Follow Slashdot stories on Twitter

 



Forgot your password?
typodupeerror
×
Google Technology

Google Nexus S Processor Overclocked To 1.2GHz 78

dkd903 writes "Though Google's Nexus S is powered by a single core Hummingbird processor, it looks like the one core would be enough to put LG's dual-core processor powered Optimus to shame. An XDA Forums user morfic has overclocked the processor on Nexus S up to 1.2GHz in a new kernel based on the Bionix NS1 mod."
This discussion has been archived. No new comments can be posted.

Google Nexus S Processor Overclocked To 1.2GHz

Comments Filter:
  • So... (Score:5, Funny)

    by fuzzyfuzzyfungus ( 1223518 ) on Sunday January 02, 2011 @09:28PM (#34740800) Journal
    How is the battery life after you install a pump for the coolant and a fan for the radiator?
    • Re: (Score:3, Insightful)

      by jsse ( 254124 )
      Funny as it seems, but overclocking could sometime save battery life as it could alter the default voltage usage of a particular frequency. At least, this is the case of Nokia N900 with Titan's Kernel Power, where we could choose 'starving' profile for overclocking with less voltage than default. Say it can run at almost half as much voltage at 600Mhz than normal. (fyi. N900 can be overclocked to 1.15GHz max. with Kernel Power)
      • Re:So... (Score:5, Informative)

        by adolf ( 21054 ) <flodadolf@gmail.com> on Sunday January 02, 2011 @09:59PM (#34740898) Journal

        Funny as it seems, but overclocking could sometime save battery life as it could alter the default voltage usage of a particular frequency. At least, this is the case of Nokia N900 with Titan's Kernel Power, where we could choose 'starving' profile for overclocking with less voltage than default. Say it can run at almost half as much voltage at 600Mhz than normal. (fyi. N900 can be overclocked to 1.15GHz max. with Kernel Power)

        That's not necessarily overclocking -- see also, undervolting or even underclocking.

        On my Droid 1, I do all three: It runs at, IIRC, 1GHz, some of the time. Its lowest clock speed is 125MHz, where it spends most of its time (half of the default lowest rate of 250MHz). And all of these speeds are at lower voltage than default.

        In the end, it's about a wash: I get a faster device for about the same battery life as I had at stock clocks. Heat generation is about the same, by my estimation, in common use.

        None of this is particularly new: I have a fanless, diskless K6-2 350 that gets used for some realtime audio processing tasks using KX audio drivers. It is equipped with a big heatsink, clocked down to 200MHz, and running at low voltage. The hard drive is a CF card on an IDE bus.

        It's stable as a champ, doesn't make a peep, and never gets too warm. (These days there's better options for that sort of work, with Atom and SSD, but it was the best I could come up with back then, and it still works just as well today as it did then.)

        • While I'm not familiar with this particular processor, in a great many of these lower-power chips you can get significant overclocking gains without noteworthy increased heat. The Pandora, for example, sports an OMAP3530 SoC (which includes an ARM Cortex V8) rated at 600mhz. People have overclocked this to 1ghz with no heat issues and only minor losses in battery life (the chip doesn't remain at maximum speed but underclocks itself when not under load).

          It's not like overclocking a normal desktop processor

      • The fastest speed at a given voltage is always the most efficient, so overclocking is really only a battery killer if you run at higher than stock voltages to reach the top speeds.
        n900 also becomes both faster and more efficient by disabling the 125 and 250 mhz frequencies. This is because they use very similar low voltages, with nokia's kernel they are close, and some of the other voltage profiles commonly used with titan's kernel they are identical. So it's better to do x cycles of work at 500mhz and slee

  • Hummingbird (Score:5, Funny)

    by MrEricSir ( 398214 ) on Sunday January 02, 2011 @09:33PM (#34740818) Homepage

    Does the Hummingbird processor require more nectar when it's overclocked?

  • by Anonymous Coward

    Don't really see this as a huge amount of news the old snapdragons have been running at 1.2GHz (well 1.19 anyways) with the same base 1GHz speed for ages.

