Sunday, August 18th 2013

HWBot No Longer Accepts Record Submissions from Windows 8

HWBot announced that it no longer accepts benchmark record submissions from PCs running Windows 8. It discovered that the way Windows 8 handles real-time clock (RTC), compromises the veracity of benchmark results. HWBot announced that is looking into possible solutions for the problem, and till such a time, it won't accept benchmark result submissions from Windows 8 PCs. This decision could affect leaderboards and records set using Windows 8 benches, and could greatly stunt adoption of the operating system among the professional overclocking community. A statement by HWBot reads;
As the result of weekend-time research, the HWBOT staff has decided to invalidate all benchmark records established with the Windows8 operating system. Due to severe validity problems with the Windows8 real time clock ("RTC"), benchmarks results achieved with Windows8 cannot be trusted. The main problem lies with the RTC being affected when over- or underclocking under the operating system. The operating system uses the RTC as reference clock, and benchmarks use it to reference (benchmark) time.
Add your own comment

81 Comments on HWBot No Longer Accepts Record Submissions from Windows 8

#1
Nabarun
by: AthlonX2
Im sure Microsoft would love to pack Windows with apps,but everytime they do manage to sneak one in they get sued for forcing it upon users.:shadedshu
IE cannot be classified as a free app - it's part of the os and the weakest link in system security.
Posted on Reply
#2
Mathragh
by: Nabarun
IE cannot be classified as a free app - it's part of the os and the weakest link in system security.
Actually, the user is the weakest link in security, just like it is in linux.

Seriously though, this thread is about HWbot's problem with win8, not Linux vs Win8:)
Posted on Reply
#3
Nabarun
by: Mathragh
Actually, the user is the weakest link in security, just like it is in linux.

Seriously though, this thread is about HWbot's problem with win8, not Linux vs Win8:)
I meant software-wise.

It's not hwbot's problem. hwbot just revealed the problem ;)
Posted on Reply
#4
Frick
Fishfaced Nincompoop
by: Nabarun
IE cannot be classified as a free app - it's part of the os and the weakest link in system security.
Naah, not if you keep the software updated, and you can remove the executable, so that IE as a browser is not on your system (at least on Vista/7, don't know how it is in 8). The underlaying parts (like Trident) remains though.
Posted on Reply
#5
Massman
by: Nabarun
I meant software-wise.

It's not hwbot's problem. hwbot just revealed the problem ;)
Just a heads-up for those who are interested in this very specific 1st world problem. From the looks of it, AMD systems might not be affected by this. We'll probably keep the OS blacklisted until we know the full scope of the problem.



Posted on Reply
#6
Solidstate89
by: Nabarun
IE cannot be classified as a free app - it's part of the os and the weakest link in system security.
It's included as part of the OS. Whether you consider that free or "as included" is academic.

Oh, and IE10/11 is one of the most secure browsers out there. The sandbox (or integrity levels as it's called in Windows) is more stringent and secure than the one Chrome uses. Chrome uses the Low_Level integrity level sandbox while IE10 and above uses the App_Locker integrity level sandbox which was what they introduced in Windows 8 to be one level even below Low_Level integrity.

My personal preference for browsers is Firefox, but if I had to pick the most secure browser to use in Windows, it would be IE10 (or 11, when it's released) without a second thought. People like you really need to educate themselves on the massive security improvements in Windows as you're all stuck in the 90's with your assumptions and bias of Windows and especially IE security. Everyone fellates Chrome and its security model when in fact it was IE that beat Chrome in both sandboxing and multi-process security.
Posted on Reply
#7
Ahhzz
by: jateruy
the problem is all about Windows 8
fixed. :pimp:
Posted on Reply
#8
d1nky
I thought this issue was just an win8 dx11 problem with calculating 3d renders.
Posted on Reply
#9
jihadjoe
by: Fourstaff

-> with SSD Win8 boots even faster
More like Win8 doesn't actually shut down properly.

The reason it boots faster is because it hibernates by default, instead of properly shutting down like you asked it to.

