iPad GPU Clock Speeds

Article: Lessons in Technology and Innovation from the iPad 3 Graphics and Display
By: Doug S (foo.delete@this.bar.bar), January 7, 2013 10:17 am
Room: Moderated Discussions
Brett (ggtgp.delete@this.yahoo.com) on January 6, 2013 11:48 pm wrote:
> ltcommander.data (ltcommander.tuvok.delete@this.gmail.com) on January 5, 2013 12:32 pm wrote:
> > I've also included FLOPS/pixel calculations for the iPhones and iPod Touches for interest.
> > The iPad 3's FLOPS/pixel rating actually ends up being close to 3rd gen devices. I used
> > a 325MHz clock speed for the SGX543MP3 in the A6 using a 1/4 multiplier ratio.
> >
> > SoC GPU Width Height Pixels FPS SIMDs MADS/SIMD Total MADs Clock Speed (MHz) GFLOPS FLOPS/pixel
> > iPad 1 A4 SGX535 1024 768 786432 60 2 2 4 250 2.0 42
> > iPad 2 A5 SGX543MP2 1024 768 786432 60 8 4 32 250 16.0 339
> > iPad 3 A5X SGX543MP4 2048 1536 3145728 60 16 4 64 250 32.0 170
> > iPad 4 A6X SGX554MP4 2048 1536 3145728 60 32 4 128 280 71.7 380
>

> The premise behind this article is that the new CPU/GPU was late and did not make it into the iPad3,
> which is wrong as anyone who followed iPad2 rumors would know. The new retina display was late and did
> not make it into the iPad2, which is what caused that huge 8X leap in iPad2 GPU performance over the
> iPad1. The GPU in the iPad2 is massive overkill and a waste of margins, but the chips were already ordered
> and in the pipeline when the retina display that was meant to go with that chip was canceled.


What you say makes sense from looking at the table, as it would have meant a nice orderly progression of 2x more flops/pixel with each model. But just because rumors said it, doesn't mean it was planned for the iPad 2. Moore Law happens on a schedule dictated mostly by the calendar, while feature introductions take place on a schedule dictated mostly by a war between marketing and accounting.

After all, after the iPhone 4 was released with a Retina screen it was not exactly rocket science to guess that the same would be done to the iPad, so the rumor mill started up. As always, every feature anyone can think of is claimed for the next generation product, then Apple delivers only a subset. That doesn't mean Apple was actually planning to do so, only that the rumor writers either believed it would happen then or didn't want to wrong if it did.

There are three categories of predictions for future Apple products. One is the dead obvious, and a Retina iPad fell into this category. Like shipping an iPhone with LTE, it was not a matter of if, only a matter of when. The second category are the ones that are not an absolute certainty but make sense based on typical Apple product directions. It's easy to guess that a future product will be thinner and lighter, Apple will always do that if it can, especially with a bulky product like the iPad 3/4, so any rumors guessing that are about on par with rumors that the next one will improve performance. The third category are rumors about things that maybe make sense based on what the competition is doing, but ignore that Apple does not necessarily follow the market. Examples here would be a bigger screen on the iPhone and a smaller iPad (they did those) or adding a mini HDMI port or SD card slot (nope, never will) or NFC (not yet, maybe later)

It is only when someone can predict something that few were expecting well in advance of product release, like the Retina screen on the iPhone 4, that they either really do have a good rumor source, or make better educated guesses than 99% of the rumor mill. If someone predicted something fairly off the wall, like for instance that the next iPhone having a sensor chip in it that can "smell" (which would be useful for example for having your phone tell you a BAC estimate, tell you how bad your breath smells or even diagnose cancer - don't laugh, some dogs can apparently smell cancer) If that came true they either really have some good sourcing/research or made a wild guess that turned out really lucky (if the next iPhone has it, put me down in the second category)

