08-10-12 05:09 PM
171 12345 ...
tools
  1. kennyliu's Avatar
    You are misinformed. PlayBook has a dual core cpu.
    He was talking about GPU and SGX540 is indeed single core.
    03-13-12 01:13 PM
  2. Scotter75's Avatar
    For the older crowd, we can remember the days when it took hours, yes hours to download a file that today would be considered relatively small. Processor speeds as well as the Internet speed are really fast. The United Nations has recently approved the specs for 4G networks that will soon be rolled out if it hasn't already started. 4G will be a thousand times faster than today's normal Internet speed. The newer PB's will run like greased lightning on that upcoming network. You can stay with the existing PB, but over time you will notice things slowing down. It has always been this way. You can't keepnup with the speed of change.
    03-13-12 01:13 PM
  3. Vindicators's Avatar
    HAHA, I agree, he is very vague about his point. I think he is referring to the way playbook does multitasking appears to be more powerful.

    Keep in mind that number of cores for the GPU or CPU does not directly translate the number of times the performance to one. Dual core 1.0GHz DOES NOT mean 1.0GHz + 1.0GHz = 2.0GHz. This will apply to the new GPU used in the new iPad. We also have to remember that the new iPad or the iPad 2 alone has more pixels to render, and the new iPad has significantly a lot more pixels to operate. In real world situation, Playbook and new iPad to render the same scene, it is not a fair fight - results may vary.
    For CPU, it is true. But game and graphic task utilize multi SIMD GPU since forever, it is not a problem. Terga 3 GPU have 12 SIMD, SGX543MP4 = 4x4 = 16 SIMD.

    Also, take games for example, they not always render shader at native resolution, so pixels may not necessary affect the performance too much.
    app_Developer likes this.
    03-13-12 01:23 PM
  4. app_Developer's Avatar
    For CPU, it is true. But game and graphic task utilize multi SIMD GPU since forever, it is not a problem. Terga 3 GPU have 12 SIMD, SGX543MP4 = 4x4 = 16 SIMD.

    Also, take games for example, they not always render shader at native resolution, so pixels may not necessary affect the performance too much.
    That's right, OpenGL implementations in particular have been shown to scale very nicely to multiple GPU cores.

    And you're right about the impact of resolution on performance as well. Even if you render at native resolution, it's still not true that this necessarily becomes the limiting factor. It depends on what and how you are rendering.
    03-13-12 01:30 PM
  5. q649's Avatar
    No its not. Web page load time will depend on network speeds and bandwidth - even with devices sitting right next to each other. Also - what is cached in RAM? How much RAM is available? Are there internal register conflicts?

    Do you honestly think you could put (2) 1 Ghz PB' next to each other and they will respond exactly the same, at the exact same speed? I can tell you from experience they dont. But according to your logic - they should. What would be the discrepancies between them then?

    You people seem to think that an increase in processor speed automatically equates to a faster moving device / machine when in fact there are multiple variables involved.
    I make no assumptions on your educational background, but you seem quite willing to make assumptions about other posters.

    By your logic there is no way to benchmark anything. I couldn't care less about your experiences or feelings when it comes to benchmarks. I care about data. No two devices will respond exactly the same, but unless they are configured differently, or have different background processes running, they should be comparable (within a small degree of error).

    Fresh reboot, cleared browser cache, same network, same OS. Multiple iterations of webpage load times, averaged out. This is real-world use, which 99% of people care about.

    You claim that it won't make any noticeable (measurable) difference. I claim that it will. It will be easy enough for CrackBerry to demonstrate, once the new PB comes out.
    03-13-12 01:54 PM
  6. FSeverino's Avatar
    FIIIIIIIIIIIIIIIIIIIIGHT!

    lol
    03-13-12 02:19 PM
  7. Rudee66's Avatar
    With the 1.5Ghz processor, the frequent browser crashes will be noticeably faster.
    03-13-12 02:20 PM
  8. morrislee's Avatar
    lol. you guys are funny. its always the same line of reasoning with complete fanboys.

    ok, look at this. you guys said the ipad2 processor is 4x as powerful as the PB (im not debating if this is true im just going with what you said). If this is true then the PB should be at the very least 2x as slow or unresponsive... yet all three ipad users i know also bought playbooks and have ALL said that the OS/GUI, multitasking and overall performance is MUCH QUICKER then the ipads. So... if the processor is so archaic then why is it doing so much more?

