mersenneforum.org  

Go Back   mersenneforum.org > Great Internet Mersenne Prime Search > Hardware

Reply
 
Thread Tools
Old 2013-08-19, 21:03   #1
M29
 
M29's Avatar
 
Dec 2003

16410 Posts
Default Windows 8 banned by world’s top benchmarking and overclocking site

Extreme Tech has this article about the accuracy of Windows 8's Real Time Clock.

The subject is beyond my competence, but it doesn't sound like anything nefarious on Microsoft's part. Nevertheless....
M29 is offline   Reply With Quote
Old 2013-08-20, 03:10   #2
Rodrigo
 
Rodrigo's Avatar
 
Jun 2010
Pennsylvania

92710 Posts
Default

Seems to me that the bottom line is that benchmarks performed on Windows 8 systems (at least those where over- or underclocking was done) may well be inaccurate and thus invalid. This could conceivably have an impact on our own benchmark database (?).

This, because of Microsoft's inane idea to devise an OS that tries to be all things to all computing devices.

Rodrigo

Last fiddled with by Rodrigo on 2013-08-20 at 03:11 Reason: additional comment
Rodrigo is offline   Reply With Quote
Old 2013-08-20, 03:37   #3
TheMawn
 
TheMawn's Avatar
 
May 2013
East. Always East.

11·157 Posts
Default

Quote:
Originally Posted by Rodrigo View Post
This, because of Microsoft's inane idea to devise an OS that tries to be all things to all computing devices.
Don't worry. Something that tries to be everything can work. The Xbox One for examp - oh wait. The PS4 for exa - oh wait.



TheMawn is offline   Reply With Quote
Old 2013-08-20, 03:49   #4
TheMawn
 
TheMawn's Avatar
 
May 2013
East. Always East.

11×157 Posts
Default

For those who aren't too sure what this is all about (actually this is a HUGE effing deal)...

A processor's "speed" or frequency is determined by two things. A base frequency and a multiplier. In Sandy Bridge and Ivy Bridge, the base frequency is 100MHz and anything other than that can cause instabilities in all sorts of places. I'll get into that later. The multiplier "multiplies" the base frequency into the CPU frequency. For example, mine is 46, so I get 46 x 100 MHz = 4.6 GHz.

A little while ago, the base clock was easy to modify. In fact, this was the golden age of overclocking, because your 1.07 GHz processor could have been 8 x 133.33 MHz, but by overclocking the base frequency to something like 200 MHz, you could get a 1.6 GHz processor, and that was quite feasible. It might have even been 250 MHz...

In Intel's second and third generation of core i processors (I won't pretend to know much about Haswell) the base clock was to remain fixed at 100 MHz. This was super important because the base clock became integral to everything in the computer. By messing around with it (some motherboards couldn't handle 104 MHz), you would get hard drive malfunctions and GPU failures because the SATA and PCI-E interfaces were controlled by the base clock. This was part of Intel's strategy to centralize the base clock to the CPU instead of having every component using its own. Good idea? Maybe. Microsoft seems to have taken a page out of their book.

Now, the Windows Clock is basing itself off the base frequency, it would seem. It only appears to be a problem if the base frequency is changed via software in the middle of a Windows session. You can set the base clock to 120 MHz and the computer will know to add one second every 120,000,000 ticks of the base frequency. However, by changing the frequency down to 110 MHz in the middle of a Windows session, the clock still thinks it needs to add one second every 120,000,000 ticks, even though in reality it needs to add one every 110,000,000. This results in a 9.2% slow-down in the windows clock.

Last fiddled with by TheMawn on 2013-08-20 at 03:50
TheMawn is offline   Reply With Quote
Old 2013-08-20, 03:51   #5
Rodrigo
 
Rodrigo's Avatar
 
Jun 2010
Pennsylvania

92710 Posts
Default

Quote:
Originally Posted by TheMawn View Post
Don't worry. Something that tries to be everything can work. The Xbox One for examp - oh wait. The PS4 for exa - oh wait.





Casual readers: Note the inclusion of (and effect on) Prime95 in the table below the video in that article.

