Windows 8 has been banned by the world’s top benchmarking and overclocking site

[Idontcare]

All my inlaws have become iPhone fans, and keep making this claim that it outsells any particular brand of Android. But the actual shipment statistics say that Samsung Galaxy S3 alone has outsold all iPhones. And I'm sure the Galaxy S4 will do the same.
iPhone SEEMS more popular, because unlike Android which has high utility value, iOS makes it more of a fashion accessory. And what is the point of a fashion accessory that not everyone sees. So everyone who has the iPhone makes a point to flaunt it as much as possible.

Everyone?

Personally I don't know anyone who flaunts their smartphone, iPhone or otherwise.

The people I know who own an iPhone do so for literally no other reason than that they are comfortable with the OS and the apps they have accumulated over the years.

Your standard legacy captured audience.

Do people really flaunt something as cheap as an iPhone? Weird if true, like flaunting that you go to Taco Bell or something.

 

[VirtualLarry]

Do people really flaunt something as cheap as an iPhone? Weird if true, like flaunting that you go to Taco Bell or something.

Smartphones are fashion accessories. Welcome to 2013.

 

[crashtech]

My social skills are not superb, but it's apparent that what kind of phone you use is a status indicator in some circles.

 

[Mark R]

so how is a rtc implemented? they don't just have a rtc module ( 32768hz xtal + battery + a ic / southbridge integration )? i thought software just reads a few registers which wouldn't be affected by an oc.

The RTC module is a battery backed quartz clock, typically similar to that in a digital wristwatch. It connects via a serial port, or similar method to a system bus, so that the OS/BIOS can read/write the time from it.

OSs have tended only to read the RTC occasionally. Historically, they would then keep track of time using a timer interrupt usually generated by the same RTC (e.g. a 60Hz timer, which is why a lot of old DOS programs could only time to the nearest 1/60 second). The OS could then perform adjustments for time zones, network sync as needed, and write them back to the RTC.

Windows has for many years been able to "skew" it's internal clock, to allow it to stay in sync with a reliable time source; if you are synced to a network time server, windows doesn't just set the time to the server value, it calculates how much the clock is gaining/losing and applies a correction in realtime, so that any drift before the next check should be minimal.

I'm not sure what changes Windows 8 has made. However, it wouldn't surprise me if MS were trying to get away from the old 60 Hz timer system, and instead were trying to push as much of their timing onto high-precision timers. In Win XP, if you used the regular programming API to get the time, you'd get 1/60 second precision; however, a second method was sometimes available to provide access to a high precision timer (originally, a CPU clock cycle counter, but later a chipset provided timer to get around problems of variable CPU clocks).

If you use MS's .NET programming framework, the default timing method is automatically via a high-precision timer, but it will automatically fallback to 1/60 sec timing if a higfh-preicisin hardware timer isn't available.

I would not be surprised if MS in Windows 8 had made high-performance timers the default hardware manager for all time-keeping requests.

The issue with high precision timers, is that they tend to be chipset or CPU based. In Intel core i platforms, the chipset timer is drive from the single master clock generator for all system buses. During OS boot, the system will read the timer frequency from the chipset; but the OS does not expect timer frequency to change in operation; a "live" bus clock reconfiguration can therefore lead to timekeeping errors.

 

[SunnyD]

Easy way to achieve higher score without BCLK. This will work on AMD, Win7, WinXP, and Win2K as well. Sorry, no Win98. Officially not supported on NT4, but I'm guessing that would probably work. I guess HWBot will need to revert to Windows 98/ME scores

int main(int argc, char* argv[]) {
  double cheat_ratio = 1.05; // 5% cheat
  DWORD time_adjustment;
  DWORD time_increment;
  BOOL disabled;
  GetSystemTimeAdjustment(&time_adjustment, &time_increment, &disabled);
  time_adjustment -= time_increment / cheat_ratio;
  SetSystemTimeAdjustment(time_adjustment, FALSE);
  return 0;
}

Note: The Windows Time service will undo (and in fact temporarily overcompensate for) this setting the next time it checks with your configured NTP server. So you will want to disable the "Internet Time" feature for this.

