Apple A5X SoC

[alent1234]

Apple ought to release a version of Xcode for iOS. That would be a useful tool for a lot of professionals and it would also lower the start up cost for new developers since they could only need to buy an iPad to start making apps vs needing a $1000+ Mac as well.

The Android SDK was already more accessible since it's cross platform but there is also a version out now that can run on an Android tablet directly with no need for a pc.

just like with professional apps, there is no way a decent version of xcode will work on iOS. the hardware is gimped compared to real computers and the OS is a lite version

there is no way you will be able to work with really large files

 

[dagamer34]

Uh, read the highest rated comment.

The general idea is that... the system reserves a certain amount of RAM for its kernel and other things, so obviously, you are right in that not all 1GB would be useable.

But it's not 512MB max. There is no limit. The limit is set because the GPU also needs to reserve a certain amount of system RAM for it to use as VRAM. When an application requests more memory, and none is available because the GPU is taking it up, the OS will try to close down other apps before it returns a NULL for memory requests.

If the app does not have enough memory to operate (NULL returned for more memory), it'll simply crash, unless the developer handles it somehow.

Low memory warning, on the other hand, are sent to background apps, not foreground (currently running) apps. I think you are getting mixed up there. Low memory warnings are sent to background apps so that they know to free up memory. If they can't free up memory, they'll have to close down. They are given a small window to comply to the free memory request. If they can't, the OS will force them to close.



It's actually simpler than that. They don't want to bring professional apps to iOS because the general pricing has gone down to $9.99 or around $19.99 to below now. If they suddenly come out with a $599.99 suite of professional apps, nobody would buy them, and so... it doesn't justify the amount of resources and time spent to develop something like that.

But if nobody starts churning those out, then we are left with $9.99 apps that are almost professional, but still not useful enough.

Considering iOS has had the low memory warning function since iPhone OS 2.0, 2 years before multi-tasking was introduced in iOS 4.0, it applies to foreground apps as well, particularly those that cache a lot of data that can be recreated.

Source: https://developer.apple.com/library...nDelegate/applicationDidReceiveMemoryWarning:

Now, I'm not sure why Apple would bother exposing that function in iPhone OS 2.0 if it only applied to background apps, which did not exist at that time. However it makes plenty of sense when your app is trying to use too much memory and signal to it to dump stuff before it gets terminated.

And I find it silly that you say "there is no limit". Of course there's a limit. With no virtual memory, the OS will eventually say "no more" and quit you if you don't comply. No one knows where that limit is (I only merely guessed at 512MB on a 1GB device, it could be higher or lower) but it varies from one device to the next, but it exists. That does NOT happen with consoles, that have a clear boundary on what they can use (which in the past used to be 100%, but with background downloading in modern consoles, is probably 95-99% now).

 

[runawayprisoner]

Shouldn't we have one "official 3rd Gen iPad thread" the same way we (try to) do with most other devices on this forum?

I don't think so. If it must be made, it should be in "All Things Apple".

Considering iOS has had the low memory warning function since iPhone OS 2.0, 2 years before multi-tasking was introduced in iOS 4.0, it applies to foreground apps as well, particularly those that cache a lot of data that can be recreated.

Source: https://developer.apple.com/library...nDelegate/applicationDidReceiveMemoryWarning:

Now, I'm not sure why Apple would bother exposing that function in iPhone OS 2.0 if it only applied to background apps, which did not exist at that time. However it makes plenty of sense when your app is trying to use too much memory and signal to it to dump stuff before it gets terminated.

And I find it silly that you say "there is no limit". Of course there's a limit. With no virtual memory, the OS will eventually say "no more" and quit you if you don't comply. No one knows where that limit is (I only merely guessed at 512MB on a 1GB device, it could be higher or lower) but it varies from one device to the next, but it exists. That does NOT happen with consoles, that have a clear boundary on what they can use (which in the past used to be 100%, but with background downloading in modern consoles, is probably 95-99% now).

Thanks for the memory jot. Looks like I need to review my information.

However, whether it happens with consoles or not is quite debatable in my opinions. Consoles with a firmware like the PS3 or 360, for instance, do use some memory for their firmwares.

http://www.dailytech.com/Sony+Shrinks+PS3+OS+Gives+Devs+70MB+of+Extra+Memory/article17775.htm

As far as virtual memory goes, it actually does exist in iOS.

https://developer.apple.com/library...tual/managingmemory/articles/aboutmemory.html

 

[Mopetar]

Shouldn't we have one "official 3rd Gen iPad thread" the same way we (try to) do with most other devices on this forum?