    • by mcvos ( 645701 )

      GigaHertzes don't mean anything. Some processors are faster than others at the same clock speed. I believe the Cortex A8 (in the iPhone 3GS, Motorola Milestone/Droid, and I think also the N900) is at 600 MHz about as fast as a Snapdragon at 800Mhz. In the Milestone/Droid, it can be overclocked to at least 1GHz, which would be faster than a 1.2GHz Snapdragon.

      So what we really need to know is how fast this Hummingbird really is. Not in GHz, but in actual computation.

  • Really?? (Score:5, Insightful)

    by mejogid ( 1575619 ) on Sunday January 02, 2011 @09:42PM (#34740840)

    Given that the LG Optimus 2x - also known as the Star - benchmarks pretty impressively, I doubt a Nexus S overclock will put it to shame. In the quadrant benchmark the Nexus S has been noted as obtaining a score of 1474 [nexusoneforum.net], whilst the Star scores 1759 [mobilecrunch.com].

    Assuming linear scaling from 1ghz, the Nexus S would obtain 1769 with android 2.3 whilst the star manages the previously stated 1759 with android 2.2 which has received fewer performance optimisations. Whilst this is only one benchmark, the more graphically focussed ones favour the Star's tegra (nVidia) processor to an even greater extent.

    Therefore far from putting the new dual core Optimus to shame, I would argue that an overclocked Nexus S can just about keep level with it whilst using a faster OS version on benchmarks that favour it. Hardly putting the Star to shame...

    • Re:Really?? (Score:5, Informative)

      by teh31337one ( 1590023 ) on Sunday January 02, 2011 @10:04PM (#34740912)

      Quadrant is a pretty flawed test. [briefmobile.com]

      That said, based on some other benchmarks and their respective specs, tegra2 has roughly 2.5x more CPU power compared to the hummingbird SOC. (1ghz A9 runs 25% faster than 1ghz A8, and tegra 2 is a dual core A9) Anadtech's Linpack scores seem to show that too. [anandtech.com] (Ignore the bloated snapdragon class scores, it has floating point performance optimisations) Article here [anandtech.com]

      GPU performance is where it gets interesting. It seems like the PowerVR 540 GPU on the hummingbird SoC is better than the GPU used in the Tegra 2 SoC. Odd considering nVidia make the tegra2. Instances where Tegra 2 outperforms the hummingbird in GPU benchmarks are as far as i can tell down to the extra CPU power (roughly 250% faster)

      Samsung's upcoming [engadget.com] Orion chip also looks promising, and is a closer match to the Tegra 2.

      • by adolf ( 21054 )

        GPU performance is interesting, I guess, but it's interesting for different reasons to me than it seems to be for you:

        I, for one, am not surprised that PowerVR is doing better in this sector than nVidia, with the former having lost all relevance in the desktop and portable market, and the latter having spent the past few years trying to beat up ATI/AMD as top dog on the desktop. That PowerVR has a superior low-performance offering is really not very interesting.

        What is interesting is that people think that

        • by Anonymous Coward

          And I have probably 80% of the top 3D games for Android on my X. It's kind of neat playing Asphalt 5/NFS-Shift on the big-screen TV from my phone (at standard definition).

          With that said, I haven't found any Android games yet that I've played for more than 1 hour at a time. None are particularly interesting.

      • by imgod2u ( 812837 )

        I'm not sure how you arrived at that conclusion. Most of the benchmarks of real-world games -- while they are a bit rare -- show the Tegra 2 as being slightly faster; take Quake 3, for instance. Neocore is particularly slow on Tegra 2 because it's a benchmark made by Qualcomm that targets tile-based rendering and also offloads a lot of the FP workload to the CPU (where Tegra 2 is particularly bad).

        Realistically, Tegra 2 is pretty much a beast of a chip, but I'm curious about its power numbers.

      • by F34nor ( 321515 )

        Is any of this related to Samsung getting the DEC Alpha after Intel picked the bones?