    Again, iOS has many features that PB does not... but that is not the issue. The issue is that it was stated the the ipad processor can STOMP all over the PBs (i am also not saying this is wrong, the ipad processor may be faster, but the fact that you think it is 8x faster leads me to my orginial argument of misinformation). SHOW ME HOW!
    Playbook has all the running apps cached to memory, that is why switch to and from apps is very fluid. This caused significant use of RAM where in one case i had as low as 44MB available!

    For the older crowd, we can remember the days when it took hours, yes hours to download a file that today would be considered relatively small. Processor speeds as well as the Internet speed are really fast. The United Nations has recently approved the specs for 4G networks that will soon be rolled out if it hasn't already started. 4G will be a thousand times faster than today's normal Internet speed. The newer PB's will run like greased lightning on that upcoming network. You can stay with the existing PB, but over time you will notice things slowing down. It has always been this way. You can't keepnup with the speed of change.
    Haha, i remember the dial up days... And the noise it makes!

    For 4G speed, if the device cant write to storage as fast as the download speed, then you will be wasting it. Websites are caches to disk first then the CPU process the data to reflect what is shown.

    Current "4G" is around 75mbps download, means 9.375MB/s write speed is required. When I copied data over to playbook, i usually get 3-5MB/s write speed, so this is capping our data speed. The new playbook spec sheet floating around only shows HSPA+ dual band 42mbps (5.25MB/s). So dont get your hopes up if they did not upgrade the memory speed
    For CPU, it is true. But game and graphic task utilize multi SIMD GPU since forever, it is not a problem. Terga 3 GPU have 12 SIMD, SGX543MP4 = 4x4 = 16 SIMD.

    Also, take games for example, they not always render shader at native resolution, so pixels may not necessary affect the performance too much.
    You are correct for GPU shader, but i was referring to more traditional rendering.

    How well the multicore management will have a significant impact in utilizing the full speed. For multi-threaded situations, apps have to distribute the jobs, and to distribute the jobs for multicore and mutithread will cause some drops of performance as the 2 or more cores will likely communicate and transfer data first before execution. Therefore, dual core 1.0GHz does not equal 2.0GHz (this applies to all types of muti core chips including GPU)

    I stand corrected.
    Last edited by morrislee; 03-13-12 at 02:33 PM.
    03-13-12 02:22 PM
  9. FSeverino's Avatar
    Playbook has all the running apps cached to memory, that is why switch to and from apps is very fluid. This caused significant use of RAM where in one case i had as low as 44MB available!

    Sooo... true multi tasking...
    I cannot play a movie and edit a document at the same time on an ipad (yes i know i cant do BOTH at exactly the same time, but if you quickly flip through the apps the movie will still play). Nothing anyone can say will change that. If i use more RAM for doing more things then im pretty sure nothing is wrong with that.
    03-13-12 02:42 PM
  10. morrislee's Avatar
    I cannot play a movie and edit a document at the same time on an ipad (yes i know i cant do BOTH at exactly the same time, but if you quickly flip through the apps the movie will still play). Nothing anyone can say will change that. If i use more RAM for doing more things then im pretty sure nothing is wrong with that.
    It is more of a preference actually if you think about it. If it eats too much RAM, it starts randomly closing my apps. On my android, it does not do the same but at least the state of the apps are saved and will resume where i have left off eve if it is no longer loaded in memory. It is just a more robust design.

    iOS uses similar muti tasking technique like Android.
    03-13-12 02:48 PM
  11. Vindicators's Avatar
    I cannot play a movie and edit a document at the same time on an ipad (yes i know i cant do BOTH at exactly the same time, but if you quickly flip through the apps the movie will still play). Nothing anyone can say will change that. If i use more RAM for doing more things then im pretty sure nothing is wrong with that.
    Still? You must love the FUD huh?

