Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

trims

macrumors regular
May 11, 2011
232
79
Nottingham, UK
I upgraded 4S to 5 and having used the new one now several weeks, I am not sure was the upgrade worth of the cost. 4S is a terrific phone and 5 don't add much to that. Slightly bigger screen and a new design. But on the other hand, both models have nice design. Perform wice I think 4S has all the power I needed.

Here in the UK it may be worth upgrading for the LTE network speeds, but coverage is patchy and the contracts eye-watering.
 

somethingelsefl

macrumors 6502
Dec 22, 2008
461
204
Tampa, FL
Considering this is one of the fastest Apple products to be discontinued...they should offer a $100 trade-in upgrade program for the 4th gen iPad or SOMETHING. It would create goodwill among the hardcore base, and then they could take customers' iPad 3's which are in perfectly good condition and slap them on the refurb store. It would such easy money lining Apple's pockets, I don't know why they don't do this.

I agree. Apple did it in the past when they announced they started shifting around the iPhone product cycle; they offered the iPhone credit program that was proposed and implemented by SJ as a way to show that 'Apple understands customer loyalty' as a way to offset a new product cycle. This really should have been no different.
 

Yvan256

macrumors 603
Jul 5, 2004
5,081
998
Canada
Comparison?

Anyone know how many gflops the PowerVR MBX Lite of the 1st generation iPod touch can do, just so I can cry a bit more? ;)
 

dashiel

macrumors 6502a
Nov 12, 2003
876
0
I agree. Apple did it in the past when they announced they started shifting around the iPhone product cycle; they offered the iPhone credit program that was proposed and implemented by SJ as a way to show that 'Apple understands customer loyalty' as a way to offset a new product cycle. This really should have been no different.

This is nothing like the iPhone price drop and subsequent credit.
 

commander.data

macrumors 65816
Nov 10, 2006
1,057
183
Now that I think of it, going from the SGX543 to the SGX554 there has to be an inherent architecture change which can also help in performance. However, the inherent argument of mine still stands, if the SGX554 (assuming no architecture change) is just double everything of the SGX543, why not use a single SGX554 in the iPhone 5 (assuming you want same performance but less space consumption hence lower TDP)?
The SGX554MP is not double everything of the SGX543MP though. In terms of ALU:TMU:ROP ratio I believe it's something like 8:2:2 vs 4:2:2. So an SGX554MP2 vs a SGX543MP3 works out as 16:4:4 vs 12:6:6. Differences in clock speed will changes things, but assuming equal clock speeds, perhaps Apple wanted to emphasize texture and fill rate for the iPhone 5 vs shader performance.

Could have, but the risk factor is not the issue. There was a risk factor with the A4, A5 and A6X chips since they are all new tech chips. The only safe bets have been the A5X and A6 (half a safe bet).
The A4 was a very low risk chip since it uses the same Cortex A8 + SGX535 architecture as in the iPhone 3GS and 3rd gen Touch and the same 45nm process first used in the 3rd gen Touch. The A5 used the same 45nm process as the A4 but changed the CPU and GPU architectures to Cortex A9 and SGX543MP2 so was a higher risk part. The A5X was also a relatively safe chip since the CPU is identical, the 45nm process is the same, and the GPU architecture is also the same architecture just doubled. The A6 was also relatively safe as it uses the proven 32nm process from the 32nm A5 and the same SGX543MP architecture just changing the CPU architecture. The issue with the A6X is that because it is introduced a little over a month after the A6, the A6X was developed directly in parallel with the A6 rather than being based off a proven design. As such when compared to the A5X, everything has changed from a 45nm process to a 32 process, Cortex A9 to Swift, SGX543MP to SGX554MP. That makes it risky.

I for one would have just used an SGX554 MP2 and give extra graphics processing and used the extra space in the A6 (since one core would have been missing) to lower the chip footprint. Perhaps even add a bit more L1 & L2 caches giving a enhanced CPU capacity. More RAM (1536 MB would have been nice; not that I'm complaining ay 1024MB) could have been added.
http://www.chipworks.com/blog/recentteardowns/2012/11/01/inside-the-apple-ipad-4-a6x-to-be-revealed/

There is no extra space saved by using the a SGX554MP2 vs a SGX543MP3. In fact you lose space. The SGX554MP is 8.7mm2 per core while the SGX543MP is 5.4mm2 per core. So a SGX554MP2 takes up 17.4mm2 whie a SGX543MP3 takes up 16.2.

