Beta

×

Welcome to the Slashdot Beta site -- learn more here. Use the link in the footer or click here to return to the Classic version of Slashdot.

Thank you!

Before you choose to head back to the Classic look of the site, we'd appreciate it if you share your thoughts on the Beta; your feedback is what drives our ongoing development.

Beta is different and we value you taking the time to try it out. Please take a look at the changes we've made in Beta and  learn more about it. Thanks for reading, and for making the site better!

Apple A4 Processor Teardown

timothy posted more than 4 years ago | from the coring-the-thing dept.

Apple 79

Plocmstart writes "Here's what EETimes.com is claiming to be the first teardown of the A4 processor. 'Apple's iPad chip is a single-core ARM A8 made by Samsung. Through various benchmarking testing, UBM TechInsights was able to find out the details of the A4 processor.'"

cancel ×

79 comments

Sorry! There are no comments related to the filter you selected.

Interesting (4, Interesting)

Pojut (1027544) | more than 4 years ago | (#32199838)

Anyone know what Google is planning on stuffing in their new tablet [gizmodo.com] ?

Re:Interesting (-1, Offtopic)

jeepmeister (241971) | more than 4 years ago | (#32199852)

Uh, FP

Re:Interesting (-1, Offtopic)

Anonymous Coward | more than 4 years ago | (#32200110)

You mean CP?

Re:Interesting (1)

sznupi (719324) | more than 4 years ago | (#32199948)

No reason for them to go with anything else than also one of the recent ARM SoC.

Re:Interesting (5, Insightful)

PopeRatzo (965947) | more than 4 years ago | (#32200616)

Anyone know what Google is planning on stuffing in their new tablet ?

Advertisements.

Re:Interesting (2, Funny)

obarthelemy (160321) | more than 4 years ago | (#32200748)

spyware !

Re:Interesting (3, Insightful)

michaelhood (667393) | more than 4 years ago | (#32204002)

Anyone know what Google is planning on stuffing in their new tablet ?

Advertisements.

Slashdot broke my mod points somehow, but this deserves to be modded to plus infinity. However, as anyone who saw the iPhone OS 4.0 presentation knows, there will be plenty of ads on the iPad too soon enough.

Re:Interesting (-1, Troll)

jo42 (227475) | more than 4 years ago | (#32200774)

Anyone know what Google is planning on stuffing in their ...

Android Butt Plug

Couldn't resist.
- T. Roll

Re:Interesting (3, Funny)

biryokumaru (822262) | more than 4 years ago | (#32203482)

Hrm... That brings a whole disturbing new level to the question "Do Androids Dream of Electric Sheep?"

8 minute abs (2, Funny)

goombah99 (560566) | more than 4 years ago | (#32201748)

Well since apple came out with the 4 minute Abs, I assume google will have to go with the 3 minute abs. ie. the A3.

Re:Interesting (0)

Anonymous Coward | more than 4 years ago | (#32205920)

Anyone know what Google is planning on stuffing in their new tablet [gizmodo.com] ?

WIN. Lots and lots of WIN.

No Way (5, Funny)

LordBmore (1794002) | more than 4 years ago | (#32199990)

This can't be right... TFA doesn't make a single reference to magic.

Re:No Way (1, Funny)

Anonymous Coward | more than 4 years ago | (#32200124)

As anyone who's dabble in electronics will tell you, opening or frying an IC lets the pixies escape.

Re:No Way (4, Insightful)

edittard (805475) | more than 4 years ago | (#32200192)

Are you saying this isn't so much a teardown as a lookat, a thinkabout or a getdrunkandspeculatewildly?

Re:No Way (3, Funny)

Anonymous Coward | more than 4 years ago | (#32201316)

Ah, yes. I once accidentally reversed the 5 V and 12 V leads on a harddrive and saw the magical blue smoke escape from controller chip. Electronics work only so long as that smoke stays inside the chips! (I wonder how they get in it in there in the factories...)

Re:No Way (1)

biryokumaru (822262) | more than 4 years ago | (#32203498)

I wonder how they get in it in there in the factories...

Oh, it's the same way they get the coke in the coke bottles, the factory is actually filled with smoke.

also, multicore? (2, Insightful)

poetmatt (793785) | more than 4 years ago | (#32200138)

also, I thought the world claimed it was multi-core? [tipb.com] This teardown summary says it's single core.