    Video from stock player, 3rd party player, video in browser, from podcast app,... all can play in the background.
    03-13-12 02:50 PM
  12. FSeverino's Avatar
    Just out of curiosity... how many apps where you running.

    I have had the battery guru, browser, message and docs to go apps open while watching a movie with ZERO problems...

    what else can you really be doing?
    (not trying to be mean, i honestly am really curious bc i thought i was doing a good job multi tasking)
    03-13-12 02:51 PM
  13. conix67's Avatar
    I overclocked my Motorola Atrix to 1.5GHz vs the stock. Speed, only shows about 500-700ms difference in the sunspider benchmark. The Atrix and the playbook gets about the same score when both are at 1.0GHz.

    The above test pretty much shows a good simulation we would get if we bumped up to 1.5GHz if it is the same CPU architecture. If the architecture is upgraded, than we might see more benefits.

    In the mean time, QNX eats RAM like cupcakes, I presume upgrading the amount of RAM will improve STABILITY to apps so they are not closed as often. Improving the storage read/write speed will greatly improve the app loading times.

    To sum up, next version of playbook should keep 1GHz for battery performance and add more RAM to increase stability and upgrade the type of storage chips to faster ones will improve visual performance (decreases loading times)

    I have read an article about improving the speed of the storage chips will improve the entire operating system's performance in most cases. This was on Gizmodo I think, cant find the article anymore.
    Are you able to confirm that the processor actually ran at 1.5Ghz static clock throughout? Modern processors are designed for power, especially on Tablets. You may set the top end clock to 1.5Ghz, but if the CPU is not designed for that frequency, it may dynamically downclock to designed range and user may not have means to "fix" the clock to certain frequency, and very possibly your results indicate that.

    Also, increasing clock speed doesn't always increase performance for other reasons, if the product is designed for that speed. Somewhere along the processing pipeline, higher frequency on one side may increase latency of certain stages, causing bottleneck that could offset the benefits of the core clock increase.

    So, when a processor is simply overclocked, it may not provide the performance level in direct comparison to actual performance you would achieve from the processor that is designed to perform at the said frequency.
    03-13-12 02:57 PM
  14. FSeverino's Avatar
    Still? You must love the FUD huh?

    Video from stock player, 3rd party player, video in browser, from podcast app,... all can play in the background.
    not when i tried. I and especially not when i tried two videos at once.
    03-13-12 02:58 PM
  15. morrislee's Avatar
    When i first got it, i was doing about 5-10 apps but i started to notice that because of my web browsing habbit (10-15 tabs at least) will cause everything to close. Most of these cases will have my musics app still running, so i am sure there must be a way to force browser to be kept in memory priority over other apps.

    BTW, an easy website to crash your playbook lol, load one of the more active development forum page on xda, about 50% of the time will make the playbook think it is a memory hog and closes the browser haha
    03-13-12 03:00 PM
  16. jwn66's Avatar
    Wow Imagine that, a 2nd product coming out over a year later with better specs. Isn't that what every business in the world does. Seems familiar,.. Get over it.
    03-13-12 03:02 PM
  17. morrislee's Avatar
    Are you able to confirm that the processor actually ran at 1.5Ghz static clock throughout? Modern processors are designed for power, especially on Tablets. You may set the top end clock to 1.5Ghz, but if the CPU is not designed for that frequency, it may dynamically downclock to designed range and user may not have means to "fix" the clock to certain frequency, and very possibly your results indicate that.

    Also, increasing clock speed doesn't always increase performance for other reasons, if the product is designed for that speed. Somewhere along the processing pipeline, higher frequency on one side may increase latency of certain stages, causing bottleneck that could offset the benefits of the core clock increase.

