Dual core SGX543 dramatically boosts iPad 2 graphics - Page 2

I'd like to see that because all the games I've seen so far use less demanding graphics when running on the iPad 1. That includes RR 2 HD, which has no problem on my iPad 1, but looks great on my iPad 2. That is, now, my wife's iPad 1.

The thing I wish they did with the cover was to have magnets hold it against the back when folded over. It's annoying having it flap back and forth. And, like the old case, it leaves smudge lines on the screen.


6 working days. You pay the fare there, and they pay it back, unless that's changed.

Hey! I though you were going to exercise some restraint while school season was still on.

It's a very fair comparison because one cannot change native display resolutions on tablets. With computers, you can arbitrarily change resolutions, monitor sizes, etc., so a head-to-head comparison at the same resolution would be a better comparison.

Each manufacturer has to weigh the benefits of screen size, resolution, CPU and GPU performance, battery life, etc.

Motorola chose poorly.

But in the end, benchmarks are rather stupid since they don't duplicate real world usage.

If benchmarks and raw performance were so important, why aren't Windows PC users happier than Mac users? The customer satisfaction surveys always have Mac users at the top. And remember, something like half of the Macs being sold in Apple's retail stores are going to first-time Mac buyers. That is, a lot of these happier Mac users are former Windows PC owners.

Apple has too many advantages here. Like Samsung, Nvidia and others, they license the basic chip designs and mod them. But other tablet manufacturers must then buy one of those off the shelf chips for their products. They do have a good choice, as there are several manufacturers making different but pretty good chips.

But Apple can mod the chip to work better with iOS, and it can modify iOS to work better with their chip. Other manufacturers can't do that because they get the OS as is, though they may be able to add their own UI overlay.

This advantage can not only be shown by the speed tests Anand did so far, but if what they found is true, the CPU is only running at 895 MHz, not 1 GHz. That makes the Xoom running at 1GHz an even worse performer relative to the iPad 2. And that makes the iPad2 and it's hardware and software even more impressive.

What great innovations would you like to see in the iPod Touch? I happen to think that it's a perfectly evolved product.

To suggest that Apple would somehow stop innovating because there is no competition belies everything about the company's history.

I'll add to that:

Costs more, runs slower, almost no apps.

Now will people start believing me when I've been saying all along that we are just one or two generations away from the iPad1 being a PS3 killer? I also did mention repeatedly a dual core ARM at 1ghz being essential for this, though of course I mentioned graphics equivalent to what I know from the PC side ie. an equivalent of an AMD 6000 class GPU.

More relevant, clearly iPad 2 was going to have a Retina Display but they were probably not going to be able to make enough of the screens, Steve was not at his best of health, and the iPad 2 is so far ahead of the competition anyway. iPad 3 = Retina Bring It ON.

Finally, I think game makers are going to be making some interesting, very near console-quality titles with these graphics.

The iPad 3 will definitely run what we consider PS3 and Xbox360 AAA "hardcore gamer" titles. There is very little reason for Apple NOT to release a gaming console with the A6.

The Xoom and iPad2 comparison is the only true comparison so far. All the others that have been made have just been from paper specs, which are useless from a performance point. It's also the only tablet that can be fairly compared. Smaller models aren't really competing directly.

If the retina display comes our, and I doubt we'll see it this year, will require new hardware, and newer drivers. With 4 times the pixels, it will require a gpu 4 times faster in order to keep up with the current model. Then, it will need a new CPU, because many apps are CPU bound, even for their graphics performance. We see this when games and graphics cards are tested.

In addition, a number of graphics functions on the iPad2 are only twice as fast, meaning a 4 times as dense display would actually be slower sometimes. This is why, despite what a few people seem to think, a high Rez display can't simply be dropped in. If Apple was able to get those displays cheaply now, and was assured production, possibly, they could have done it, because performance would have been measured against the old, slower model. But not any more.

In most of the EU, the movie rate is 25FPS because of the line frequency, which is 50 Hz. Their Tv has a frame rate of 50FPs because of that.

30FPs is usually considered to be the slowest rate that a game is playable.

But movies, that is film, is doubled. That 24FPs is actually running at 48FPS. What is done is they double every frame, and run it twice. That way, the frame rate is really 48FPS, but with only 24 different frames per second. If you have 24 frames per second, it will look jumpy.

I think we're only just scratching the surface with the iPad 2 graphics. We're going to see some very interesting stuff in the next 8 months.

Big question: Do we need OpenGL3 in iPad 3?

Why stop there?

Costs more, weighs more, runs slower, almost no apps...