Knock yourself out and make one then. Until it actually launches and gets into the hands of users and reviewers most of the things that are posted will still be baseless speculation, much like this thread. No one has any clue what the A5X is like unless they've actually used the hardware.

I don't think so. If it must be made, it should be in "All Things Apple".

Why? It's an Apple product, but it's clearly a mobile device and/or gadget.

That said, I'm not even sure a single thread is a good idea. Eventually it becomes entirely too large and no one really bothers keeping track of what's been discussed so it inevitably runs in circles. It also makes it difficult to find any one particular piece of information that someone might find useful, which also contributes to the run-around-in-circles problem.

 

[finbarqs]

let's not forget that the VITA also is a Quad Core CPU Cortex A9...

 

[ChAoTiCpInOy]

Good post about A5X...

http://www.anandtech.com/show/5663/analysis-of-the-new-apple-ipad/

 

[zsdersw]

What I want to see is how the A5 performs on a screen with as many pixels as those that are used to benchmark Tegra 3.

I'd hope the A5 could win handily against the Tegra 3 @ the 1024x768 pixels of the iPad 2.

Seems to me to be pretty pointless to show how the iPad 2 performs against Tegra 3 tablets when the pixel count is nowhere near the same.

 

[MrX8503]

What I want to see is how the A5 performs on a screen with as many pixels as those that are used to benchmark Tegra 3.

I'd hope the A5 could win handily against the Tegra 3 @ the 1024x768 pixels of the iPad 2.

Seems to me to be pretty pointless to show how the iPad 2 performs against Tegra 3 tablets when the pixel count is nowhere near the same.

Anand predicts that its 3-4x the performance of Tegra 3 at 1024. Although he's doubtful its really 4x and is only theoretical.

At the iPad3's native res, it performs a little slower than the current iPad2....per Anand's preview. However he said the performance is "generational" and has a similar advantage on the competition as what the A5 had a year ago. Simply, any future tablet that wants to go retina style needs a powerful GPU.

 

[zsdersw]

Anand predicts that its 3-4x the performance of Tegra 3 at 1024. Although he's doubtful its really 4x and is only theoretical.

At the iPad3's native res, it performs a little slower than the current iPad2....per Anand's preview. However he said the performance is "generational" and has a similar advantage on the competition as what the A5 had a year ago. Simply, any future tablet that wants to go retina style needs a powerful GPU.

Yes, I read that.

What's missing from Apple's graph and claims about being 4x faster than Tegra 3 is the phrase: because it needs to be.

 

[MrX8503]

Its missing from their graph, but in their promo videos they said "because we have 4x the pixels we need quad core graphics". Or something along those lines.

Whatever the case its a higher resolution display and a faster gpu than any competitor right now. We may see a repeat of what we saw in the A5 in terms of performance lead.

 

[zsdersw]

Its missing from their graph, but in their promo videos they said "because we have 4x the pixels we need quad core graphics". Or something along those lines.

What videos?

 

[MrX8503]

I think its in the video where all the engineers/designers are sitting in a white background and talk about how amazing it is. On apple.com home page.

 

[zsdersw]

I think its in the video where all the engineers/designers are sitting in a white background and talk about how amazing it is. On apple.com home page.

Yes, that's in there... but it's not exactly as omnipresent in their marketing as it should be.

 

[MrX8503]

Even if they were, you'd find something else to nitpick at anyways.


Warning for a non-technical personal attack post. This is a technical forum, please discuss technical details and not each other.

Moderator PM

 

[zsdersw]

Even if they were, you'd find something else to nitpick at anyways.

And you'd find something else to trumpet.


Warning for a non-technical personal attack post. This is a technical forum, please discuss technical details and not each other.

Moderator PM

 

[MrX8503]

And you'd find something else to trumpet.

I don't like everything Apple makes. I skipped the iPad1/2 because I didn't think they were good enough, but I think the iPad3 is a great tablet. You have your red marker ready in hand before anything is announced.

 

[smartpatrol]

Yes, that's in there... but it's not exactly as omnipresent in their marketing as it should be.

Huh? What should they say? "DISCLAIMER: higher resolution screens need more powerful GPUs."?

I guess the anti-Apple crowd are really grasping at straws by this point.

 

[Pliablemoose]

Huh? What should they say? "DISCLAIMER: higher resolution screens need more powerful GPUs."?