A4 is a System-on-a-Chip, or SOC, that integrates the main processor [ARM Cortex-A9 MPCore i.e. Multi-Processing Core, identical to ones used in nVidia Tegra and Qualcomm Snapdragon]

meanwhile, from the summary:

number of ARM cores: 1

hmm.

Re:also, multicore? (2, Informative)

atrus (73476) | more than 4 years ago | (#32200330)

Multicore was wild speculation. It was pretty much obvious that it was a Cortex-A8 once the first units hit the market.

Re:also, multicore? (0, Troll)

the linux geek (799780) | more than 4 years ago | (#32201166)

It's also an A8, not an A9 as has occasionally been claimed. Apple fans will do anything to hype their fairly weak chip, it seems.

Re:also, multicore? (2, Insightful)

gumbi west (610122) | more than 4 years ago | (#32202072)

Find a review that says the UI slow.

Re:also, multicore? (3, Insightful)

negRo_slim (636783) | more than 4 years ago | (#32202268)

What does the UI have to do with processor performance?

Re:also, multicore? (3, Insightful)

soppsa (1797376) | more than 4 years ago | (#32207322)

It's all that matters in the end. This device isn't for you to play Crysis on or compile your gentoos.

Re:also, multicore? (0)

Anonymous Coward | more than 4 years ago | (#32209062)

Tons, think of playing a game of chess with a timer vs playing over email.

Re:also, multicore? (2, Interesting)

devjj (956776) | more than 4 years ago | (#32202296)

Apple's just proving the point that consumers don't care how "fast" a chip is if the experience is that the device is fast.

lol (-1, Troll)

Anonymous Coward | more than 4 years ago | (#32203432)

that sounds like the kind of tripe your wife tells you to console you about your toddler penis.

"it feels big, even if it doesn't look big..."

you are pathetic.

Re:lol (1)

Tak_1 (1684082) | more than 4 years ago | (#32210416)

So, basically, you are saying that anon trolls are insecure children who see everything in direct proportion to their "minuscule" pens size? Thank you, I know understand the driving force of trolls. Explains a great deal really.

Re:also, multicore? (2, Insightful)

trouser (149900) | more than 4 years ago | (#32204040)

Apple's just proving the point that consumers don't care how "fast" a chip is if the experience is that the device has an Apple Logo.

Re:also, multicore? (4, Insightful)

BasilBrush (643681) | more than 4 years ago | (#32205144)

Brand loyalty first has to be earned. Apple didn't get lucky, they created a lot of great products and earned their brand loyalty.

Re:also, multicore? (2, Insightful)

tehcyder (746570) | more than 4 years ago | (#32207544)

Hi Steve!

Re:also, multicore? (1)

BasilBrush (643681) | more than 4 years ago | (#32210434)

Hi Bill. Are you stalking me again?

Re:also, multicore? (0, Redundant)

jo_ham (604554) | more than 4 years ago | (#32205560)

Wait, I thought all "Apple fans" were mindless sheep who thought their Apple products were magical white shiny boxes powered by unicorn farts. Why would they know the difference between an ARM A8 and A9?

Or are you saying that Apple fans do actually have some discerning knowledge about the hardware they use?

Re:also, multicore? (1)

SolusSD (680489) | more than 4 years ago | (#32206386)

Weak chip? The intent was good performance and good battery life. The iPad gets ~10 hours on a single charge and nothing about the iPad feels sluggish. A lot of this has to do with a very optimized software stack (xnu + darwin + cocoa touch ui layer) with apps written in Objective-C (managed memory instead of garbage collection, natively compiled). It is quite a bit more efficient than even Android's optimized software stack, which runs apps on a stripped down JVM. There's more to building a responsive system than throwing ever faster CPUs at it.

Re:also, multicore? (1)

guhknew (123675) | more than 4 years ago | (#32216294)

Sorry. Undoing a bad moderation.

Re:No Way (2, Funny)

WrongSizeGlass (838941) | more than 4 years ago | (#32200172)

MOV r0, 'Magical' ;
LDR r1, r0 ;


There you go, all better ;-)

Re:No Way (2, Informative)

marcansoft (727665) | more than 4 years ago | (#32201644)

Does not assemble. Try this:

get_magic:
                ADR r0, magic
                BX lr

magic: .ascii "Magical\0"

No but it does have neon. (4, Informative)

leuk_he (194174) | more than 4 years ago | (#32200354)

It does have Neon [arm.com] but no powerpc. Strange that this information did not came from someone with a compiler. Does apple withhold information what code can be generated? Are devs so spooked by the apps license?