After a while, it starts to feel mean...

Thanks! This is exactly what I was wondering. I thought there was a chance with all this talk of graphics firepower that it might indicate that Apple had a roadmap for a drop-in retina display update on the iPad2.
Clearly, if that would put the iPad2HR behind the original iPad in some graphics functions than it will not happen!

Edit: I re-read this and I think it makes me look naive for hoping for a HR update. In my defense, Apple was throwing around 9X graphics performance increases. In my mind, that sounded a little like overkill unless they had some plan for using that capability...

For just rendering the iOS UI and 2D graphics, the iPad 2 would already handle a retina resolution right now. For games, it could have a 2x mode that looks just like the current iPad screens. Games benefit more from smoother frame rates and better textures and models anyway, than they do from higher pixel densities, and 1024x768 on a 10" screen is comparable to typical PC gaming pixel densities.

For the browsing experience and text rendering alone a retina screen would already make sense, and technically that's already possible.

Comparing a 1024 x 768 resolution benchmark with a 1280 x 800 resolution benchmark.

This is like saying "wow! my graphics card runs 120FPS at 640x480 but yours runs it at 34 at 1440x900! My card is superior!"

You will see similar performance from this new chip to the Xoom when the resolution of the iPad 2 is bumped up to 1280 x 800.



In order to get similar or faster performance compared to this new chip, the newer chip inside the iPad 3 will need at least 4 times the graphics power of the current one just to keep up with benchmark scores of the current iPad 2. Therefore, increasing resolution only degrades benchmark scores.

Perhaps, Apple maintained the same resolution of the display scree for this purpose: to show people how much of a performance boost they get.

Apple is the competition.

And as long as it continues to beat itself, it will continue to lead. As they, i.e. Apple has done for so long now.

Your contention that Apple could have had cameras in the iPod Touch right from the start is ludicrous. Could you list the lenses that could fit into it at the time and at the cost that wouldn't raise the price? Was there an mp3 video player/camera that matched in every aspect the first iPod touch. Heck, is there one now?

Keep in mind that 'invention' and 'innovation' have different meanings. Some of the greatest innovators never invented anything. It appears that Apple, i.e., Jobs, innovates first and out of his innovations, comes inventions.

FYI
To become the market leader you must be the`Competition`
http://www.ippmedia.com/frontend/index.php?l=20315

http://en.wikipedia.org/wiki/Innovation

Seeing that the iPad 1 and 2 use OpenGL ES 2 and not full-blown OpenGL, and there is no such thing as OpenGL ES 3, I'd say no, iPad 3 doesn't need OpenGL 3.

GPU hardware is all fully programmable right now, and every GPU vendor can expose specific fixed-function capabilities through extensions, so sticking with OpenGL ES 2 will not limit iPad graphics at all I think.

Yeah right... 30% more pixels == 4 times lower graphics performance... Also, triangle throughput is independent from resolution...

But keepin hanging on to every last bit of justification for your own assertion that the Xoom and Tegra 2 are really as great as all the Motorola and NVidia PR has lead you to believe.

The CPU part in the Xoom is on par with the iPad 2 and other comparable chips. The GPU part is actually a lot less sophisticated which it tries to make up for by throwing more brute force behind it. But the real difference in GPU performance as you see it here is probably software-related, I think the way Android works with it's Java-based frameworks is simply hurting performance of things like graphics.

Let me add to that that from the user perspective it doesn't matter why or how their frame rates are shitty if they are. If Tegra 2 can't drive 1280x800 at decent frame rates in games, while it would magically be so much faster with 30% less pixels, then Motorola screwed up when they picked their GPU-screen combo.

Anyway, it doesn't matter that much anyway, as these benchmarks clearly show Tegra 2 really is not the mobile powerhouse that NVidia has always marketed it as, and that NVidia still has years of catching up to Imagination Technologies when it comes to mobile graphics. A good desktop-grade GPU maker does not imply a good mobile GPU maker.

Honestly, I'm beginning to get confused over all the standards and compatibility questions.

Supposedly, OpenGL 3 is OpenCL compatible. Plus, it adds a number of efficiencies that 2 doesn't have, in addition to having more abilities. We can always use newer standards developers will take advantage of.

I remember reviewers saying that the IBM PC with a 286 could do all that businesses required. Later, it was said that a 1HGz Pentium was as fast as would be needed. There are a lot more statements like that. Software follows hardware. When the hardware can work the math quickly enough, we'll see the software be upgraded. So we should see OpenGL 3 at some point. Do we want it? Sure we do!

Jobs did say "up to".