And RAM capacity has little relationship to the SoC die area. In the iPhone the RAM chip(s) is(are) independently stacked on top of the SoC. You can put any size RAM you want. The limitation is RAM technology. I believe the iPhone only uses 1 RAM chip in it's stack and the largest readily available capacity in a single RAM chip is 1GB hence that's what the iPhone 5 uses.
 

truettray

macrumors 6502
Sep 7, 2012
386
268
USA
I've owned an iPad 3rd gen since launch and think this is awesome news.

I like what they've done to update the iPad, and it doesn't mean mine will suddenly stop working.

Agreed!! My 3rd Gen iPad was an excellent upgrade, and the 5th gen will be excellent as well, especially seeing how good the 4th gen is.
 

bawbac

macrumors 65816
Mar 2, 2012
1,232
48
Seattle, WA
Considering this is one of the fastest Apple products to be discontinued...they should offer a $100 trade-in upgrade program for the 4th gen iPad or SOMETHING. It would create goodwill among the hardcore base, and then they could take customers' iPad 3's which are in perfectly good condition and slap them on the refurb store. It would such easy money lining Apple's pockets, I don't know why they don't do this.
18475_steve-jobs-apple.jpg
 

teknikal90

macrumors 68040
Jan 28, 2008
3,346
1,901
Vancouver, BC
Last edited by a moderator:

teknikal90

macrumors 68040
Jan 28, 2008
3,346
1,901
Vancouver, BC
In CF-Bench, the Nexus 10 managed a total score of 10,814 (17,377 native; 6,439 Java), while finally we looked at browser JavaScript performance, with SunSpider coming back with 1,419.9ms (faster is better). It’s worth noting that the 4th-gen iPad with Retina display scored 879.2ms in SunSpider.

Almost 40% faster...Quite a far cry from an overpriced garbage. You pay for quality. In this case, $100.

also, you quotes the Verge's article above to confirm the battery life on the nexus... but read this on that same article:
Yes, the Nexus 10 wins on price — $399 for this incredible display is a nice deal. But consider the extra $100 you'll spend to get the iPad an entry fee to the App Store, and its many apps and accessories that just aren't available to the Nexus 10.

Google's now proven conclusively that it can design great Android hardware, but until developers prove they can design great Android software it's still hard to recommend the Nexus 10 over an iPad.

They've been playing with it and they cant recommend it.... you dont have it and youre fighting for it lol
 
Last edited by a moderator:

jav6454

macrumors Core
Nov 14, 2007
22,303
6,257
1 Geostationary Tower Plaza
The SGX554MP is not double everything of the SGX543MP though. In terms of ALU:TMU:ROP ratio I believe it's something like 8:2:2 vs 4:2:2. So an SGX554MP2 vs a SGX543MP3 works out as 16:4:4 vs 12:6:6. Differences in clock speed will changes things, but assuming equal clock speeds, perhaps Apple wanted to emphasize texture and fill rate for the iPhone 5 vs shader performance.

Makes sense.


The A4 was a very low risk chip since it uses the same Cortex A8 + SGX535 architecture as in the iPhone 3GS and 3rd gen Touch and the same 45nm process first used in the 3rd gen Touch. The A5 used the same 45nm process as the A4 but changed the CPU and GPU architectures to Cortex A9 and SGX543MP2 so was a higher risk part. The A5X was also a relatively safe chip since the CPU is identical, the 45nm process is the same, and the GPU architecture is also the same architecture just doubled. The A6 was also relatively safe as it uses the proven 32nm process from the 32nm A5 and the same SGX543MP architecture just changing the CPU architecture. The issue with the A6X is that because it is introduced a little over a month after the A6, the A6X was developed directly in parallel with the A6 rather than being based off a proven design. As such when compared to the A5X, everything has changed from a 45nm process to a 32 process, Cortex A9 to Swift, SGX543MP to SGX554MP. That makes it risky.

