"Only utilizing one core" would mean that the software only uses one, but the hardware has two.
"Binning parts" means they're either actually hardware-disabling ("diking out") the extra core, or the extra core was bad to begin with (making use of otherwise bad chips).
If those images are real, both cores are definitely on the chip, but whether one is hardware-disabled or not is not certain.
Would make more sense to software disable it? Then when new hardware comes out the older hardware can be upgraded to support a new OS and utilise the second core rather than being left behind.
The simplest and most plausible explanation here is that they're using the Apple TV's relatively low performance demands as a way to get some use out of A5s with defective cores
As mentioned if they have switched all their 4S production over to this chip they may be able to support their Apple TV line primarily with dies that have a borked processor (otherwise those get thrown out). In that case only upgrading a portion of the people who bought Apple TVs with a software update might be counterproductive.
Apple TV was not about profit, Another possibility they might do a software update which would increase the performance (by activating the second core) when some new capabilities(server/cloud side) are available
That sounds bizarre to do with products in the wild. With chip yield being so fickle at such high concentrations, is having a core being disabled for most of the products' lifestyle a bad idea to activate suddenly, in case one core wasn't functioning properly?
Disabling hardware and then enabling it through software later to pretend that it's an update isn't how Apple operates and would be a really weird thing for them to do. The only thing I can think of that even comes close is when some Macs shipped with really early draft 802.11n hardware when the software wasn't ready, so they only supported 802.11g until a software update was made later. However, that was a case of the software and spec actually not being ready yet, not simply a matter of disabling something that would otherwise work fine just to have something to brag about later.
It wouldn't be unprecedented for them to release an update which enables features which were always in a chip but were unusable before, although it's not an identical situation:
The fact that this feature of that firmware update was so hidden that it was only noticed when people saw different behaviors makes it rather the opposite of what's proposed. Additionally, I seriously doubt that Apple deliberately withheld the ability to use 8GB of RAM in software only to add it back later. Much more likely it was a bug or just something that they didn't have time to finish. Neither would apply to running a single core of a dual-core processor in an Apple TV.
How about Bluetooth on the second generation iPod touch? That's a prime example of "chipset supported this, Apple locked it out with software and unlocked it with an update."
I wouldn't really call it "locking out". The Bluetooth stack in iOS was extremely incomplete before that update. If Apple had offered it at release, it would have only supported the Headset profile which is near useless on a touch. At the time Apple genuinely believed that with their accounting scheme, they couldn't give away hardware features without charging. To strike a balance between those concerns, we got Bluetooth in the iPhone OS 2 update for a fee.
Either way, that isn't what is happening here. This is just CPU manufacturing 101. New chip, new processes, high bin rate. You mitigate this as much as possible by disabling features on binned chips to make them sellable. Intel has done it a number of times, typically when making a big change like the introduction of the Core Duo.
Not related to the subject matter, but I bought an Apple TV just after the new generation was released (last week) and it is probably the best purchase I've made for my media set up. The ability to just play stuff straight from itunes on my display device (projector) wirelessly is amazing. Really cool piece of technology, which I had never heard about before I went looking for it. Apple TV seems very under advertised / talked about.
I really like mine as well, but for most people it's probably not the best streaming box on the market. I also have a Boxee for the codecs AppleTV doesn't support, and when I bought a streamer for my inlaws as a Christmas gift the Boxee is what I got them. It even does Airplay in the latest firmware.
I don't understand the love that boxee is getting. I also have one and it's a terrible device (albeit just cheap enough that some people might ignore the flaws).
It's loud, it's slow, the UI is pretty bad and unpolished. Sometimes it boots to a black screen or locks up randomly. Heck, 1 out of 5 times it doesn't even shutdown (just hangs on a black screen infinitely).
These are all documented issues (and really only the tip of the iceberg) that have gone unfixed for years.