Re:No but it does have neon. (2, Interesting)

ShakaUVM (157947) | more than 4 years ago | (#32200448)

>>Does apple withhold information what code can be generated?

The only officially approved code generation method for the iPad is to send a nicely worded letter to Apple, along with X dollars. In Y months your app will be compiled for you and put up on the app store, for which you'll receive no royalties.

Re:No but it does have neon. (2, Informative)

FuckingNickName (1362625) | more than 4 years ago | (#32200958)

I wish it were that easy! In the real world, developers go through all this and have to worry about their app being rejected.

Re:No but it does have neon. (0)

Anonymous Coward | more than 4 years ago | (#32201264)

The only option I've seen in xcode is Armv7 code... I haven't seen any info on the instruction set from Apple, but I also haven't really looked.

iPhone OS now targets both Armv6 (original iPhone, iPhone 3G and first 2 generations of iPod touch), as well as Armv7 (supported by iPhone 3GS + 3rd gen iPod touch). Additionally, Apple requires building for Armv7 if you wish to distribute your app as a native app to the iPad.

Basically, in order to build an app for both iPone and iPad (Universal), you need to compile your executable for both instruction sets. This is really easy. It is just an option in xcode and your executable ends up around twice the size. It is actually pretty cool that you can do it.

Based on the fact that we already need to think about multiple instruction sets now, I would think that Apple discourages any form of direct assembly code, but to be honest I really don't know.

Re:No but it does have neon. (1)

sznupi (719324) | more than 4 years ago | (#32203876)

Neon, which...pretty much every ARM Cortex has, I believe.

Re:No but it does have neon. (0)

Anonymous Coward | more than 4 years ago | (#32204016)

Well, sort of. Every Cortex A perhaps. Cortex M microcontrollers... not so much!

Re:No Way (1)

gyrogeerloose (849181) | more than 4 years ago | (#32201098)

This can't be right... TFA doesn't make a single reference to magic.

It's a given. All electronic components operate via the Magic Smoke they contain. That's why it's such a big concern when see smoke rising from a device--once the Magic Smoke has leaked out, they stop working.

Re:No Way (1)

negRo_slim (636783) | more than 4 years ago | (#32202310)

I have a feeling a vast majority of people do think of computers and electricity as not necessarily the dictionary's definition of magic, but something very similar.

Re:No Way (2)

gyrogeerloose (849181) | more than 4 years ago | (#32202368)

I have a feeling a vast majority of people do think of computers and electricity as not necessarily the dictionary's definition of magic, but something very similar.

Indeed. The obligatory Arthur C. Clarke quote:

"Any sufficiently advanced technology is indistinguishable from magic."

definitely applies. Even when you know something about the technology, it can still seem like magic.

Re:No Way (1)

blueZ3 (744446) | more than 4 years ago | (#32207948)

IMO, that's true of pretty much any "technology" these days.

Even something as "simple" as a car--ask your average person about how their car runs and they'll probably be able to tell you that the engine uses gasoline and that's pretty much it. You put the key in the ignition, turn it, and the magical transportation fairies start working.

A big part of this is due to specialization. Most folks have some particular thing that they're good at/focused on. But ask for details about how something works that's outside their area of expertise and you're likely to get an explanation that makes little sense to someone who really knows the answer.

Not that there's anything necessarily wrong with specialization, it's just something to be aware of. And anyway, as long as you get the results you expect, does it really matter how it works? That information is only really "useful" when something goes wrong. Which is all too often, I know :-)

Re:No Way (1)

michaelhood (667393) | more than 4 years ago | (#32204014)

This can't be right... TFA doesn't make a single reference to magic.

I wish we had the original diagrams, maybe we could at least find the revolution generators..

(why parent post is funny [apple.com] )

ARTICLE IS GOATSE REDIRECT (-1, Troll)

Anonymous Coward | more than 4 years ago | (#32199996)

Do not click! Not Safe For Work!!

how do these articles keep getting through?

Re:ARTICLE IS GOATSE REDIRECT (2, Insightful)

Darkness404 (1287218) | more than 4 years ago | (#32200096)

This is /. we do not read TFA, therefore it doesn't matter if it is Goatse or not, no one is going to see it.

Congratulations. (4, Funny)

spun (1352) | more than 4 years ago | (#32200304)

You found a way to make Slashdotters actually click through to the article. Now, if only you could find a way to make them read it.