Now that Samsung Galaxy phones are so popular people are playing this game with them. They see Samsung talk about a flexible display and immediately the rumor mill cranks up about the GS4 having a flexible screen. Had it been LG talking about it no doubt the rumors would have been that the next iPhone will have a flexible screen. Nevermind that electronics don't bend and a plastic screen that flexes may not break but inherent in the chemical composition of all plastics is the fact they'll scratch much more easily than glass (cue the comeback of the screen protector if this rumor is true) There are useful things you could do with a more flexible screen with haptic feedback or even shaping "buttons" out of the screen, but none of the rumor writers go there, despite that being the only reason I can see you'd want to go back to the bad old days of non-glass phone screens. There was a reason phone makers all abandoned polycarbonate screens coatings to "protect" the LCD glass, which the "ooh shiny, flexible screens won't shatter!" crowd don't seem to understand.
< Previous Post in ThreadNext Post in Thread >
TopicPosted ByDate
Lessons in Technology and Innovation from the iPad 3 Graphics and DisplayDavid Kanter2013/01/01 04:14 AM
  Games at 2048 x 1536? Can you tell?Mark Roulo2013/01/01 02:22 PM
    Games at 2048 x 1536? Can you tell?David Kanter2013/01/01 07:58 PM
      Games at 2048 x 1536? Can you tell?Anon2013/01/03 12:04 AM
        Games at 2048 x 1536? Can you tell?Andy2013/01/03 04:05 AM
          Games at 2048 x 1536? Can you tell?Anon2013/01/06 02:10 AM
            Games at 2048 x 1536? Can you tell?EBFE2013/01/06 11:07 PM
    Games at 2048 x 1536? Can you tell?Seni2013/01/01 09:17 PM
    Games at 2048 x 1536? Can you tell?anon2013/01/02 04:00 AM
      Games at 2048 x 1536? Can you tell?EduardoS2013/01/02 09:07 AM
        Games at 2048 x 1536? Can you tell?anon2013/01/03 12:21 AM
          Games at 2048 x 1536? Can you tell?EduardoS2013/01/03 01:03 AM
      Games at 2048 x 1536? Can you tell?Ricardo B2013/01/02 11:01 AM
        Games at 2048 x 1536? Can you tell?Michael S2013/01/03 01:35 AM
          Games at 2048 x 1536? Can you tell?Andy2013/01/03 04:10 AM
            Games at 2048 x 1536? Can you tell?Michael S2013/01/03 06:33 AM
              Games at 2048 x 1536? Can you tell?rwessel2013/01/03 11:56 AM
                Games at 2048 x 1536? Can you tell?rwessel2013/01/03 11:56 AM
                  Oh, for crying out load, would you fix this stupid thing to make it a tiny bit harder to double post (NT)rwessel2013/01/03 12:01 PM
                    /load/loud/ grrr... (NT)rwessel2013/01/03 12:02 PM
                    Oh, for crying out load, would you fix this stupid thing to make it a tiny bit harder to double postDavid Kanter2013/01/10 10:07 AM
                      Oh, for crying out load, would you fix this stupid thing to make it a tiny bit harder to double postrwessel2013/01/10 09:33 PM
                        Oh, for crying out load, would you fix this stupid thing to make it a tiny bit harder to double postrwessel2013/01/10 09:35 PM
                          Oh, for crying out load, would you fix this stupid thing to make it a tiny bit harder to double postrwessel2013/01/10 09:40 PM
                          Me, too! :-)Mark Roulo2013/01/11 08:57 AM
              85 Hz CRT monitorsDoug S2013/01/04 02:12 PM
                85 Hz CRT monitorsAnon2013/01/06 02:16 AM
                  85 Hz CRT monitorsDoug S2013/01/06 12:40 PM
                    The benefits of (non-perceived) flickerMax2013/01/07 01:43 PM
                      The benefits of (non-perceived) flickerAnon2013/01/07 02:59 PM
                        The benefits of (non-perceived) flickerPaul2013/01/07 05:40 PM
                      The benefits of (non-perceived) flickerUngo2013/01/07 03:57 PM
                        The benefits of (non-perceived) flickerDoug S2013/01/07 04:40 PM
                        The benefits of (non-perceived) flickerMax2013/01/07 06:16 PM
                          The benefits of (non-perceived) flickerAnon2013/01/09 09:50 PM
