 3 came out with a bunch of faster options close on its heels.

NVidia has great hardware engineers, but awful software driver people on their mobile platform. They have done a terrible job supporting their chipsets after release with Android, or getting good manufacturers to adopt them.

so is it fast or is it not? the terrible sw excuse was employed by intel for many years as to why their gpu's sucked since they weren't as fast as their promises..

Re:First HTC drops its quad core chip for a dual c

[Anonymous Coward]

The problem is Tegra does not support LTE - this is a problem for sales in the USA.
Europe will continue to get the full spec, full speed, Nvidia Tegra3 devices

Re:First HTC drops its quad core chip for a dual c

[Sollord]

It has nothing to do with the quad core aspect of the Tegra3 and everything to do with Nvidia being stupid and not bothering to get it supported/certified by any of the LTE chipset manufacturers.

Misinformed

[TheBogBrushZone]

They announced two Infinity models. Once of them has LTE/3G and the dual-core S4, the other is Wi-fi only and is still toting a 1.6GHz Tegra 3.

Re:Misinformed

[Anonymous Coward]

^^ This. Completely - see here http://www.everythingabouttablets.net/2012/02/27/asus-launches-new-transformer-pads/

"shrounded"?

[Anonymous Coward]

Really?

Re:"shrounded"?

[Anonymous Coward]

"shrounded", adj. both radiused and neuroactive, like 'shrooms.

Substandard abstract now standard?

[Anonymous Coward]

>In a move that will shock and disgust bleeding-edge technophiles everywhere

You obviously don't know what technophile even means. Why would they be shocked and/or disgusted by someone using the latest and greatest in technology.

LTE?

[RenHoek]

What the hell is an LTE? Nowhere can I find what this acronym means.

Re:LTE?

[Anonymous Coward]

http://en.wikipedia.org/wiki/3GPP_Long_Term_Evolution

Re:LTE?

[EmagGeek]

It's the "latest and greatest" 3G cellular data network.

Re:LTE?

[gbjbaanb]

seriously?

http://lmgtfy.com/?q=lte

its a network technology like GPRS, but better. They call it "4G" in some places, but it should be considered 3.5G.

Re:LTE?

[gl4ss]

hspa(+) is more like 3.5G, been for years marked like that on actual selling shipping devices.

if what's commonly called LTE had things like.. well, voice spec, standard way to handle video and all that, then it would be more like 3.5G. it's just another network for the yank operators to ask for mo' money whilst not upgrading their hspa network to bearable levels where they could sell it without transfer limits..

a more proper name for currently out there LTE networks would be short term evolution though! it's so unfinished spec that it's not a spec..

Re:LTE?

[Svartalf]

Long Term Evolution

Next time, though, Just F'n Google It.

Re:LTE?

[RenHoek]

I did Google it, I just had no clue it had anything to do with cellular technologies. I thought it was a CPU feature, so I disregarded the 3G link.

Re:LTE?

[RenHoek]

I did Google it, I just had no clue it had anything to do with cellular technologies. I thought it was a CPU feature, so I disregarded the 3G link..

Re:LTE?

[Anonymous Coward]

um. LTE *is* the upgrade from hspa, with more bandwidth and lower latencies, but requiring new network hardware and software. It's also not "a spec", LTE networks are operational e.g. in Korea and Japan.

Skype on standby

[21mhz]

Standby battery life with Google sync, a few IM clients (I run Skype and imo.im), Whatsapp, Viber and so on, should be around 4-5 days.

I wonder what Skype did to achieve this. On N900 and N9, the Skype engine is the monstrous wakeup hog that drains your battery in a day and exchanges packets with various hosts on the network all the time.
Did they subscribe to some push wakeup mechanism where the app can be launched on incoming activity that needs user interaction?

Re:Skype on standby

[bemymonkey]

Android has some sort of a built-in low power push-like mechanism that was implemented starting with Android 2.2 (Froyo), called C2DM. It's not quite real push, but the battery life is stupid good.

I'd assume Skype uses C2DM, as do most IM apps...

http://code.google.com/android/c2dm/

Re:Skype on standby

[gbjbaanb]

not on Windows Phone it doesn't.

Skype calls can be made and received when it's running in the foreground, but as soon as you switch away - even mid-conversation - the application goes offline. Want to check a detail in an e-mail so you can tell the person you're calling? You'll have to hang up first.

$8.5bn for Skype, shame they couldn't chuck a few dollars at getting multitasking for the phone working.

Re:Skype on standby

[TonyMillion]

Because of the way the Skype network works (its a p2p mesh system), depending on whether the client elects to be a leaf, a node or a super node will depend on the amount of network activity you see.

The protocol was *really* not designed for mobile devices, and we are starting to see how badly it can treat the battery. There isn't much room to expand it to use push notifications either.