As I said, I've got an AppleTV and the Boxee. The AppleTV is the one I prefer to use - the interface is the nicest of any streaming box I've used so far. Unfortunately, it really doesn't play anything except h.264. The Boxee is basically there to play whatever the AppleTV can't handle, and at that it excels. I rarely encounter a file Boxee isn't capable of decoding. The interface isn't the greatest, but having come from a cheap Popcorn Hour knock-off, it's certainly better than a lot of its competitors.
I really doubt that it's just the case of a software update to unlock cores. Beyond the questionable benefit of that second core, it's likely the reason that one core is running disabled (however it is) is cause Apple is trying out a new manufacturing process. Likely yields weren't consistent enough.
I agree. My hunch is yields and they will use the higher quality (verified both cores working) for the iphone5 expected to come out later this year, or maybe for the ipanel tv thing that is rumored.
Interesting that the same part is now being used in iPads 2. Sounds like they're getting ready for the next iPhone, maybe a speed bump for the retinal iPad.
I'd like to see if the new iPad 2s get better battery life. It stands to reason that they would, but Apple doesn't really talk about the iPad 2 anymore. I'm not willing to by an iPad 2 just to perform this test but some tech sites should do some battery benchmarks on the new iPad 2.
It makes sense to me that Apple is testing out yields for the next iPhone. This new, smaller manufacturing process will allow Apple to put LTE in the next iPhone by having the CPU take up less space and consume less power.
I don't envision a speed bump to the Retina iPad this year. Rather, I expect the iPad to be revised again next year. It should be on the new, smaller manufacturing process, but I'd expect the new iPad to be ARM A15 based, not A9 based like Apple's current iOS products. The next iPad may have double the CPU power while also getting better battery life.
In fact, with the new GPUs that Apple will be using next year and the new CPUs, I expect the next iPad to be more powerful than the XBO 360 and PS3. That'll be a very interesting day for gaming.
The cpu is such a small consumer of battery compared to the screen that it would take a non-trival sample of iPad2s to say anything statistically significant. Remember, no two batteries are alike.
It make sense to test the new die shrink on a CPU that has been around awhile and on one where disabling broken parts is an ok thing to do. I wonder if any of the shrunk A5's that have both cores working are ending up in shipping products yet (iPhone 4, iPad 2)?
From the article: And, lo! and behold, when we looked at a new iPad 2 (v4), inside it was the APL2498; presumably with both A9 cores enabled this time.
This is reminiscent of the Intel 486 processor and 487 co-processor.
Intel sold 486SX and 486DX parts - the latter had a built-in floating point unit, whereas the former did not. Or at least that's how they were sold. If you had the SX, you could get a 487 "math co-processor" as an add-on to get your floating point unit.
As it turns out, the 486SX, 486DX, and 487 were actually the same chip, with different functionality enabled (although I think the 486SXs were actually ones with defective FP units).
Anyway, if you upgraded your 486SX with the 487, really what happened was that your 486SX got disabled completely, and the 487 became your one and only processor, with a FP unit.
I had one of these (The SX), it was AWFUL! I couldn't even play games that required the math co-processor. I ended up upgrading the processor to one of those Evergreen ones. Worked like a dream. I wish I can remember which game wouldn't run. I believe it was the original Demolition Derby?
Really?
I also had 486SX and the only time I needed the co-processor was when I wanted to run some 3D software (mostly 3D Studio).
And for it to work the only thing I needed was coprocessor emulator.
In case of games it wouldn't probably be a viable solution as it was obviously slower than the real thing.
Quake was, IIRC, the first mass market game which used the floating point unit. 486SX owners just couldn't run it. It also suffered more than most games on 486es and non-Intel chips, since the Pentium FPU was streets ahead of any other x86 chip at the time.
I had a 486DX-50 at the time, the fastest chip ever produced in the series with a 1:1 clock ratio, and it couldn't really run Quake. The FPU wasn't designed to be exercised much, probably intended for light-duty use. In practice it behaved like an "embedded co-processor".
A Pentium 75, perhaps the slowest Pentium ever made, was at least ten times faster because of the floating point gains.
