Announcement

Collapse
No announcement yet.

Allwinner A80 Octa-Core Hardware Coming Next Month

Collapse
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • Allwinner A80 Octa-Core Hardware Coming Next Month

    Phoronix: Allwinner A80 Octa-Core Hardware Coming Next Month

    In June we should start seeing the Allwinner A80-based hardware designs hitting the market for a much needed performance boost for Allwinner SoCs...

    http://www.phoronix.com/vr.php?view=MTY4MjE

  • #2
    As an owner of several Chinese tablets using the Allwinner and other chipets, they are slow. They might have the clock speed and the cores and the instruction set, but the IPC is low. THat combined with REALLY slow NAND storage makes the whole unit feel sluggish.

    Don't get me wrong, it'll play 1080p and all that stuff fine (mostly, most codecs) but switching between apps, launching new apps etc is dog slow. But, that's why they are cheap. So, I don't hold much hope for this one other than, hey, it's nice to see improvements.

    I know I won't be picking up any more Chinese tablets anyway. Nexus are too cheap.

    Comment


    • #3
      What's the difference? They're all made there, all using variations of Arm CPUs. I suspect RAM and the flash storage have more to do with application switching performance than CPU speed anyway.

      But an octa-core with the performance of a quad-core. Doesn't deserve the name.

      Comment


      • #4
        With a quad a15 (the fastest ARM designed core available) and a pretty beefy GPU, this may be Allwinner trying to move up from the budget end? All their previous chips were bargin-basement CPUs, so this is a big change.

        Comment


        • #5
          Originally posted by jonnyh View Post
          With a quad a15 (the fastest ARM designed core available) and a pretty beefy GPU, this may be Allwinner trying to move up from the budget end? All their previous chips were bargin-basement CPUs, so this is a big change.
          It's one thing to honour the instruction set etc, anotherthing to be fast. The allwinner were never fast, at all. Like I said before, the whole tablet experience is going to suck unless you've got fast RAM and NAND as well.

          Comment


          • #6
            "Making the Allwinner A80 more appealing too is that its graphics are upgraded to the PowerVR G6230."

            Uh WUT.....

            have you been smoking?



            SERIOUSLY!!!

            Comment


            • #7
              Originally posted by droidhacker View Post
              "Making the Allwinner A80 more appealing too is that its graphics are upgraded to the PowerVR G6230."

              Uh WUT.....

              have you been smoking?



              SERIOUSLY!!!
              What a worthless comment. Wish I could make it so I never have to see your comments again.

              Comment


              • #8
                @Cyborg16

                I can see why you might think that, but the main difference in this SoC configuration is that all 8 cores can run at the same time, whereas other bigLITTLE configurations haven't allowed for this. Hence the reason they are advertising this as 8 core, and the first of it's kind. Though I suspect we'll see others soon.

                Having all 8 cores able to run at the same time, rather than just switching from a15/a7 should provide significant performance boosts.


                Originally posted by Cyborg16 View Post
                What's the difference? They're all made there, all using variations of Arm CPUs. I suspect RAM and the flash storage have more to do with application switching performance than CPU speed anyway.

                But an octa-core with the performance of a quad-core. Doesn't deserve the name.

                Comment


                • #9
                  Ignore list man, ignore list.

                  Though I agree with him on this one, Mali has better prospects with good open drivers than anything PowerVR.

                  Comment


                  • #10
                    Where there is smoke, therse fire

                    PowerVR. Burn it with fire. And don't inhale

                    While the PowerVR delivers power there is little hope for contiguous usage, even Intel couldn't cope with it. Driver development is left to the final vendor. Imagination is hopeless. Their Marketing proved it once again:
                    http://www.cnx-software.com/2014/02/...rvr-g6230-gpu/

                    I wished they'd licensed something like an ARM Mali, Vivante GC, Quallcomm Ardreno, NVIDIA SoC GPU. All with some binary contiguously and hopes for free (mainline) drivers.

                    There might be hope, as for the A23, which was released after the A31 (PowerVR), again a Mali GPU was used.

                    The A23 seems more of an iteration of the A31 than A20 looking at the messages on the linux-sunxi mailinglist

                    So let's hope for a next one with Mali.

                    Comment


                    • #11
                      Wrong link

                      Oops wrong article referenced
                      http://www.cnx-software.com/2014/01/...ig-little-soc/

                      Comment


                      • #12
                        Originally posted by Falcon1 View Post
                        Comments section in that link are hilarious. Powervr coming in to try to defend themselves. Their argument boiled down to "open source won't help you because other parts of the soc might be closed source too".
                        They also tried to say that you can't compare them to amd or nvidia because they're a "gpu maker not a chipset maker".... lol.

                        Comment


                        • #13
                          PoweVR Graphics...The Horror....The Horror...

                          Comment


                          • #14
                            Originally posted by bmourit View Post
                            @Cyborg16

                            I can see why you might think that, but the main difference in this SoC configuration is that all 8 cores can run at the same time, whereas other bigLITTLE configurations haven't allowed for this. Hence the reason they are advertising this as 8 core, and the first of it's kind. Though I suspect we'll see others soon.
                            I lie ARM and they have had great ideas in the past, but big.LITTLE has to be the result of a corporate stupid moment. I just see the idea as huge waste of die space that offers little gain for most users.
                            Having all 8 cores able to run at the same time, rather than just switching from a15/a7 should provide significant performance boosts.
                            Not anywhere near the boost you could get for 8 A15 cores and an architecture designed to support those cores. Hell you wouldn't even need 8 A15 cores to do the job. Combine that with freeing up space for cache and other hardware and you would be far better off.

                            I'm not seeing a big rush to big.LITTLE in the rest of the market so maybe my opinions are shared by some of the other SoC developers out there.

                            Comment


                            • #15
                              Originally posted by wizard69 View Post
                              I lie ARM and they have had great ideas in the past, but big.LITTLE has to be the result of a corporate stupid moment. I just see the idea as huge waste of die space that offers little gain for most users.


                              Not anywhere near the boost you could get for 8 A15 cores and an architecture designed to support those cores. Hell you wouldn't even need 8 A15 cores to do the job. Combine that with freeing up space for cache and other hardware and you would be far better off.

                              I'm not seeing a big rush to big.LITTLE in the rest of the market so maybe my opinions are shared by some of the other SoC developers out there.
                              big.LITTLE is because of energy saving. Using more die space doesn't matter because it gets turned off and doesn't use energy. A simple core like A7 needs far less energy to do the same stuff as an A15. It is only when an A7 doesn't have enough peak performance that the A15 gets enabled. If users were willing to wait for calculations in-order cpus would be the best and most power efficient cpus. But users want peak performance when scrolling etc.
                              I think big.LITTLE hasn't been used yet because the implementation wasn't smart enough to provide the possible energy savings. Tegra3 was not very power efficient.

                              Comment

                              Working...
                              X