    • Thanks for tossing numbers to this, as the article, the articles links, xda, etc make no mention or comparison to the Optimus that I could see. Couldn't you also assume a similar situation with an overclocked Tegra2 as well? 20% increase would put it at over 2100.
      • by Dr Max ( 1696200 )
        Mod the man up. Its comparing apples and oranges if only one of them has been overclocked.
  • And? (Score:5, Informative)

    by teh31337one ( 1590023 ) on Sunday January 02, 2011 @09:42PM (#34740842)
    It's a hummingbird. And it's been overclocked to 1.6ghz on Galaxy S phones. Besides, 1.2ghz OC is pretty standard. I've got my Galaxy S i9000 overclocked to 1.2ghz, and it's using the same voltage as 1ghz would. I've undervolted the other frequencies, and the battery life is great - better than stock 1ghz.
  • by msauve ( 701917 ) on Sunday January 02, 2011 @09:44PM (#34740852)
    move along.

    How is this news? The year old Motorola Droid (also using an ARM Cortex A8 core) was overclocked to 1.2 GHz quite a while ago (as with all overclocking, it works fine with some units, not so well with others). I understand the Hummingbird has had some microcode tweaks, and so is a bit more efficient than the OMAP (5-10%), but that's mice nuts. This isn't news, or even very interesting outside of Nexus S owner circles.
    • I was about to post this too. My milestone (same hardware as the droid) runs just fine at 1.2 GHz.
    • ... as with all overclocking, it works fine with some units, not so well with others ...

      Judging an overclocked CPU to be working fine is quite tricky. Not all overclocking induced failures are catastrophic or otherwise noticeable. Sometimes the failure is as simple as giving an incorrect answer, a 1+2=4 for example.

      • ... as with all overclocking, it works fine with some units, not so well with others ...

        Judging an overclocked CPU to be working fine is quite tricky. Not all overclocking induced failures are catastrophic or otherwise noticeable. Sometimes the failure is as simple as giving an incorrect answer, a 1+2=4 for example.

        Very true, and the CPU isn't the only issue. The stock speed that most phone vendors run their phones is often less than half of the processor's rated speed. So you can generally run the CPU quite a bit faster without it having any problems. Unfortunately, the rest of the components in the system may not be rated for the higher clock rate. Memory, for example.

        • The stock speed that most phone vendors run their phones is often less than half of the processor's rated speed. So you can generally run the CPU quite a bit faster without it having any problems.

          To elaborate on "generally" ...
          Running the CPU at its rated speed may generate more heat than the design of a particular device can accommodate. The design may only be within its thermal specs when the CPU is running at its reduced clock rate. Excess heat can cause incorrect results even if the clock rate is OK.

          • The stock speed that most phone vendors run their phones is often less than half of the processor's rated speed. So you can generally run the CPU quite a bit faster without it having any problems.

            To elaborate on "generally" ... Running the CPU at its rated speed may generate more heat than the design of a particular device can accommodate. The design may only be within its thermal specs when the CPU is running at its reduced clock rate. Excess heat can cause incorrect results even if the clock rate is OK.

            Very true, although if your OS has a CPU governor that dynamically adjusts the CPU clock based upon demand, you can probably get away with a lot. Average heating will be a lot less if the CPU is only responding to peak loads.

    • I have an OMAP 3 (Beagle board) running over clocked at 720MHz tasked with a Knight's tour solver it's been running for 6 weeks, on Angstrom. It draws about 1Watt (Video & audio turned off). To date it has solved 38M tours and 8.4B traversals from start pos . The current minimum depth is 24. I haven't even gotten the DSP into the mix yet. With a bit more work I could get this system to run 24/7/365 off a battery backed supply and add DSP solving support. I can let this sucker run for years if

  • by Lord Byron II ( 671689 ) on Sunday January 02, 2011 @10:31PM (#34740990)
    Not matter how fast a single core goes, there is still a cost associated with having to perform a bunch of context switches and you have to share all of the cache. There are benefits to having a real dual core implementation.
    • I want to see duals in Androids as much as you do, but laptops took forever to get dual cores compared to desktops, and netbooks are a year or two older, but are *still* waiting for that tech (and they're waay underpowered and under-RAM'd as is.)

      Cellphones, which don't have "core fever" due to different marketting than PCs, will take forever to justify the performance/battery life drawbacks. Forever, unless phones REALLY become just another PC running full-blown Windows. We do not want to go back to Window

      • by mlts ( 1038732 ) *

        Dual core devices are not new. The HTC Wizard had this in '05. However, the phone had separate cores so the radio could do what it needed to while the OS was able to do what it needed, both cores running at a fairly low clock speed. The device had a pretty amazing battery life though.