    So, when a processor is simply overclocked, it may not provide the performance level in direct comparison to actual performance you would achieve from the processor that is designed to perform at the said frequency.
    Yeap, and gradually increasing the clock speed will gradually see different result at each higher clock speed. I also tried locking it to 1.5GHz vs dynamic and they both showed up with the same result, so it is jumping to 1.5GHz properly

    Btw, Cortex-A8 is designed up to 2.0GHz, only if there is enough power supplied for it
    03-13-12 03:03 PM
  18. Vindicators's Avatar
    So now it is play a movie and edit a document or play two video at once?

    Lol, when one FUD fail, quicky change to another, exactly like the other thread.
    03-13-12 03:09 PM
  19. conix67's Avatar
    Yeap, and gradually increasing the clock speed will gradually see different result at each higher clock speed. I also tried locking it to 1.5GHz vs dynamic and they both showed up with the same result, so it is jumping to 1.5GHz properly

    Btw, Cortex-A8 is designed up to 2.0GHz, only if there is enough power supplied for it
    If that's the case, could you run the same set of tests when 1.5Ghz units come out? I'd like to see what the results will be.

    The processor can be designed for certain speed, but there could be set power budget for it, where exceeding certain frequency may not really push the speed internally. I'm not familiar with ARM CPU design, but I've seen similar things on other products, so just saying that it is odd to not see the speed increase, not much at all, with 50% difference in effective clock frequency.
    03-13-12 03:16 PM
  20. morrislee's Avatar
    If that's the case, could you run the same set of tests when 1.5Ghz units come out? I'd like to see what the results will be.

    The processor can be designed for certain speed, but there could be set power budget for it, where exceeding certain frequency may not really push the speed internally. I'm not familiar with ARM CPU design, but I've seen similar things on other products, so just saying that it is odd to not see the speed increase, not much at all, with 50% difference in effective clock frequency.
    THey have moved on to Cortex-A15 for the next generations of mobile devices, I am afraid that we might see A8 at stock 1.5GHz.

    I have tried at 1.1GHz (increased performance and heat), 1.2GHz, 1.3GHz, 1.4GHz, and eventually 1.5GHz. I see increase in benchmarks in all increments as well as heat. 1.5GHz gets really hot, and it was also required to push the supplying voltage to 1.4v from 1GHz stock at 1.0v. How well the architecture handles the software has the most impact on the performance. My netbook first gen Atom N270 1.6GHz single core scored 1300ms on sunspider, dual core tegra 2 in my atrix scores about 1550ms at 1.5GHz (2200ms at 1GHz) this shows the architecture and software it is using will have different efficiency (both the netbook and the Atrix used webkit based browsers, so it should produce a fair comparison)

    There was also a 1.6GHz overlocking mod, but I have not seen performance gain over 1.5GHz and I also noticed that some diagnostics apps shows it is still maxxed at 1.5GHz through logs.

    Edit: Actually, I recalled that there were a line of Samsung Galaxy S2 on Cortex-A8 at stock 1.5GHz, later on, it was finally released at 1.2GHz. After release, overclocking mods were made for it, but it shows similar benchmarking results in the same overclocked frequencies as the Tegra 2

    And I think I miss understood your question, I believe you are referring to 1.5GHz models of playbook. I dont think I will be getting one when it is released.... I JUST got this playbook not too long ago and it does what i need in school (notes taking and web browsing). But if I found an answer from a reliable source, and if I remember to, I will post it here. I do doubt a great difference will be seen in performance from 1.0 to 1.5GHz upgrade. I actually think Playbook's performance in battery life is great, I have it on pretty much consistent use over my school hours and when I get home (8-3:30PM + 1.5hours for the public transit traveling use). If we increased the clock speed on the same CPU, we will kill the battery life.
    Last edited by morrislee; 03-13-12 at 03:57 PM.
    03-13-12 03:29 PM
  21. lnichols's Avatar
    Think I'll wait for the OMAP5 or S4 CPU's that should be in the BB10 devices to show up in the Playbooks before I'll upgrade mine. I don't see a clock increase or built in 3G/4G as being a selling point since I have a BB phone and the bridge is working awesome now for everything. NFC would be cool, but not enough at this point to justify a replacement. I can't even get the NFC in my T-Mobile Bold to work because of RIM bending over to T-Mobile's ridiculous request to strip out the functionality.
    03-13-12 03:47 PM
  22. missing_K-W's Avatar
    I like how everyone is an arm chair technological benchmarking genius on these boards.