You are right on ... Just look at the upgrades of Macbook ... prices have been dropping some, but, performance is better, better, better with each iteration. The White Macbook for students is being replaced with Macbook air. Instant on, light weight, longer battery life, the essential legacy ports for EZ data shifting. Macbook Pro with discrete graphics is the most powerful laptop on the planet. OS Lion is bringing revolution to integration within the Mac Apple lines of products. Their wallet is growing, and the revolution continues. iPAD is fabulous for many things people do with their laptops. But many want the features of Macbook Air in iPAD. But try to hold Macbook Air with one hand and read a book, etc vs iPAD. All things in one device loses. That is why my workshop has many tools. Each one for the specific job it needs to do. ShopSmith tried to integrate many great functions into one system. It failed, because of setup time and reset and realign. I'll stop here with analogies.

For some things it might work. But even for browsing, it would be slower. Remember that the gpu would have to render four times the pixels. That still requires more power. I'm noticing on my iPad2 that most things are faster. Browsing is faster, and so is quickly scrolling down a long page. Where before I would always get the checkerboard for a second or two, now I either don't get it at all, or I get it for a fraction of a second. That's due to faster hardware.
Put a four times the pixel display in there, and that will all slow down again.

No. As we've said, the resolution is just 23-25% higher, but the tests have shown performance of the iPad2 to be anywhere from 100% to 400% higher. Changing the screen would make little difference in the speed of the graphics.

For a retina display, it would be different.

All this has been interesting to watch. Before this time last year, the iPad was pretty much the worst kept secret in Silicon Valley. Everyone was expecting Apple to do something, but nobody else, save the ill-launched Joojoo dared try to beat them to market. Android has up until that time been very successful with the "fast follower" approach. But then a few things happened competitors did not expect.

First, Apple leaked probably one of its most memorable bits of disinformation...a "head fake" as you will. They leaked to reputable news outlets that they were thinking of pricing said tablet close to $999. A lot of this made sense...Apple was known for premium priced products and had outright said this was their strategy in general. I'm sure that more than a few CEOs were beginning to make launch plans in the following months with products of their own.

Then Apple dropped the final product....a sleek tablet with a funny name...and then everyone's jaw hit the floor when they said you could have one for $499. For perspective, that's less than an iPhone off contract. The "head fake" literally froze out everyone and their respective supply chains while they scrambled to figure out how to sell a product that cheaply and still make money at it. Weeks turned to months and it took late into the year for even the first credible alternative (the Samsung Galaxy Tab) to even make it to market, and that was a 7" tablet....at the same price of Apple's 10". it didn't help that Google was very late in getting a tablet version of Android out the door when they of all people should have seen this coming.

Now we're into 2011 and most of the Android tablets were finally poised to drop. Motorola, which had literally clawed its way back from oblivion riding the back of Android began previewing the Xoom. It promised to address iPad shortcomings, run faster and provide a better tablet experience. Markets and most pundits were expecting Apple to quickly cede majority share of the new market it created faster than the phone market. Motorola, combined with Google and Nvidia with the Tegra graphics platform took their BEST SHOT.

So what happened. Well...

1. Steve Jobs dropped the iPad 2 shortly thereafter and sucked all the marketing air out of the room. That was going to happen anyway. But...

2. The Xoom still came in too expensive. This after a year of trying to get something price competitive out there. Steve Jobs was almost gleeful hammering this fact home during his keynote. It's a testament to how Apple rules the supply chain these days so that it's hard for even powerhouses like Samsung to keep up...and they provide Apple will many key components!

3. The iPad 2 came in 33% thinner and 15% lighter. Some people poopoohed this but people who know industrial design basically knew that Apple had raised it's proverbial middle finger to anyone who thought they could out-engineer them in terms of hardware. Motorola is a company with decades of experience making small consumer electronics...they invented the Star*tac and the RAZR.

4. The iPad 2's graphics smoked the Nvidia-based Tegra 2. This was a big surprise....not that it was faster but that it was "another league" faster. Again, Nvidia has tons of experience in graphics and mobile graphics. But they still have lots to learn about handheld apparently.

5. Honeycomb wasn't ready. This was another surprise. I would have thought that Honeycomb would have been better baked by launch but it wasn't. You don't want want people like Walt Mossberg saying that the Xoom couldn't get through a review without crashing the whole unit multiple times. Google has no excuse outside of trying to make a launch window. If Apple had launched something that buggy, they would have been raked over the coals by everyone. Geeks and Fandroids might forgive shortcomings like that, but everyone else who buys a tablet won't. If you walk into an Apple store on a typical day, the place isn't packed with fandrois. It's crowded by ordinary folks who buy Apple stuff because it works for them enough that their geek friend doesn't have to explain it to them.

