Next-Gen iPhone & iPhone Nano Speculation

[french toast]

Sorry i should have been more clear..i meant Exynos 5250, as he was using the TI 5xxx as an example about samsung also having A15 ready early..

Saying that though, 2x 544 @300+ is going to be very fast isn't it? how does a 544 compare to a 543? is it just dx9.3 that the difference?

 

[Lazy8s]

Basically, yes. 544 and 543 target similar performance.

544MP2 @ 300+MHz won't be terribly fast by that time, so they'll be targeting a much higher clock. Still, I expect the competition to be capable of pulling ahead of, by at least a small margin, suped-up Series5XT solutions later in 2012 before the hammer drops with Series6.

 

[french toast]

Basically, yes. 544 and 543 target similar performance.

544MP2 @ 300+MHz won't be terribly fast by that time, so they'll be targeting a much higher clock. Still, I expect the competition to be capable of pulling ahead of, by at least a small margin, suped-up Series5XT solutions later in 2012 before the hammer drops with Series6.

Well i think alot of people are going to be suprised with the Mali t604, alot of people were dumbstruck last year when they got the benchmarks of the galaxy s2, no one could believe that arm could of made something that powerfull, even now the mp4 @400 would be the most powerfull out there, the t604 brings 5x the performance, ALL the APIs some of which even the rogue wont get till 2014 (like dx11).

I am itching to get a look at the adreno 3xx series...that it shaping up to be quite good.

 

[Lazy8s]

Read some of the archives on this forum... not everyone was caught by surprise at Mali's (ex-Falanx) proficiency! On several occasions, I had to point out that Exynos 4210 performance would be outside the range of Tegra 2 and existing SGX540 solutions. Mali will definitely be a top contender, but I still don't project the gap closed to PowerVR's TBDR.

 

[french toast]

Oh right, i keep forgetting the peeps on this forum are usualy a step ahead of most others

Well the mali t658 is looking like a beast, i would bet that they are going to be competitive with rogue, t658 will be another 2/4 times the power of t604 so i cant wait for that..

Adreno 2xx series is getting long in the tooth now, it suffered from poor efficiency compared to IMG TECH USSE2. the average drivers and the poor memory bandwidth Qualcomm equiped it with didn't help.
Adreno 225 should show us what 2xx series is capable of, the new drivers put on the 220 in the sammy skyrocket gave a glimpse.

 

[roninja]

If Mali was so nailed on I don't understand Samsung licensing SGX XT cores recently. Did I miss something? Why is Exynos not in all Galaxy SII variants? Is something amiss?

 

[french toast]

If Mali was so nailed on I don't understand Samsung licensing SGX XT cores recently. Did I miss something? Why is Exynos not in all Galaxy SII variants? Is something amiss?

Yea thats what i thought, but someone on here pointed out that they licensed a display ip, not for graphics, so i thing the t604 is nailed on as it supports DX 11 and also covers the 4x that sammy claims for next exynos.

Sammy had to use different SOCs because i think they had supply problems and also becuase to cover 4g in different markets.

 

[Helmore]

Sammy had to use different SOCs because i think they had supply problems and also becuase to cover 4g in different markets.

I thought that mainly had to do with the networks they could support with their Exynos chip. I'm not sure what radio chip they're using along side Exynos though.

 

[JohnH]

Yea thats what i thought, but someone on here pointed out that they licensed a display ip, not for graphics

You're confusing Samsung with someone else http://www.imgtec.com/corporate/newsdetail.asp?newsid=656

 

[french toast]

You're confusing Samsung with someone else http://www.imgtec.com/corporate/newsdetail.asp?newsid=656

Yea John your right (obviously) i got mixed up with Qualcomm.

Maybe there is 2 versions, the 4412 for smartphones and the 5250 for tablets?

 

[Lazy8s]

The T-Mobile US Galaxy S II used a Qualcomm S3 SoC instead of Exynos because they felt using a baseband with support for T-Mobile's 42mbps network was a bigger selling point.

I consistently pull down around 16mbps on it in my market... it's pretty nice. I'd still trade off some connection speed for more graphics, though... but that's me.

 

[Arwin]

If Apple does go quad core CPU + quad core GPU, presumably they are going to have an inventory of partially defective chips. The iPad and iPhone are going to need full chips for competitiveness, but the next iPod Touch could presumably make do with spec reduced, fused off parts. A quad core Apple A6 disabled to dual core CPU and dual core GPU would look much like an Apple A5, but presumably would have battery life benefits from being 32nm, additional clock speed room, and make use of chips that would otherwise be thrown away. I don't suppose a triple core CPU configuration is easy to support? I'm guessing PowerVR doesn't support a MP3 configuration?