                            The benefits of (non-perceived) flickerMax2013/01/10 06:18 AM
                              The benefits of (non-perceived) flickerSM2013/01/10 12:37 PM
                        The benefits of (non-perceived) flickerdmsc2013/01/07 07:36 PM
                          The benefits of (non-perceived) flickerEBFE2013/01/07 11:08 PM
                          The benefits of (non-perceived) flickerAnon2013/01/09 10:07 PM
                            The benefits of (non-perceived) flickerdmsc2013/01/10 04:41 PM
                              The benefits of (non-perceived) flickerUngo2013/01/11 05:20 PM
                    perception of flickerCarlie Coats2013/01/08 04:39 PM
                      perception of flickerDoug S2013/01/08 10:25 PM
                        perception of flickerAnon2013/01/09 10:09 PM
        Games at 2048 x 1536? Can you tell?anongamer2013/01/03 04:31 AM
      Games at 2048 x 1536? Can you tell?Richard Cownie2013/01/02 02:56 PM
        Tablet vs. console GPUsDoug S2013/01/02 04:09 PM
          Tablet vs. console GPUsRicardo B2013/01/02 04:50 PM
      Games at 2048 x 1536? Can you tell?Anon2013/01/03 12:13 AM
        Games at 2048 x 1536? Can you tell?anon2013/01/03 12:54 AM
          The Hobbit is at 48 fpsRichard Cownie2013/01/03 07:02 AM
            Thanks for the interesting data point! (NT)Paul A. Clayton2013/01/03 09:26 AM
            The Hobbit is at 48 fpsRicardo B2013/01/03 11:17 AM
              The Hobbit is at 48 fpsRichard Cownie2013/01/03 11:55 AM
                The Hobbit is at 48 fpsRichard Cownie2013/01/03 12:44 PM
                  The Hobbit is at 48 fpsRicardo B2013/01/03 01:42 PM
                    The Hobbit is at 48 fpsRichard Cownie2013/01/03 02:31 PM
                      The Hobbit is at 48 fpsy2013/01/04 07:47 PM
                        Exposure time traditionsNathan Monson2013/01/05 12:21 PM
                          Exposure time traditionsy2013/01/05 09:59 PM
                            Exposure time traditionsanon2013/01/06 01:46 AM
                              Exposure time traditionsJukka Larja2013/01/07 08:52 AM
                            Exposure time traditionsdmsc2013/01/06 09:34 AM
                              Exposure time traditionsy2013/01/06 09:11 PM
                                Exposure time traditionsdmsc2013/01/07 07:42 PM
                                  Exposure time traditionsRichard Cownie2013/01/09 08:23 AM
                        The Hobbit is at 48 fpsMichael S2013/01/05 12:30 PM
                          The Hobbit is at 48 fpsRichard Cownie2013/01/05 03:45 PM
                            The Hobbit is at 48 fpsAnon2013/01/06 02:03 AM
                              The Hobbit is at 48 fpsRichard Cownie2013/01/06 06:43 AM
                                The Hobbit is at 48 fpsJukka Larja2013/01/07 09:18 AM
                                  The Hobbit is at 48 fpsRichard Cownie2013/01/07 05:21 PM
                                    The Hobbit is at 48 fpsJukka Larja2013/01/08 09:29 AM
                                      The Master 65mm/70mmRichard Cownie2013/01/08 12:52 PM
                                        The Master 65mm/70mmJukka Larja2013/01/11 02:07 PM
                                          The Master 65mm/70mmRichard Cownie2013/01/11 07:05 PM
                                            The Master 65mm/70mmJukka Larja2013/01/11 11:35 PM
                                              The Master 65mm/70mmDoug S2013/01/12 02:16 PM
                                                Pixar and 3DRichard Cownie2013/01/13 07:07 AM
                                            The Master 65mm/70mmDavid Hess2013/01/12 02:06 PM
                                            The Master 65mm/70mmMegol2013/01/16 01:02 PM
                                        The Master 65mm/70mmRakesh Malik2013/01/23 12:28 PM
                                    The Hobbit is at 48 fpsAnon2013/01/09 10:18 PM
                                      The Hobbit is at 48 fpsRichard Cownie2013/01/10 01:07 PM