That doesn't affect QueryPerformanceCounter() or QueryPerformanceFrequency(), which is what any legitimate benchmark is going to use to accurately* measure frame rates and system cycle times as it should hook either HPET or the CPU base clock generator directly (via HAL) depending on the platform as an independent stable* clock source to avoid the exact situations you describe.

* Asterisk to denote that it's obvious the HPET is not exactly stable or accurate when used in conjunction with an improperly coded HAL.

 

[glugglug]

That doesn't affect QueryPerformanceCounter() or QueryPerformanceFrequency(), which is what any legitimate benchmark is going to use to accurately* measure frame rates and system cycle times as it should hook either HPET or the CPU base clock generator directly (via HAL) depending on the platform as an independent stable* clock source to avoid the exact situations you describe.

* Asterisk to denote that it's obvious the HPET is not exactly stable or accurate when used in conjunction with an improperly coded HAL.

QueryPerformanceCounter and QueryPerformanceFrequency used to be tied to CPU cycle counter (RDTSC instruction) by default (up until XP SP3), which we all know varies at runtime nowadays, so not really useful for this. In fact, when multicore CPUs were first introduced, timing in that way would sometimes result in negative values, because the counter on the different cores could get out of sync, so when your app got resumed on a different core it could look like going back in time, causing all sorts of wierd issues. Because of this, its normally only used for multimedia and games that need the finer granularity, not benchmarks.

 

[Idontcare]

Smartphones are fashion accessories. Welcome to 2013.

My social skills are not superb, but it's apparent that what kind of phone you use is a status indicator in some circles.

Interesting. This must be true for demographic circles of which I am not a member.

 

[Arkaign]

Interesting. This must be true for demographic circles of which I am not a member.

I try to avoid those folks myself

That brings me to the irritating trend of overpriced phone plans that subsidize phones that people can't really afford, but somehow they think they're getting a 'deal' with a $100+/mo smartphone plan. Insanity.

 

[Chipfiref]

im fairly sure most of us who did it as a pure sport all had technet accounts which allowed us to install windows 7 on a machine which is based at home and never see's business legally.



And yes another fail on windows 8.
Which is why im having a very hard time going to windows 8, or even migrating my servers on server2012.

Ah yes! I am not a big overclocker, but I love me some Technet subscription - $249 per year for several development licenses to every version of every MS product there is ( except Visual Studio ) I will renew it for one last year since I want to do a lot of experimenting, and I want Office on most of my machines, and I have a few servers at my house!

I am not sure yet how much I like Win 8 but my friend loves 8.1, it is a free preview and will be free in October. He got the 20" Sony Tap 20 tablet/all in one and loves 8.1 on that. He is a big Mac user.

So if Win 8.1 fixes the bug or whatever it is there will be a huge online celebration honoring Windows 8.1 for overclocking, right?

 

[BallaTheFeared]

I like 8.1, had to revert back to 8 for the stttttttttutter driver though.

I don't like it more than I like fluid frames >.<


Disclaimer: I purchased Start8 probably when Windows 8 was still in beta.

 

[csbin]

Update 20/08 – Windows 8.1 Affected, AMD Problem-Free(?)
http://hwbot.org/news/9824_breaking_windows_8_benchmark_results_no_longer_accepted_at_hwbot/



We have continued our testing on more platforms. Windows 8.1 is, just like its older brother, affected by at-runtime base clock frequency adjustments. Intel platforms back to LGA775 are affected. From the looks of it, AMD systems are unaffected. In fact, the clock drifting on Intel based systems seem to resolve itself when switching from an Intel based system, to AMD, and back to Intel. Later today Christian Ney, Head Moderator, will publish an update to the rules at HWBOT concerning the usage of Windows 8. Stay tuned!
The following combinations of platforms have been tested

 

[CHOPPER GOD]

good to know

 

[Jaydip]

Windows 8 RTC Bug Isolated, Fixed
http://www.techpowerup.com/189607/windows-8-rtc-bug-isolated-fixed.html