My laptop already goes into hibernate by default (usually 10-15 minutes after going into sleep, which happens when I close the lid on it). My desktop is on 24/7. When I go into the shutdown menu on either one, it's usually because I want a proper shutdown and cold boot, not a suspend-to-disk event and a resume shortly afterward.
Posted on Reply
#10
Tatty_One
Senior Moderator
Off topic is an understatement here, please keep on track and cut the insults otherwise my patience will dissolve as quickly as the OP topic did :D
Posted on Reply
#11
TRWOV
For any(every)one not bothering to read the post:


This issue is caused by changing BCLK within Windows

So this is what happens:

1- "X" boots his PC at stock BCLK
2- W8 makes RTC calculations at boot time
3- "X" changes BCLK while in Windows.
4- W8 doesn't sync the RTC with the new BCLK value
5- "X" benches and sends dodgy result to HWBOT

AMD CPUs don't seem to be affected as Massman posted.
Posted on Reply
#12
EarthDog
by: TRWOV
For any(every)one not bothering to read the post:


This issue is caused by changing BCLK within Windows

So this is what happens:

1- "X" boots his PC at stock BCLK
2- W8 makes RTC calculations at boot time
3- "X" changes BCLK while in Windows.
4- W8 doesn't sync the RTC with the new BCLK value
5- "X" benches and sends dodgy result to HWBOT

AMD CPUs don't seem to be affected as Massman posted.
QFT.

Who does this affect here? Like 10 people total that benchmark and post to Hwbot?? Though I suppose now that it was posted you may have some people benching it here in those scores... so, like Hwbot, just do not allow W8 to be used until it gets fixed. :toast:

by: Jorge
Wow, a real big deal over "leaderboard" results". Who frigging cares about leaderboard results?
You would probably care if you were actually participating/competing there. Its cheating if used, plain and simple.
Posted on Reply
#13
unclewebb
RealTemp Author
I wrote a program a while ago called WinTimerTester that compares the high performance timer to the regular timer in Windows. When adjusting the BCLK in Windows, the two timers get out of sync and WinTimerTester should show the percentage difference between the two timers.

This was originally a problem in Windows 7. Linking the high performance timer directly to the BCLK was introduced as a new feature for Windows 7 to conform to ACPI. If you adjust the BCLK in Windows 7 or Windows 8, try the WinTimerTester program and post your results.

Here's some background info.

http://forum.notebookreview.com/asus-gaming-notebook-forum/568525-setfsb-game-timing-problem-g60jx-2.html#post7364805
Posted on Reply
#14
cheesy999
by: jihadjoe
More like Win8 doesn't actually shut down properly.

The reason it boots faster is because it hibernates by default, instead of properly shutting down like you asked it to.

My laptop already goes into hibernate by default (usually 10-15 minutes after going into sleep, which happens when I close the lid on it). My desktop is on 24/7. When I go into the shutdown menu on either one, it's usually because I want a proper shutdown and cold boot, not a suspend-to-disk event and a resume shortly afterward.
Theres a large difference between saving the kernel to disk and saving the memory in it's entirety to disk
Posted on Reply
#15
xvi
So, is this something that should be handled with hardware and double-checked with software?

by: cheesy999
Theres a large difference between saving the kernel to disk and saving the memory in it's entirety to disk
Let it die. If you need encouragement, see Tatty's post above.
Posted on Reply
#16
EarthDog
by: xvi
So, is this something that should be handled with hardware and double-checked with software?
yes nd no as there are still easy ways around it. For example, lets say you lower bclk and run a bench. Clearly, the timer is off and so is the benchmark. So you raise the bclk back and possibly raise the mutli up a notch or two to compensate for the increase in score making it more believable.
Posted on Reply
#17
Hood
by: Mathragh
What does this have to do with,,.. anything here? =D
I think he means that after spending thousands on high end hardware, some have no use for it except to squeeze the highest benchmarks out of it so they can brag to their friends. Not sure whether it's Haswell or Windows 8 that's making these machines into pieces of crap in his estimation...
Posted on Reply
#18
riffraffy
Maybe it would not be a problem if it's accurately wrong , just add or subtract ...or not .
Posted on Reply
#19
Frick
Fishfaced Nincompoop
by: TRWOV
For any(every)one not bothering to read the post:


This issue is caused by changing BCLK within Windows

So this is what happens:

1- "X" boots his PC at stock BCLK
2- W8 makes RTC calculations at boot time
3- "X" changes BCLK while in Windows.
4- W8 doesn't sync the RTC with the new BCLK value
5- "X" benches and sends dodgy result to HWBOT

AMD CPUs don't seem to be affected as Massman posted.
How is it affected from throttling and boost then?
Posted on Reply
#20
Aquinus
Resident Wat-man
by: Frick
How is it affected from throttling and boost then?
It wouldn't because neither touches the bclk. It drops or boosts the multiplier in most cases.
Posted on Reply
#21
naoan
by: remixedcat
I do like the improvements in the UI actually.

  • The file transfer dialog finally makes sense with the speed graph
  • blink blink! but wow why is everything else looks so dead flat and drab now...?
  • The Task manager is improved greatly
  • You mean you haven't used Process Explorer available since XP time?
  • Built-In ISO mounting/burning
  • Great!
  • Hyper-v
  • Great for average joe I guess (not)?
  • Improved explorer UI. The ribbon is so much easier to do a lot of file ops on
  • And take even more spaces, also hide even more icons.
  • Very fast boot
  • That saves me about 5 seconds of my life every few weeks. Based MS.
    Fast boot also known to cause possible damage.
    Posted on Reply
    #22
    Mathragh
    Why are this many people so busy with trying to convince others windows 8 is bad :wtf:

    Seriously, this isnt even a thread about windows 8, but still it feels like people come at it to bash win8 like flies are attracted to shi.. Poo!

    I mean, its not a bad OS to use(seriously), it was actually really cheap(30 dollar/euro when it came out), and technically one of the most advanced OS-es ever made.

    Noone comes to linux or old windows related threads to bash those OS-es because of their worse support or worse user friendliness now do they.



    If you don't own windows 8, and didn't ever bother to properly test/use it for like a month of daily use, then imho you simply don't have a lot of right to say stuff about it.
    So quit posting stuff about windows 8 already!





    On topic: Has anyone heard anything about if this has been found on any platform besides Haswell yet? Looks to me atm that the problem is confined to users combining Haswell overclocking during the running of windows 8.
    Posted on Reply
    #23
    TRWOV
    by: Frick
    How is it affected from throttling and boost then?
    throttling and boost don't change BKCL only the multiplier.
    Posted on Reply
    #24
    Frick
    Fishfaced Nincompoop
    by: Aquinus
    It wouldn't because neither touches the bclk. It drops or boosts the multiplier in most cases.
    by: TRWOV
    throttling and boost don't change BKCL only the multiplier.
    Yeah, obviously. D'oh! :o
    Posted on Reply
    #25
    Nabarun
    by: Mathragh
    Why are this many people so busy with trying to convince others windows 8 is bad :wtf:

    Seriously, this isnt even a thread about windows 8, but still it feels like people come at it to bash win8 like flies are attracted to shi.. Poo!

    I mean, its not a bad OS to use(seriously), it was actually really cheap(30 dollar/euro when it came out), and technically one of the most advanced OS-es ever made.

    Noone comes to linux or old windows related threads to bash those OS-es because of their worse support or worse user friendliness now do they.



    If you don't own windows 8, and didn't ever bother to properly test/use it for like a month of daily use, then imho you simply don't have a lot of right to say stuff about it.
    So quit posting stuff about windows 8 already!





    On topic: Has anyone heard anything about if this has been found on any platform besides Haswell yet? Looks to me atm that the problem is confined to users combining Haswell overclocking during the running of windows 8.
    And we complain so that the vendors hear us and understand that whatever they may think is good, isn't necessarily so. Your opinion matters, so do ours. It's important that opinions are made known, so that it benefits us all in the future.
    Posted on Reply
    Add your own comment