Re:Skype on standby

[bemymonkey]

Not on Windows Phone it doesn't... what? Use C2DM? No surprises there, that's an Android thing that's not available on Windows Phone ;)

It really is unfortunate that Microsoft kludged the system like that - VoIP is a no-brainer for multitasking... :(

Re:Skype on standby

[21mhz]

It really is unfortunate that Microsoft kludged the system like that - VoIP is a no-brainer for multitasking... :(

Lync works on Windows Phone. I don't think it has VoIP, but for background lurking IM should be served the same way. What makes you think it's a system limitation?

Re:Skype on standby

[bemymonkey]

The article he linked to.

"Though Windows Phone 7.5 introduced new multitasking capabilities, allowing applications to do things like play music in the background and run periodic scheduled tasks, it only supports certain scenarios. VoIP is not one of those supported scenarios."

Re:Skype on standby

[21mhz]

The article he linked to.

"Though Windows Phone 7.5 introduced new multitasking capabilities, allowing applications to do things like play music in the background and run periodic scheduled tasks, it only supports certain scenarios. VoIP is not one of those supported scenarios."

In contrast, perhaps, with the author of this article, I know a thing or two about how VoIP protocols typically work. Receiving a "riser" message signaling an incoming call is not entirely dissimilar to receiving a chat message. If an application can subscribe to the latter, it could use the same mechanism to be notified about incoming calls. After being activated, the application can proceed to establish the VoIP session. Hence my doubts about multitasking being the problem.

Re:Skype on standby

[bemymonkey]

It's not the notification that's the problem, but rather that the call is ended when you switch to a different app.

Re:Skype on standby

[21mhz]

I see. I guess this piece of the puzzle can only be solved with baked-in Skype support, not in an installable app. Unless they have offered the "remain as an obtrusive overlay on top of the screen" behavior in public APIs, that the stock phone UI uses when switched away from.

Re:First HTC drops its quad core chip for a dual c

[fearofcarpet]

I bought a Prime as soon as they were available where I live. The first time I switched it on, it updated to ICS--great. I headed over to XDA-developers to see how to root it, and found stickies dedicated to the various problems that the devices have--random reboots, lockups, terrible WiFi performance, and so on. It seems that these problems are related to the serial numbers, too. ASUS even has an "official" support thread on XDA in which (what I presume is) an engineer fields questions about said random problems. The long and short is that there appears to be some serious quality control issues, not just with the aluminum cases (and the antennas not making contact internally) but with the chipset.

I am by no means saying that there is something wrong with Tegra 3, but it would explain the seemingly-binary mix of "my Prime works great" and "well, it hasn't rebooted in three hours since the update--damn, there it goes again," the frequent (twice just last week) OTA updates, and the apparent correlation to serial numbers; and now ASUS dropping it altogether. I've been lucky in that my Prime only suffers from terrible WiFi and nonexistent GPS reception (but, really, GPS in a tablet?), but the WiFi is getting better with each update (even though ASUS claims the cause is the aluminum case.)

Whatever--two cores, five cores, 11 cores--my Prime is faster than greased lighting, doesn't eat the battery, and multitasks like a boss. I can't imagine ASUS would release an improved version that wasn't even faster.

Re:First HTC drops its quad core chip for a dual c

[EmagGeek]

The problem for nVidia is that, the microsecond someone comes out with a similarly-performing platform on a smaller process, the larger process product is obsolete.

Halving the process size equates to a quartering of cost, because 4 times as many elements can be placed in the same area. So, shrinking from 40nm to 28nm about halves the cost for the same performance, and increases battery life, which we all know is everything in the tablet market, for some stupid reason.

Parallelization Overhead

[mysidia]

It is a net application performance increase if you can halve the number of CPUs and double the performance of each CPU.

Mem BW limited

[Anonymous Coward]

Based on actual real world performance analysis memory bandwidth was an issues already with a single core A8, no wonder they have scaling issues with quad A9s.

Asus is failing me

[drinkypoo]

Transformer prime with a shitty GPS antenna, now the descendants with hopefully fixed antenna have an inferior CPU. I personally want lots of cores, they make sense for me. Asus is being really creative at finding ways not to take my money.