Also used the coprocessor emulator for 3D Studio, nearly had forgot how that went till I read this post. Now I work on two quad core workstations and some stuff needs to go to a render farm in the states. What a flashback. :)
A more direct and recent example would be the Intel Core Solo, which was an identical situation. New chip, high bin rate while working out the kinks, so they disabled dead cores on Core Duos and sold them as Core Solos.
This is pretty common practice to increase yield by using only partially working parts that are still usable in limited way. While 486DX/SX is prime example, there are lots of other:
Original Xbox used marginal DRAM chips and detected by software on each boot which regions of memory are actually usable.
AMD's current CPUs with "weird" number of cores (3, 6, 12..) are produced by disabling cores that does not pass tests.
Cell BE as used in Playstation 3 has one SPU disabled for same reason.
While in most cases such disabling is done by blowing some fuses in hardware I can imagine that in embedded systems software only solution (like the Xbox one) is actually more convenient.
Adittionally, probably the most famous example is the Celeron range.
Celerons were originally Pentium IIs with large segments of the cache fused off, unusable due to defects. However they turned out to be so popular that Intel started to disable the cache of otherwise good cores, and then later began specifically fabbing them, to meet the demand.
For me, the most famous example is DRAM chips in Sinclair ZX Spectrum. Sir Clive got them cheap because they were ones rejected at the production test. Spectrum issue 1 48 K model used 64 kbit chips where one half was broken, and the motherboard was then wired to only use the other half.
That was a time when hardware, particularly memory, was expensive. This was in 1982.
Not exactly. Celerons began as straight up Pentium IIs designed to operate without any L2 cache. They didn't have great performance but were crazy cheap to make and super easy to overclock. Then they started cramming in just a little bit of on-die L2 cache (about 1/4 of the minimum for a regular Pentium II) to help out performance. This turned out to be wildly successful, and eventually this formulation (one chip with on-die L2 cache) became the predominant design for the mainline Intel micro-processors. Afterwards the Celeron family of processors were identical to retail Pentium/Core cpus but with sections of cache disabled, reduced clock rates, etc. (either due to binning by defects or intentional deactivation).
Interestingly, Intel now offers "upgrade cards" for low-cost CPUs whereby components are deactivated via firmware or suchlike and special codes can be entered into the bios screen to turn them on (for a price, of course).
And of course GPU vendors also employ similar techniques. It is not uncommon to see the same chip in multiple models, with lower-end models with parts of the chip disabled either due to defects or for product differentiation.
It's not necessarily a "social good" to provide tools like this. It obviously annoys nerds to no end that companies use software to detune hardware for price segmentation; the nerd worldview is, "what hardware you sell me is mine to do with as I please".
That's utterly sensible. But in the long term, it means that either consumers will have to pay more, or that companies will have to waste time specializing hardware to perform the same segmentation goals. In other words: it's a win for a tiny minority of hardcore nerds, but a loss for everyone else.
Also: consider carefully that similar techniques are used throughout many different industries. For instance: recreational boat engines are detuned in software to provide multiple performance categories at different prices.
After I hacked the Emotiv EPOC headset to get the raw EEG data out of it, lots of people referred to me as a pirate. However, I firmly believe that customers have the right to do with their property what they will, and I ended up writing this in response to all of it: http://daeken.com/the-hardware-hacker-manifesto
As other comments have said, some (possibly most) of these chips likely have a core disabled because of manufacturing issues, so it's probably not a good idea... Apart from power usage (which the AppleTV doesn't have to worry about, being plugged into AC) there really isn't much reason to do it otherwise.
59 comments
[ 2.7 ms ] story [ 106 ms ] threadEither Apple is only utilizing one core or they are binning parts
So, there are possibly two cores, but it's not confirmed.
"Binning parts" means they're either actually hardware-disabling ("diking out") the extra core, or the extra core was bad to begin with (making use of otherwise bad chips).
If those images are real, both cores are definitely on the chip, but whether one is hardware-disabled or not is not certain.