Process cycles have no saying into the working of a SoC. A process is simply the way the chip is made. True in the end it dictates the potential clock speeds and thermals, but in terms of chip design no. 45nm or 32nm makes no difference during design. The true face of fab process is seen once the chip is made. If its a small chip, a smaller process will drastically decrease power usage and give it low thermals. Also, less real estate; which is good.

As per the A4, yes, you are right, it is the old Cortex A8 architecture found in the 3GS. I forgot about that. Now that I remember, the A4 is just a bundled CPU+GPU. Saves spaces and chip count in the end.

As per the A5, thats entirely new. Better GPU, better CPU architecture. Hell, even dual core GPU integration even though there was no need (at the time since the SGX535 handled everything nicely in the 3GS).

The A6 is still a safe bet like you said (not because of the fab), but because it held to the same GPU but added a core (not as hard) and just integrated hand-made CPU.

Ah yes, I don't know why I didn't see the A6 and A6X parallel development. That can certainly explain why the differences.

http://www.chipworks.com/blog/recentteardowns/2012/11/01/inside-the-apple-ipad-4-a6x-to-be-revealed/

There is no extra space saved by using the a SGX554MP2 vs a SGX543MP3. In fact you lose space. The SGX554MP is 8.7mm2 per core while the SGX543MP is 5.4mm2 per core. So a SGX554MP2 takes up 17.4mm2 whie a SGX543MP3 takes up 16.2.

And RAM capacity has little relationship to the SoC die area. In the iPhone the RAM chip(s) is(are) independently stacked on top of the SoC. You can put any size RAM you want. The limitation is RAM technology. I believe the iPhone only uses 1 RAM chip in it's stack and the largest readily available capacity in a single RAM chip is 1GB hence that's what the iPhone 5 uses.


I know RAM is separate from the SoC die area, I just figured that with a smaller SoC, maybe they could push in another RAM chip. Anyways, it yes, I checked, the SGX554 is bigger, I didn't know that. Well, perhaps when we go to 28nm we can get a smaller die area.
 

truettray

macrumors 6502
Sep 7, 2012
386
268
USA
Anyone know how many gflops the PowerVR MBX Lite of the 1st generation iPod touch can do, just so I can cry a bit more? ;)

Such a good looking device. 1st gen iPhone and iPod touch will both always hold a special place in my heart. lol
 

Anti-Lucifer

macrumors 6502a
Mar 9, 2012
776
2
Where is the microsoft surface tablet in the comparison chart?? :confused::confused::confused:


I can't wait to see how the ipad 5 will annihilate the gSheeps even further.:D
 

Tigger92

macrumors member
Sep 22, 2009
77
0
North Carolina
This makes me glad I did not trade up my iPad2 for the 3. Now I have to decide if I want to wait until the redesigned iPad Gen5 to see what new things they will add. ;)
 

Ubele

macrumors 6502a
Mar 20, 2008
888
332
Remember when the first G4 Mac to break a GFLOP came out - the commercial with the tanks surrounding it because it was "so powerful, the US Government classified it as a weapon." Now our iPads' graphics chips can do 76.8 GFLOPS.

Yes, to kill an enemy combatant with the G4 Power Mac, you had to pick it up, lift it over your head, and bring it down hard on the enemy's head. It was neither quick nor easy. With the iPad, all you have to do is ram one of the sharp edges – all 76.8 GFLOPS of it – into the enemy's throat. How far Apple has come with its weapons technology! (Or so my covert black-ops friends tell me.)
 

whitesand

macrumors regular
Oct 10, 2011
196
17
Apple enjoying a very large performance advantage in the benchmarks over the android nexus 10....How ya like them apples fandroids!:apple:
 

whitesand

macrumors regular
Oct 10, 2011
196
17
I think the fellow is extremely upset by the latest benchmark results favoring apple by a HUGE margin...Even the iphone 5 is beating the nexus 10 in several benchmarks...:apple:
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.