Re:Asus is failing me

[Sollord]

Actually according to theverge.com and gizmodo.com the descendants will have a Tegra 3 if you only want WiFi it's the 3g and LTE versions that will use the S4 due to Tegra 3 lacking LTE at this time and based on the Anandtech reviews of the S4 it's dual cores pretty much win all the CPU tests and most of the GPU tests compared to a Tegra3.

http://www.anandtech.com/show/5563/qualcomms-snapdragon-s4-krait-vs-nvidias-tegra-3

maybe the article is flat out wrong

[way2trivial]

quadcores CAN'T work with US cellular networks- presently
http://www.pcworld.com/article/250769/atandts_htc_one_x_flashy_smartphone_quadcore_everywhere_but_the_us.html

" Nvidia's processor won't be compatible with LTE radio chipsets for at least a few months and ,with the One X due to launch stateside within 60 days, AT&T wants a version of the phone that supports 4G LTE."

now the above url is about a quadcore android phone planned for US release only as a two core, but the same datum likely applies- the quadcore is not compatible.

there was a technical limitation of existing 4cores and US mobile tech. They dolled the argument to fit, " the fact that quad-core processors don't really have a use case in mobile devices is one reason" sounds like marketing speak in this circumstance.

Re:maybe the article is flat out wrong

[Guspaz]

You've taken a big (and wrong) leap to go from "Tegra 3 doesn't currently work with any LTE radios" to "Quadcores CAN'T work on US cellular networks- presently"

You could factually say "There are not currently any quad core SoCs on the market that are compatible with LTE".

Re:maybe the article is flat out wrong

[fast turtle]

The problem is, the Tegra does not have an LTE radio/modem on the SoC. This means it's a seperate chip and the current chips do not pla nicely with the Tegra. Qualcomn has pulled a major upset with their new design as the SnapDragon is faster, uses less power and includes the needed LTE radio on the die. The only area where the Qualcomn offering sucks tits on a worm is graphics as the Tegra beats the hell out of it like a red headed step-child but that's the only thing it wins.

Re:maybe the article is flat out wrong

[ChronoReverse]

That's not necessarily true. The Adreno 225 in the current Krait is slower than the 320 that'll be coming later in the year but it still trades benchmarks with Tegra3 in graphics tests.


Notably, the GLbenchmark offscreen test where the Tegra3 wins the most by, has a severe bug somewhere.

'Shrounded' voted best new word.

[Anonymous Coward]

Shrouded? Surrounded? Why not take a bet each way?

Multicore misunderstandings

[Anonymous Coward]

Some of the comments are pathetic. Here's the tradeoff. More cores means more apps can run concurrently. That isn't a big deal with tablets yet because most of the operating systems only let you run one or a few apps at the same time. However, there are many background notification systems, chat and other things that can benefit. There's a certain upper limit where it's just crazy to have more cores for a tablet because they won't get used.

Amdal's law kicks in after a time and it is slower to have 4 cores than 1 for an app but with lockless algorithms and some other research areas, it's becoming less of a performance hit to do SMP. It's a no brainer that any computing device can benefit from 2 cores. I think it's even reasonable to have 4 cores in a high end tablet. They will get used for games, mail apps, and especially web browsing.

the S4 is much faster than the Tegra's A9 cores

[Anonymous Coward]

Its been benched on Anandtech. Take a look here http://www.anandtech.com/show/5563/qualcomms-snapdragon-s4-krait-vs-nvidias-tegra-3.

So yes, 2 faster cores (almost by a factor of 2), is much better than 4 slower cores.

Re:First HTC drops its quad core chip for a dual c

[jader3rd]

I've been waiting and salivating at the promised "Quad" core offerings for smartphones.

Why? Unless you know for sure that a regular task on your phone is CPU bound, adding more cores will just add overhead without making anything faster.

Re:First HTC drops its quad core chip for a dual c

[AlienIntelligence]

I've been waiting and salivating at the promised "Quad" core offerings for smartphones.

Why? Unless you know for sure that a regular task on your phone is CPU bound, adding more cores will just add overhead without making anything faster.

Thanks for asking... I did actually imply, while not going further, that I was aware that my phone was CPU bound.

And so... going further... yes, I see some processes lock both cores up tightly... for seconds at a time. 100% util
to (all) cores, is a good sign that adding cores will help, even if partially.

This happens most frequently... when I need to use the phone the most. Such as, turning it on, after not using it
for let's say, over an hour. When I turn it on, all of the apps which have been happily asleep stretch their arms and
legs and download everything they missed.

And, I'm obviously turning the phone on cause I need to use it for something, so let's say that I already hit my
unlocker to go to text messages, or docs, or the camera. Now, I'm standing there, looking at my task manager
just shoot up to the top 100% util for countless seconds... it's countless if you REALLY wanted that camera
open RIGHT THEN. If it's an extended update that the apps are doing, then I might as well just give up.

Addressing further comments about how much I actually use my phone... my battery app says my phone is
idle 52% of the time. So, a lot.

-AI

Re:First HTC drops its quad core chip for a dual c

[LoveMuscle]

Yes there a "secret".. It's Ahmdal's Law. (http://en.wikipedia.org/wiki/Amdahl%27s_law)
Mobile computing tends to be single task centric, and doesn't yet have the high level of parallelism to utilize a high core count..
It's not rocket science, it's just regular science, and they don't have to play the marketing core count game since there's a dual core now that outperforms...