I guess the anti-Apple crowd are really grasping at straws by this point.

LOL, I know, it's awesome to watch them squirm and bitch.

I'll bench one of mine next week, hell, I might keep both of them

 

[zsdersw]

I don't like everything Apple makes. I skipped the iPad1/2 because I didn't think they were good enough, but I think the iPad3 is a great tablet. You have your red marker ready in hand before anything is announced.

Doesn't look that way when reading your posts. You always defend/credit Apple and rarely, if ever, criticize anything they make or do.

 

[zsdersw]

Huh? What should they say? "DISCLAIMER: higher resolution screens need more powerful GPUs."?

I guess the anti-Apple crowd are really grasping at straws by this point.

They shouldn't make the "4x the performance of Tegra 3" claim.

 

[runawayprisoner]

Erm... I think it's best to just talk about the SoC here rather than about what Apple does or what they should say, no?

----

Good post about A5X...

http://www.anandtech.com/show/5663/analysis-of-the-new-apple-ipad/

That's an interesting read indeed.

What I want to see is how the A5 performs on a screen with as many pixels as those that are used to benchmark Tegra 3.

I'd hope the A5 could win handily against the Tegra 3 @ the 1024x768 pixels of the iPad 2.

Seems to me to be pretty pointless to show how the iPad 2 performs against Tegra 3 tablets when the pixel count is nowhere near the same.

The way I see it, since Anand's benchmark is off-screen (resolution independent), it gives a good indication of how A5 stacks up against Tegra 3 in a similar resolution situation. It's pretty much independent of the resolution of the tablet, after all:

To be fair, it's obvious it's not a 2x gap in every situation, so the gap between Tegra 3 and A5X is likely not at all 4x.

But what I'm more interested in, is how much faster A5X is compared to Tegra 3 when both have to handle 2048 x 1536, and I'm also interested in how much performance A5X loses when it has to handle such a monstrous res.

 

[zsdersw]

But what I'm more interested in, is how much faster A5X is compared to Tegra 3 when both have to handle 2048 x 1536, and I'm also interested in how much performance A5X loses when it has to handle such a monstrous res.

That's why I think Anand's benchmarks at this point are useless.

It would be as useless as comparing the horsepower/torque ratings of a supercar on paper as a judge of how the car performs, when what counts is how much power it gets to the wheels, whether or not it's AWD, etc.

 

[runawayprisoner]

That's why I think Anand's benchmarks at this point are useless.

It would be as useless as comparing the horsepower/torque ratings of a supercar on paper as a judge of how the car performs, when what counts is how much power it gets to the wheels, whether or not it's AWD, etc.

His benchmark for A5 vs Tegra 3 is still valid and is a good indication of the performance difference between the two.

But as I noted at the very beginning, I sincerely have a hard time believing A5 is 2x faster than Tegra 3, and Anand also agrees. So A5X can't be 4x faster than Tegra 3, or at least not on average.

But the way I look at it, since A5 is about on par or faster than Tegra 3 in some situations, it's safe to assume that A5X should be faster than Tegra 3 by a sizable margin. Maybe not 4x, but I don't doubt it'll be about 2x to 3x.

 

[zsdersw]

His benchmark for A5 vs Tegra 3 is still valid and is a good indication of the performance difference between the two.

But as I noted at the very beginning, I sincerely have a hard time believing A5 is 2x faster than Tegra 3, and Anand also agrees. So A5X can't be 4x faster than Tegra 3, or at least not on average.

But the way I look at it, since A5 is about on par or faster than Tegra 3 in some situations, it's safe to assume that A5X should be faster than Tegra 3 by a sizable margin. Maybe not 4x, but I don't doubt it'll be about 2x to 3x.

The buck doesn't stop at a useless "on paper" comparison. Where the rubber hits the road is where it counts.

 

[lothar]

I don't get the point of doing "Offscreen" benchmarks...
If "Offscreen" benchmarks are the holy grail, then why don't we start using them in benching CPUs/GPUs instead of doing numerous resolution changes, numerous AA/AF/HDR changes, numerous game platform changes, and so on?

It seems to me that "Offscreen" benchmarks benches theoretical performance, not actual performance.

It's safe to say that Apple just took a single benchmark that had 4x performance out of the dozens of other benches they tested.
Nvidia does that a lot on the PC discrete graphics side too. In fact, they've been doing that for decades. It's an art for them.