                                        The Hobbit is at 48 fpsAnon2013/01/10 03:50 PM
                                          The Hobbit is at 48 fpsRichard Cownie2013/01/10 06:05 PM
                                            The Hobbit is at 48 fpsAnon2013/01/13 12:48 AM
                                              I don't care about workflowRichard Cownie2013/01/13 06:55 AM
                                                I don't care about workflowAnon2013/01/13 04:15 PM
                                                  I don't care about workflowRichard Cownie2013/01/13 08:02 PM
                                            The Hobbit is at 48 fpsRakesh Malik2013/01/23 01:38 PM
                              The Hobbit is at 48 fpsRakesh Malik2013/01/23 12:15 PM
                  The Hobbit is at 48 fpsJukka Larja2013/01/05 07:00 AM
                    Realism and fiction?Paul A. Clayton2013/01/05 09:42 AM
                      Realism and fiction?Jukka Larja2013/01/06 02:32 AM
                        Or the second factor?Paul A. Clayton2013/01/06 04:35 AM
                          Or the second factor?Jukka Larja2013/01/07 09:07 AM
                        further reading (was: Realism and fiction?)hobold2013/01/06 04:47 PM
                          Thanks. I'll probably check that out soon. (NT)Jukka Larja2013/01/07 09:13 AM
            The Hobbit is at 48 fpsrwessel2013/01/03 09:57 PM
            VR and latencyCarlie Coats2013/01/04 05:30 PM
              VR and latencyRicardo B2013/01/04 07:49 PM
                VR and latencyJoe2013/01/05 02:01 AM
            The Hobbit is at 48 fpsBernd Schmidt2013/01/08 03:42 AM
        Games at 2048 x 1536? Can you tell?anongamer2013/01/03 05:01 AM
        Games at 2048 x 1536? Can you tell?Ricardo B2013/01/03 08:06 AM
    Games at 2048 x 1536? Can you tell?Marco Salvi2013/01/02 10:31 AM
    Games at 2048 x 1536? Can you tell?Ungo2013/01/02 04:28 PM
    Games at 2048 x 1536? Can you tell?Jukka Larja2013/01/05 07:13 AM
  Lessons in Technology and Innovation from the iPad 3 Graphics and DisplayRichard Cownie2013/01/02 10:29 AM
  iPad GPU Clock Speedsltcommander.data2013/01/05 12:32 PM
    iPad GPU Clock SpeedsBrett2013/01/06 11:48 PM
      iPad GPU Clock SpeedsDoug S2013/01/07 10:17 AM
      iPad GPU Clock SpeedsDavid Kanter2013/01/09 01:48 PM
    iPad GPU Clock SpeedsUngo2013/01/07 04:26 PM
      iPad GPU Clock SpeedsDavid Kanter2013/01/09 01:42 PM
  Lessons in Technology and Innovation from the iPad 3 Graphics and Displaympx2013/01/06 05:35 AM
    Lessons in Technology and Innovation from the iPad 3 Graphics and DisplayMegol2013/01/07 04:45 AM
      Lessons in Technology and Innovation from the iPad 3 Graphics and Displaympx2013/01/07 12:56 PM
        Lessons in Technology and Innovation from the iPad 3 Graphics and DisplayDoug S2013/01/07 04:53 PM
        Lessons in Technology and Innovation from the iPad 3 Graphics and Displayanon2013/01/07 06:26 PM
      Lessons in Technology and Innovation from the iPad 3 Graphics and DisplayMaynard Handley2013/02/13 12:19 PM
  Lessons in Technology and Innovation from the iPad 3 Graphics and DisplayVincent Diepeveen2013/01/06 03:29 PM
    Lessons in Technology and Innovation from the iPad 3 Graphics and DisplayDoug S2013/01/07 10:53 AM
      Lessons in Technology and Innovation from the iPad 3 Graphics and Displaympx2013/01/07 12:48 PM
      Lessons in Technology and Innovation from the iPad 3 Graphics and DisplayVincent Diepeveen2013/01/08 09:53 AM
        Lessons in Technology and Innovation from the iPad 3 Graphics and DisplayDoug S2013/01/08 01:29 PM
          Lessons in Technology and Innovation from the iPad 3 Graphics and DisplayLlew2013/01/26 12:06 PM
            Lessons in Technology and Innovation from the iPad 3 Graphics and DisplayDavid Kanter2013/01/26 01:43 PM
              Lessons in Technology and Innovation from the iPad 3 Graphics and DisplayLlew2013/01/26 02:09 PM
            Wii U has three cores clocked at about 1.2 ghz or so. (NT)I.S.T.2013/01/29 07:37 PM
Reply to this Topic
Name:
Email:
Topic:
Body: No Text
How do you spell avocado?