Re:First HTC drops its quad core chip for a dual c

[jader3rd]

This happens most frequently... when I need to use the phone the most.

Well it's great that you actually have evidence that this is your problem vs. the attitude of bigger-number-must-be-better. It's interesting though how we can see the same problem and think about how to solve it differently. You see this problem and think "Throw more hardware at it." I see that and think "Needs better software".
It's my opinion that a device OS should know that it's a device and not a general computing machine. As a result, not all tasks are equal. The platform should have ways of making sure that the background tasks don't interfere with you using the device. Do apps need to check for updates and status changes? sure. But a good platform wide notification system, and aggresive thread scheduler, could also solve the problem of background tasks stomping on the user experience, and still run on your existing hardware.

The could use Tegra 3 with LTE

[tlambert]

Qualcomm does sell discrete LTE components separate from their CPUs, like the MDM9200 or MDM9600. Juniper also sells discrete components, and probably would not force the bundling.

-- Terry

Re:First HTC drops its quad core chip for a dual c

[AlienIntelligence]

This happens most frequently... when I need to use the phone the most.

Well it's great that you actually have evidence that this is your problem vs. the attitude of bigger-number-must-be-better. It's interesting though how we can see the same problem and think about how to solve it differently. You see this problem and think "Throw more hardware at it." I see that and think "Needs better software".

Ahhh but we do actually agree... 100% in fact. It is, definitely, the software.

However.

I am practical, in so far as... I don't have time to waste, to message a dozen app devs that
their app isn't playing nicely with others. When I can just hope that a phone manufacturer
hears my pleas and makes the hardware, faster.

It's like herding cats vs gently nudging an already rolling boulder in a more favorable direction.

-AI

Re:First HTC drops its quad core chip for a dual c

[AlienIntelligence]

The problem is Tegra does not support LTE - this is a problem for sales in the USA.
Europe will continue to get the full spec, full speed, Nvidia Tegra3 devices

It has nothing to do with the quad core aspect of the Tegra3 and everything to do with Nvidia being stupid and not bothering to get it supported/certified by any of the LTE chipset manufacturers.

Ahhh, wow, I wish they would have asked their power-user customers, such as myself.

I could give two craps about LTE right now. I'd rather have the hardware that is ready
to go, in favor of something, that I need to wait for them to "roll-out" and make usable.

I'm a data glutton... but I'm usually near enough to a usable wi-fi that my cell speeds
are moot. And when I'm not... I'm usually not interested in using high-bandwidth, since
that is a large portion, directed at entertainment. [Ugh, I just saw their rationale, dammit]

oh well, no sense crying over spilled cores.

-AI

Re:First HTC drops its quad core chip for a dual c

[Sollord]

Lucky for you then since only the 3g and 4g models sold by carriers will use the S4 the Wifi model will be Tegra3 though the Tegra3 seems to have a slightly better GPU most tests show the S4 having a far better CPU see anandtech.com for details

Dual Core..

[Anonymous Coward]

Dual Core ought to be enough for everybody

Re:First HTC drops its quad core chip for a dual c

[AlienIntelligence]

Lucky for you then since only the 3g and 4g models sold by carriers will use the S4 the Wifi model will be Tegra3 though the Tegra3 seems to have a slightly better GPU most tests show the S4 having a far better CPU see anandtech.com for details

I did actually... very impressive.

It does seem at least for now, that throwing cores at the issue,
isn't a problem solver. The S4 was pretty impressive in HD vid
decoding.

I wish there was a real-world test, not just simple benchmarks.

A script that would turn a phone on... so it will start sync'ing,
then run a series of apps that one would 'commonly' run.

The page loading for the browser was nearly identical to the
Tegra 3. So, most of that action must run on a single thread
model. And the decoding abilities helped the S4 as well.

Being on a 28nm is a benefit too, to battery life. But I really
just hate the idea of wasting my upgrade on a dual core,
when I already have one.

I guess I'll wait til the end of the year when the quad Tegra 3
with LTE comes out.

-AI

On Flashing and Bootloaders

[rdnetto]

Frankly, I'm more interested in the fact that they're using a Qualcom chip instead of a Tegra. All the Tegra series devices can be reflashed and repartitioned using nvflash, but only if the manufacturer decides to release 128-bit AES key. It'll be interesting to see how this one can be flashed, if at all. Asus released a fastboot-based method of reflashing the prime, but as I understand it doesn't support repartitioning.

Addendum: To those of you wondering why you'd want to be able to repartition the device, consider that Linux runs rather well on it and gets insane battery life compared to x86. If you can only reformat existing partitions, then you lose access to somewhere between 2 and 5 GB of smaller partitions.