6. Flash wasn't ready. Really, Adobe? It's been, you know, 4 years now.



What's happened is that Apple is learning from it's mistakes that allowed Android to poach market share away from them in the phone space. Apple is moving faster to market and locking up key supply chain products early before anyone can get their hands on them. Everyone knows displays will be next. If they can lock up that space for an extended period of time, then they will own the lion's share of this space for some time to come.

2011 should be fun to watch. We still don't know how Apple acquisitions of outfits like P.A. Semi and others will differentiate the playing field. Apple doesn't talk about this stuff alot, but I'd be curious to find out how different the A5 is from other Cortex-A9 designs.

Just for perspective-

Tegra 2 - 4.8 GFLOPS (8, 1-way ALUs @ ~300MHz)

PowerVR SGX543MP2 - 19.2 GFLOPS (8, 4-way ALUs @ ~300MHz??)

Radeon 9700 Pro - 33.8 GFLOPS (8, 4-way ALUs (pixel) + 4, 5-way ALUs (vertex) @ 325MHz)

Radeon 2400 Pro - 42 GFLOPS (8, 5-way ALUs @ 525 MHz)

Radeon 5450 - 104 GFLOPS (16, 5-way ALUs @ 650MHz)

Xenos (Xbox 360) - 240 GFLOPS (48, 5-way ALUs @ 500MHz)

RSX (PS3) - 255.2 GFLOPS (24, 2 x 4-way ALUs (pixel) + 8, 5-way ALUs (vertex) @ 550MHz)

Radeon 6970 - 2703.4 GFLOPS (384, 4-way ALUs @ 880MHz)

Radeio 6990 - 5100 Gflops


With the quad core version of this same chip (MP4) which will be in the Sony NGP, we'd be getting into the raw GFLOP range of high end cards from 7 years ago.

LOL, looks like a slogan for the Mac segment of PCs.

Like the Mac, right?
oh, we get it...you can use BootCamp.

I agree with most of what you said here, but not with this. There is no indication that Apple leaked anything. There were analysts and tech writers who were expecting an Apple tablet to be expensive for several reasons. One was that it was expensive to manufacture a high quality tablet, as we see. And they knew of plans from other manufacturers, including tentative pricing. Even some new book readers were going to be priced at $650.

And two, it was a fair assumption that an Apple product would be a very good, but premium priced model. Apple didn't have to leak anything for them to think that. I was hoping that they would have the price below $800 for a base model myself, but wasn't certain they would.

Certainly, if Apple maintained the same margins for the tablet they maintain for the IPhone and Touch, the base model would be closer to $600. But Jobs said that they accepted lower margins to price this to sell. That was a surprise. But it strangled everyone else. We could hear the sound of the other shoe dropping in the executive offices around the world of the manufacturers of different tablets and book readers.

Very funny. If it were only true.

Did you just come here to troll. I'm the wrong person to troll. Do it again, and you're gone.

I think you try so hard here that i can see it 1000 miles away. Keep trying ...

It really is a complete roll reversal of the Mac market. Not just from Apple’s end, but from their competitors too.

I think even those who understand what Apple is doing here should be amazed by the 2nd generation product having 18 models for the iPad ranging from $499 to $829 and 10 different accessory covers designed in tandem with the device.

The other vendors can barely muster one model to sell and are either going with half the display real estate (which had certain people saying was ideal as the display on the iPad was too big) or going with a single expensive 10”model (which I think has those same people saying is better than the iPad because it’s a slightly bigger display).*

I am certainly dumbfounded by this roll reversal. The iPad 2 event also had a lot of digs at other tablets yet I don’t see them as a big threat. I can only assume this level of aggression is Apple wanting to create another iPod in terms of market dominance. If it can do than I think the iPad could be Apple’s most profitable business segment within a couple years.


* Xoom has one model with a second one planned for WiFi with no release date, right? Only one Playbook planned, right? And the Galaxy Tab only has one, right? The Asus tablets are the only I can think of that have any diversity, but it’s not capacity diversity, but unique qualities to the OS type and/or casing for specific usage needs. At least the Asus tablets all have IPS panels. Go Asus!