Rodrigo

Last fiddled with by Rodrigo on 2013-08-20 at 03:51
Rodrigo is offline   Reply With Quote
Old 2013-08-20, 04:50   #6
Primeinator
 
Primeinator's Avatar
 
"Kyle"
Feb 2005
Somewhere near M50..sshh!

2×3×149 Posts
Default

Quote:
Originally Posted by Rodrigo View Post


Casual readers: Note the inclusion of (and effect on) Prime95 in the table below the video in that article.

Rodrigo
So the 1.5% is not much...but noticeable for really big exponents on fast processors or for smaller exponents on slower processors? What would cause this malfunction in the clock? Would a restart fix the problem if you realize your iteration time rises unexplained?
Primeinator is offline   Reply With Quote
Old 2013-08-20, 05:01   #7
kracker
ἀβουλία
 
kracker's Avatar
 
"Mr. Meeseeks"
Jan 2012
California, USA

41658 Posts
Default

Quote:
Originally Posted by Rodrigo View Post


Casual readers: Note the inclusion of (and effect on) Prime95 in the table below the video in that article.

Rodrigo
Tehehee... We're famous now!(NOT)
kracker is offline   Reply With Quote
Old 2013-08-20, 07:19   #8
LaurV
Romulan Interpreter
 
LaurV's Avatar
 
Jun 2011
Thailand

2·3·31·47 Posts
Default

Quote:
Originally Posted by TheMawn View Post
For those who ... <snip>
Good explanation and summary.

edit: the part with "100MHz" is called "standardization" and it is a good move, as it allows other small factories, like the one I work for, to make hardware that will work in any computer, and not "this oscilloscope card only work on a motherboard which has 173.14MHz FSB clock". This eliminates the necessity of user intervention, remember the "jumpers" you had to set few years ago, for example for ethernet cards or sound cards, before mounting them in the computer...

Last fiddled with by LaurV on 2013-08-20 at 07:27
LaurV is offline   Reply With Quote
Old 2013-08-20, 11:57   #9
jasonp
Tribal Bullet
 
jasonp's Avatar
 
Oct 2004

DC916 Posts
Default

It's hilarious that SuperPi still gets top billing for PC benchmarks, almost 25 years after it was written. Maybe they can benchmark office suites using Wordperfect or something
jasonp is offline   Reply With Quote
Old 2013-08-20, 15:08   #10
kracker
ἀβουλία
 
kracker's Avatar
 
"Mr. Meeseeks"
Jan 2012
California, USA

5×433 Posts
Default

Quote:
Originally Posted by jasonp View Post
It's hilarious that SuperPi still gets top billing for PC benchmarks, almost 25 years after it was written. Maybe they can benchmark office suites using Wordperfect or something
Maybe ms calc would be a good benchmark now..

y-cruncher is better and faster... AND multi-threaded
kracker is offline   Reply With Quote
Old 2013-08-20, 20:05   #11
kladner
 
kladner's Avatar
 
"Kieren"
Jul 2011
In My Own Galaxy!

3×5×23×29 Posts
Default

Quote:
Originally Posted by kracker View Post
Maybe ms calc would be a good benchmark now..

y-cruncher is better and faster... AND multi-threaded
I coulda sworn I posted this yesterday in response to jasonp-

How about Visicalc?
Attached Thumbnails
Click image for larger version

Name:	Visicalc.png
Views:	103
Size:	2.9 KB
ID:	10141  
kladner is offline   Reply With Quote
Reply

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
WHY out of the entire university did only Davieddy get banned?! Stargate38 Forum Feedback 61 2014-07-08 18:54
World’s first water cooled smartphone from NEC kladner Hardware 3 2013-05-20 23:38
GMP 5.0.1 vs GMP 4.1.4 benchmarking unconnected GMP-ECM 5 2011-04-03 16:16
Banned user on online users list? 10metreh Forum Feedback 9 2009-03-22 12:21
Not banned anymore ET_ Forum Feedback 3 2008-09-13 16:46

All times are UTC. The time now is 20:02.

Sun Sep 27 20:02:00 UTC 2020 up 17 days, 17:12, 0 users, load averages: 1.55, 1.51, 1.43

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2020, Jelsoft Enterprises Ltd.

This forum has received and complied with 0 (zero) government requests for information.

Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.2 or any later version published by the Free Software Foundation.
A copy of the license is included in the FAQ.