Making use of defective parts gives Apple more configuration flexibility while maintaining the 1 SoC/year development cycle. The iPod Touch is probably the lower volume platform compared to the iPad and iPhone which should reduce concerns that they'll need to disable many good chips just to supply the Touch as yields improve.

http://www.eurogamer.net/articles/2012-01-16-ipad-3-march-release-date-evidence-builds

If true that it will dramatically increase resolution (from the looks of things to something like 2000x1500) then they'll need the quad-core performance to keep everything smooth.

 

[Ailuros]

http://www.eurogamer.net/articles/2012-01-16-ipad-3-march-release-date-evidence-builds

If true that it will dramatically increase resolution (from the looks of things to something like 2000x1500) then they'll need the quad-core performance to keep everything smooth.

The fabled 2048*1536 resolution has been floating around for a long time now in the internet. It makes sense since it has a 4:3 aspect ratio just like 1024*768.

 

[silent_guy]

If true that it will dramatically increase resolution (from the looks of things to something like 2000x1500) then they'll need the quad-core performance to keep everything smooth.

Because screen resolution and CPU performance are highly correlated?

It'd be just fine for me to have simply scale up 2x initially. I want the high res text, though. Should be no problem with the current SOC.

 

[Arwin]

The fabled 2048*1536 resolution has been floating around for a long time now in the internet. It makes sense since it has a 4:3 aspect ratio just like 1024*768.

Considering how important the apps are and the simplicity if handling a clear multiplier makes that res very likely, I agree.

 

[wco81]

Sounds like Android tablets will go a different way. Samsung is rumored to go to over 2500x1600, topping what is expected from Apple, while Asus will have a 1080p or 1920x1200 display.

Android apps. are suppose to scale better to different resolutions without a rewrite. But design and layout (and UI flow) may not be optimal if they depend on resolution scaling rather than redesigning for higher resolutions.

 

[Helmore]

http://www.eurogamer.net/articles/2012-01-16-ipad-3-march-release-date-evidence-builds

If true that it will dramatically increase resolution (from the looks of things to something like 2000x1500) then they'll need the quad-core performance to keep everything smooth.

I think that an iPad with 4 times the resolution of the iPad 2 would run fluently on an Apple A5. I'd be perfectly happy if the Apple A6 chip was simply a die shrink and clock speed increase of the Apple A5. That said, I'll never complain about getting more than that.

 

[Ailuros]

Because screen resolution and CPU performance are highly correlated?

No but GPU performance (ie fill-rate) is. Once you scale up GPU performance by N degree it wouldn't necessarily be a bad idea to go also for a stronger CPU config to keep a finer balance between the two. How to achieve either/or is of course up to Apple's engineers.

It'd be just fine for me to have simply scale up 2x initially. I want the high res text, though. Should be no problem with the current SOC.

Either just increasing frequencies with a shrink of the current SoC or increasing core amounts or both shouldn't be a problem. It's not like it'll come with any major architectural changes between A5 and its successing SoC.

I think that an iPad with 4 times the resolution of the iPad 2 would run fluently on an Apple A5. I'd be perfectly happy if the Apple A6 chip was simply a die shrink and clock speed increase of the Apple A5. That said, I'll never complain about getting more than that.

Then (as I mentioned before) there's always the competition and how Apple's future SoC will compare to it. I severely doubt that they're willing to give up their current lead at least in terms of graphics performance.

 

[silent_guy]

No but GPU performance (ie fill-rate) is. Once you scale up GPU performance by N degree it wouldn't necessarily be a bad idea to go also for a stronger CPU config to keep a finer balance between the two. How to achieve either/or is of course up to Apple's engineers.

My pet theory has always been that the A5 was so overbuilt for 1024x768 because it was intended to be used for 2x resolution also (or at least as a fall back...)

That said, I didn't check specifics, such as fill rate, maybe I'm just full of it.

But you don't need fill rate (for games) if you double up the resolution during scan out. There all you'd need is a line buffer.

 

[Entropy]

Because screen resolution and CPU performance are highly correlated?

Depends. How are typefaces/sizes/scaling handled for instance? There's quite a lot going on with the content of the screen that is resolution dependent and handled by the CPU. And it is difficult to see from a user perspective just what would turn out to be limiting if the number of pixels to be handled was quadrupled. It is probably less than totally clear across all apps for Apple as well, so beefing up the CPU resources a fair amount makes sense if it means reducing the risk of having 50 million customers complaining about crummy responsiveness.