I think the current iPad 2 GPU should easily be able to render at that resolution already. Remember that HTML rendering is not so much about fill-rate as it is about laying out the different elements of the web page, which is mostly CPU bound, and rendering fonts, for which really fast and smart-caching implementations exist already and scales pretty well with higher resolutions (ie: rendering fonts at 4x the pixels takes less than 4x the time). The rest is all compositing and bitblts (GPU copying) of images, flat, non-shaded rectangles and scrolling (which again is mostly bit-blitting. You'd need more RAM for off-screen content to keep scrolling smooth, but GPU blitting is already so insanely fast and cheap to do that the iPad 2 GPU can handle 4x the pixels. At least that's what I think.

The checkerboard you see in the iOS browser is mostly because the device can't keep up loading the page over the network and calculating the layout, I don't think it's GPU bound.

If only everyone could "get this" .... they would truly understand why Apple is "the best" and others are just "the rest".

I believe there a second Playbook planned. But with the availability of the first one being just somewhere in the spring, when the second will get here is a good question.

Apple has a lock with a number of manufacturers. Being able to pay a lot in advance helps too. The large numbers get better pricing, and also allows the R&D cost to be spread around. They get benefits all over. Shipping costs per device are lower because of volume.

In addition, Apple has an advantage that other manufacturers don't have. Apple sells a lot of devices through their own web site. They also sell a lot through their own stores. This allows them a higher margin on those sales. It then allows them to lower the price and average out that margin with other third party retail sales. That gives them the ability to lower the price again. No one else can do that. I'll bet that Moto will sell just a small percentage through their own website, as I do assume they can do that. Same thing for Samsung.

A great deal of what you're saying is true, but not entirely. There are still pixel based graphics and images, as well as textures that must be rendered. When I'm quickly scrolling down a long page, that page is already downloaded. If it were network issues, then there would be little difference between the two models, but the difference is great. I find the same thing to be happening with offline apps, so there is no network issue there.

At any rate, even if some things weren't affected much, the things that would be, would be noticed. And as games are very important on the iPad, any slowdown wouldn't be acceptable.

This is 100% right (Google being the old Microsoft in the sense that all it is doing is copying everybody else in all areas outside of its search engine business). If Google's search results weren't so good (the best) I would use something else (Bing sucks).

It amazes me how people deride the iPad for various reasons but don't acknowledge just how influential that product was in the tablet market. If the iPad was not that revolutionary, then (1) how come no one released anything remotely like it before it was introduced and (2) how come everyone is now essentially copying it (most of the hardware form factors are almost exact replicas of the ipad).

What was in the market before the iPad and how well were those devices selling?

Don't forget the "ES" part of Open GL ES 1/2/3/4/whatever...!!!

But, like I said, we'll see the gaming gurus push this iPad 2 hardware to the max. We'll definitely be seeing some interesting stuff in the next 6 months. I'm surprised Steve didn't feature a top-tier game developer in his keynote as usual. But I'm sure things were quite tight up to that point.

Now why the iPhone 5 would need the GPU horsepower of the iPad 2 is quite beyond me. But then, I've never been a big phone or apps guy. Ah, whatevs. I'm running around in circles now trying to digest all the info.

BTW the 320M is brilliant compared to the 9400M, in case anyone is wondering, in my personal testing (picked up a price-dropped MBP 13" 2010 model a few weeks ago). Twice as good, at least, in Valve Source-driven games. So if you're into any kind of PC or Mac gaming, the new MacBook Pros (13") with Intel graphics that are sub-320M performance (plus other issues) is NOT YOUR OPTION.

But if you're talking raw CPU power then the new MBPs can't be beat.

The best MBP for those that actually would harness that power would be IMO the
$3,649.00
2.2GHz Quad-core Intel Core i7
8GB 1333MHz DDR3 SDRAM - 2x4GB
512GB Solid State Drive
SuperDrive 8x (DVD±R DL/DVD±RW/CD-RW)
MacBook Pro 15-inch Hi-Res Antiglare Widescreen Display

But... we are now in a "post-PC" era

You guys deal in the rumor business on a day-by-day basis but I'm going by my memory at the time that the Wall Street Journal was the one who went with that rumor at the time. The WSJ isn't known for publishing rumors unless they come from high end sources at companies. In Apple's case, they have been known to get tips straight from the top before when Apple wants them to know something. I still might be wrong, but how do you remember it?

The customer will pay more... for less... of worse...
-The Palo Alto Postulate-

...because worse is undesirable, the customer wants less of it, and will pay more for what he wants!
-The California Ave. Corollary

We'll always want faster stuff. It's in our genes. Maybe a hundred years from now they will have reached some sort of practical limit, but maybe not. Maybe it will be in two hundred years. At some point a limit will be reached, where going faster isn't possible, or will serve no purpose for personal devices. But until then, improvements come fast, and we can use the results.