    Like I said before - I could stick 2 tablets in front of you one with a 1 ghz and one with a 1.5 ghz processor and you would not be able to perceive the difference. To say that you can is ridiculous.
    Let's see some facts? Or do you not roll that way lol
    03-13-12 03:48 PM
  23. emirozmen's Avatar
    Browsing would make a difference.
    03-13-12 03:50 PM
  24. howarmat's Avatar
    Just for comparison and clarification, i have ran my transformer overclock to 1.4-1.6 ghz range many times before. You can maybe say its "quicker" and sure it will benchmark higher but its not like i cut 2 seconds of loading webpages and apps loaded several seconds faster. It might improve flash and video playback and that is probably the biggest thing that one might notice.

    In honestly it ran about the same as another kernel that was tweaked at 1 to 1.2 ghz. The modifcations to the kernel code itself can make a much better impact sometimes than cranked up the cycles.

    There are certain instances where 1.5 ghz PB will make a difference, i am going to guess than 90% of people when handed the devices arent going to be able to see the differences without doing a benchmark though. The 1 GB of RAM is still going to be the limited factor and cause the bottlenecks just as it does now.

    Just like with any line of tech, cost will become a big factor. how many people are going to be willing to sell their current 16gb PBs for around 175-225 and then go get the new one for 400+ given the only changes are a cell radio, processor bump and NFC? probably not many really. I doubt I would and I upgrade my phone every 8 months generally.
    app_Developer likes this.
    03-13-12 03:55 PM
  25. missing_K-W's Avatar
    THey have moved on to Cortex-A15 for the next generations of mobile devices, I am afraid that we might see A8 at stock 1.5GHz.

    I have tried at 1.1GHz (increased performance and heat), 1.2GHz, 1.3GHz, 1.4GHz, and eventually 1.5GHz. I see increase in benchmarks in all increments as well as heat. 1.5GHz gets really hot, and it was also required to push the supplying voltage to 1.4v from 1GHz stock at 1.0v. How well the architecture handles the software has the most impact on the performance. My netbook first gen Atom N270 1.6GHz single core scored 1300ms on sunspider, dual core tegra 2 in my atrix scores about 1550ms at 1.5GHz (2200ms at 1GHz) this shows the architecture and software it is using will have different efficiency (both the netbook and the Atrix used webkit based browsers, so it should produce a fair comparison)

    There was also a 1.6GHz overlocking mod, but I have not seen performance gain over 1.5GHz and I also noticed that some diagnostics apps shows it is still maxxed at 1.5GHz through logs.

    Edit: Actually, I recalled that there were a line of Samsung Galaxy S2 on Cortex-A8 at stock 1.5GHz, later on, it was finally released at 1.2GHz. After release, overclocking mods were made for it, but it shows similar benchmarking results in the same overclocked frequencies as the Tegra 2

    And I think I miss understood your question, I believe you are referring to 1.5GHz models of playbook. I dont think I will be getting one when it is released.... I JUST got this playbook not too long ago and it does what i need in school (notes taking and web browsing). But if I found an answer from a reliable source, and if I remember to, I will post it here. I do doubt a great difference will be seen in performance from 1.0 to 1.5GHz upgrade. I actually think Playbook's performance in battery life is great, I have it on pretty much consistent use over my school hours and when I get home (8-3:30PM + 1.5hours for the public transit traveling use). If we increased the clock speed on the same CPU, we will kill the battery life.
    The OMAP 4 4430 currently in the PB, is a cortex A9 . We are unaware at this point whether it will be the 4460/4470 which both are A9

    The next PB could infact be the 4470 clocked to 1.5ghz

    OMAP™ Mobile Processors - OMAP™ 4 Platform
    Last edited by missing_K-W; 03-13-12 at 04:08 PM.
    morrislee likes this.
    03-13-12 04:01 PM
171 12345 ...
LINK TO POST COPIED TO CLIPBOARD