Re:Congratulations. (0, Troll)

commodoresloat (172735) | more than 4 years ago | (#32200760)

You found a way to make Slashdotters actually click through to the article. Now, if only you could find a way to make them read it.

Say it's a two girls one cup redirect instead?

Useful / single page URL (5, Informative)

BitZtream (692029) | more than 4 years ago | (#32200008)

http://www.eetimes.com/showArticle.jhtml?articleID=224701036&printable=true&printable=true [eetimes.com]

Next time guys, save us the effort and use the link that doesn't require us to click next 4 times to read an article that fits on half a page.

Oh ... timothy, nevermind.

Re:Useful / single page URL (1)

Hatta (162192) | more than 4 years ago | (#32200284)

I was too busy giggling about "UBM" to notice. UBM? I do!

Re:Useful / single page URL (1)

Bearhouse (1034238) | more than 4 years ago | (#32202458)

Yup. Timmy is the new Zonk

Here's The iPad Killer (0, Troll)

Anonymous Coward | more than 4 years ago | (#32200014)

iPad Killer [qi-hardware.com]

Take the crapPad and Blend [youtube.com] it !

Yours In Smolensk,
Kilgore Trout.

Re:Here's The iPad Killer (1)

WrongSizeGlass (838941) | more than 4 years ago | (#32200220)

Wow, that iPad Killer comes with a free can of Coke? That's pretty impressive. It looks like a Franklin Electronics address book.

Re:Here's The iPad Killer (1)

Antiocheian (859870) | more than 4 years ago | (#32200548)

No, it's the Coke that comes with a free iPad Killer

Re:Here's The iPad Killer (1)

TheRaven64 (641858) | more than 4 years ago | (#32201370)

I played with a Ben NanoNote at FOSDEM and, while it's an amazing device, it's nothing like the iPad. The screen is really tiny - only 320x200 pixels, which is not enough even for most terminal apps (40x15 characters), let alone graphical apps. The keyboard is really tiny too and it has a very slow CPU and little RAM.

The strength of the NanoNote is that it really is tiny. It's smaller than a typical wallet - you could slip it into your trouser pockets and take it everywhere. Running a web browser on it would be painful, but if you want a computer that you can take absolutely anywhere, then it might be interesting. It's also so cheap that you won't worry too much about losing it.

Re:Here's The iPad Killer (1)

tehcyder (746570) | more than 4 years ago | (#32207638)

In your useful review, you seem to have overlooked the points of comparison with the iPad. That's because there aren't any, so fuck knows what the GP was talking about.

Upshot: display tech dominates battery life (5, Interesting)

Anonymous Coward | more than 4 years ago | (#32200574)

They show that power usage is:
0.5 Watts at idle with the display disconnected
1.75 Watts at idle with display at min brightness
4.25 Watts at idle with display at max brightness

They estimate that the A4 CPU uses 0.5 to 0.8 Watts when browsing the web over WiFi.

Conclusion, FTA:

Until display technologies make big moves downward in power consumption the consumer experience of battery life may be driven as much or more by the LED drivers, opto-mechanical design of the back-lighting, display settings and wireless connectivity employed versus the CPU itself.

Re:Upshot: display tech dominates battery life (0)

Anonymous Coward | more than 4 years ago | (#32202492)

If a man has a strong faith he can indulge in the luxury of skepticism. -- Friedrich Nietzsche

Re:Upshot: display tech dominates battery life (1)

sznupi (719324) | more than 4 years ago | (#32203816)

Goof thing Pixel Qi is finally ramping up production to appreciable levels, with several upcoming products using their screen (also tablets)

Sure, they are "cheating" by turning the backlight down in bright ambient light situations...so? The effect is what counts.

Re:Upshot: display tech dominates battery life (1)

afidel (530433) | more than 4 years ago | (#32204294)

PixelQI fixes this today, the cost needs to be worked on, but the technology is a solved problem.

Re:Upshot: display tech dominates battery life (1)

quenda (644621) | more than 4 years ago | (#32205016)

Considering its a cellphone CPU attached to a laptop screen: No shit!

teardown my ass (2, Funny)

veg_all (22581) | more than 4 years ago | (#32200728)

tfa:

With further analysis, including chip-level reverse engineering, we may be able to identify whether innovations such as Intrinsity’s patented cell libraries were used to optimize the critical paths in the ARM core itself.