The simplest and most plausible explanation here is that they're using the Apple TV's relatively low performance demands as a way to get some use out of A5s with defective cores
It does seem like the images are real, because apparently the site is in the business of selling very high-res versions of such chips:
http://www.chipworks.com/en/technical-competitive-analysis/r...
"Die photos are available in the Chipworks Report Store at top metal and lower metal/poly"
Sadly, it's one of the few semiconductor segments in which Ottawa, where I live, is still something of a leader.
And, weirdly, the owner of the company used to live in the condo below me.
http://blog.macsales.com/9102-secret-firmware-lets-late-08-m...
Either way, that isn't what is happening here. This is just CPU manufacturing 101. New chip, new processes, high bin rate. You mitigate this as much as possible by disabling features on binned chips to make them sellable. Intel has done it a number of times, typically when making a big change like the introduction of the Core Duo.
It's loud, it's slow, the UI is pretty bad and unpolished. Sometimes it boots to a black screen or locks up randomly. Heck, 1 out of 5 times it doesn't even shutdown (just hangs on a black screen infinitely).
These are all documented issues (and really only the tip of the iceberg) that have gone unfixed for years.
It makes sense to me that Apple is testing out yields for the next iPhone. This new, smaller manufacturing process will allow Apple to put LTE in the next iPhone by having the CPU take up less space and consume less power.
I don't envision a speed bump to the Retina iPad this year. Rather, I expect the iPad to be revised again next year. It should be on the new, smaller manufacturing process, but I'd expect the new iPad to be ARM A15 based, not A9 based like Apple's current iOS products. The next iPad may have double the CPU power while also getting better battery life.
In fact, with the new GPUs that Apple will be using next year and the new CPUs, I expect the next iPad to be more powerful than the XBO 360 and PS3. That'll be a very interesting day for gaming.
Then double that for the old iPad2s.
Intel sold 486SX and 486DX parts - the latter had a built-in floating point unit, whereas the former did not. Or at least that's how they were sold. If you had the SX, you could get a 487 "math co-processor" as an add-on to get your floating point unit.
As it turns out, the 486SX, 486DX, and 487 were actually the same chip, with different functionality enabled (although I think the 486SXs were actually ones with defective FP units).
Anyway, if you upgraded your 486SX with the 487, really what happened was that your 486SX got disabled completely, and the 487 became your one and only processor, with a FP unit.
[EDIT: typo]
http://www.team5150.com/~andrew/carmack/johnc_interview_1997... (search for "float")
A Pentium 75, perhaps the slowest Pentium ever made, was at least ten times faster because of the floating point gains.
Original Xbox used marginal DRAM chips and detected by software on each boot which regions of memory are actually usable.
AMD's current CPUs with "weird" number of cores (3, 6, 12..) are produced by disabling cores that does not pass tests.
Cell BE as used in Playstation 3 has one SPU disabled for same reason.
While in most cases such disabling is done by blowing some fuses in hardware I can imagine that in embedded systems software only solution (like the Xbox one) is actually more convenient.
Celerons were originally Pentium IIs with large segments of the cache fused off, unusable due to defects. However they turned out to be so popular that Intel started to disable the cache of otherwise good cores, and then later began specifically fabbing them, to meet the demand.
That was a time when hardware, particularly memory, was expensive. This was in 1982.
Interestingly, Intel now offers "upgrade cards" for low-cost CPUs whereby components are deactivated via firmware or suchlike and special codes can be entered into the bios screen to turn them on (for a price, of course).
"You're making us lose business by providing a tool to allow customers to use more than what they expected and paid for."
It's not necessarily a "social good" to provide tools like this. It obviously annoys nerds to no end that companies use software to detune hardware for price segmentation; the nerd worldview is, "what hardware you sell me is mine to do with as I please".
That's utterly sensible. But in the long term, it means that either consumers will have to pay more, or that companies will have to waste time specializing hardware to perform the same segmentation goals. In other words: it's a win for a tiny minority of hardcore nerds, but a loss for everyone else.
Also: consider carefully that similar techniques are used throughout many different industries. For instance: recreational boat engines are detuned in software to provide multiple performance categories at different prices.