        I'd like to see multiple cores... perhaps fast and slow (but battery saving) cores and a smart process scheduler. This would go a way to help with battery life, as well as provide smoother performance for apps.

        • by Mulder3 ( 867389 ) on Sunday January 02, 2011 @11:33PM (#34741140)
          "Dual core devices are not new. The HTC Wizard had this in '05. However, the phone had separate cores so the radio could do what it needed to while the OS was able to do what it needed, both cores running at a fairly low clock speed. The device had a pretty amazing battery life though."

          That's not dual core... All current smartphones use that... One Application CPU that runs Android/Symbian/iPhoneOS/whatever and another CPU inside the modem that runs the GSM stack on a RTOS... Only very low end phones run the GSM stack on the app processor...
          • One Application CPU that runs Android/Symbian/iPhoneOS/whatever and another CPU inside the modem that runs the GSM stack on a RTOS... Only very low end phones run the GSM stack on the app processor...

            Actually, of those OSs, Symbian is the only one capable of running the OS and baseband on the same CPU. Not that Nokia uses that feature for the high end phones.

      • by 0123456 ( 636235 )

        netbooks are a year or two older, but are *still* waiting for that tech (and they're waay underpowered and under-RAM'd as is.)

        Atom-based netbooks can already run two threads with hyperthreading, so it's less of an issue. A dual-core Atom adds a sizable amount of power consumption and needs four threads to max it out; because of in-order execution there are usually plenty of stalls where the single core Atom can run a second thread at no cost to the first.

        Cellphones, which don't have "core fever" due to different marketting than PCs, will take forever to justify the performance/battery life drawbacks.

        The company I was working for in 2005 was building dual-core ARM chips for such lower-power devices; putting two low-frequency ARMs into a single chip ended up giving us the same

      • Motorola and LG IIRC are releasing ones in early 2011.

    • by Xyde ( 415798 )

      I know linux is obviously SMP enabled etc but does anyone know how well optimized Android is for taking advantage of multiple cores and scheduling? How about iOS? (I assume it has Grand Central/blocks etc) How useful could it possibly be? As best practices backgrounded apps should either be suspended or at the very least certainly not burning CPU with reckless abandon. I understand that maybe games and certain uses (video encoding? isn't it hardware assisted already?) would get increased performance from pa

  • I'm not seeing how a 200MHz overclock puts a phone with a 1000MHz aggregate clockspeed advantage to shame...

    Sure, performance doesn't scale linearly, but I'll still take a dual-core 1GHz Cortex A9 over a 1.2GHz single-core Cortex A8-alike any day.

  • An XDA Forums user morific has

    How many morifics are there on the XDA Forums?

  • I remember like 10 years ago when somebody converted a big ass A/C unit into a chip cooler and overclocked a Pentium III to 1.0 GHz, and it was a big deal...now here we are discussing PHONES that come OUT OF THE BOX with 1.0 GHz, running Quake 3, like it ain't no thang. Gotta love Moore's Law.

    • by quenda ( 644621 )

      meh, it was a big deal when an el cheapo Celeron 300 was overclocked to 450MHz routinely, with no special hardware. Almost enough to decode DVDs in real time :-)

  • Sooo since it is a google product does this mean I will get email up dates and promises of support only to find out later that I used MY personal computer that doesnt adhere to some "moral TOS" google has decided on and thus is banned from said support?
  • by brunes69 ( 86786 ) <slashdot@nOSpam.keirstead.org> on Monday January 03, 2011 @09:05AM (#34743130)

    The Samsung Galaxy S and it's cousins have the exact same processor and it was overclocked too 1.2 Ghz 3 months ago.

    Is it news because it has Google in the title? Seeing how Samsung makes the phone and it is basically a copy of the galaxy S with a few minor changes, I don't see why this is new news.

  • Maybe soon they will catch up to my Droid X which I usually run at either 1.3 or 1.45GHz...

Adding features does not necessarily increase functionality -- it just makes the manuals thicker.

Working...