IOW, "It looks kinda neat, wonder what's inside!"

Re:teardown my ass (0)

Anonymous Coward | more than 4 years ago | (#32201108)

Not only is this not a teardown, its not the first by a long shot.

Chipworks seems to have beaten them to the punch - most of this report was available the day after the iPad was launched.

Do your homework EETimes!

http://chipworks.com/Apple-iPAD-Teardown.aspx

Re:teardown my ass (0)

Anonymous Coward | more than 4 years ago | (#32203096)

Not only is this not a teardown, its not the first by a long shot.

Chipworks seems to have beaten them to the punch - most of this report was available the day after the iPad was launched.

Do your homework EETimes!

http://chipworks.com/Apple-iPAD-Teardown.aspx

This may surprise you, but the iPad is not its own CPU.

L2 Cache! (1, Insightful)

Anonymous Coward | more than 4 years ago | (#32200734)

Did you see the size of the L2 cache? It ought to be enough for everybody!

Next steps (2, Insightful)

Yvan256 (722131) | more than 4 years ago | (#32201642)

A4 for iPad
A4 for next iPhone
A4 for next iPod touch
A4 for next AppleTV, possibly let it play games too.

Re:Next steps (1)

JamesP (688957) | more than 4 years ago | (#32201848)

And then A5 maybe?!

Re:Next steps (1)

nyctopterus (717502) | more than 4 years ago | (#32205024)

I'm waiting for the A5 Powerbook, it's gonna be awesome!

Re:Next steps (3, Funny)

Anonymous Coward | more than 4 years ago | (#32202314)

A4 for President

Re:Next steps (2, Funny)

duckbillplatypus (596100) | more than 4 years ago | (#32204132)

i prefer u.s. letter size.

So that's why... (0)

Anonymous Coward | more than 4 years ago | (#32203290)

They can only run a single app at a time.. Only a single core. This explains everything.

Re:So that's why... (1)

Anonymous Coward | more than 4 years ago | (#32205974)

They can only run a single app at a time.. Only a single core. This explains everything.

Yes, before Intel's Core Duo CPUs, people on desktop computers could also only run one app at a time.

Oh? (1)

FredMenace (835698) | more than 4 years ago | (#32204800)

"Here's what EETimes.com is claiming to be the first teardown of the A4 processor

Oh really? [ifixit.com]

They didn't see the hardware flash accelerator (1)

alfredos (1694270) | more than 4 years ago | (#32205316)

Which is disabled by software anyway

Interesting in a left-handed sort of way (1)

Whuffo (1043790) | more than 4 years ago | (#32205374)

I think we can all agree that the iPad isn't magical. I've got one and it doesn't support any kind of magical function that I can find. But the claim that it's revolutionary or world-changing seems to gain some support from the subject of this article. If it's nothing special, why is a detailed analysis of its CPU important or interesting?

No mystery why the iPad is faster than iPhone 3GS (1)

default luser (529332) | more than 4 years ago | (#32210610)

The question has been asked: how is the iPad so much faster than the iPhone 3GS [arstechnica.com] , despite having the same processor (sometimes more than double the speed)? The simple answer is: double the memory bandwidth!

iPhone 3GS: 32-bit memory bus, 600 MHz core
iPad: 64-bit memory bus, 1000 MHz core.

And this is assuming the memory technologies / clock speeds did not change. If they also increased the memory clock, the bandwidth increase could be 3x or more. And since rich web media craves memory, bandwidth is a key limiter for rendering complex websites. Thus, it's not hard to see how a %66 faster core speed could translate to a double-or-more average benchmark improvement.

Re:No mystery why the iPad is faster than iPhone 3 (1)

edxwelch (600979) | more than 4 years ago | (#32213322)

There's a huge flaw in that Arstechnica artical: they are using Safari to benchmark. Who's to say that Safari itself hasn't been optimised for the iPad? They should have tested using a custom benchmark app.

Check for New Comments
Slashdot Login

Need an Account?

Forgot your password?
or Connect with...

Don't worry, we never post anything without your permission.

Submission Text Formatting Tips

We support a small subset of HTML, namely these tags:

  • b
  • i
  • p
  • br
  • a
  • ol
  • ul
  • li
  • dl
  • dt
  • dd
  • em
  • strong
  • tt
  • blockquote
  • div
  • quote
  • ecode

"ecode" can be used for code snippets, for example:

<ecode>    while(1) { do_something(); } </ecode>