Jump to content
Sign in to follow this  
Devil 07

A bug in Prime95?

Recommended Posts

I was doing some stability testing over the past few weeks and discovered a rather interesting bug in Prime95. It is possible to cheat the display time at the end of a run. I wasn't trying to cheat Prime95 this was discovered by accident.

 

The way I discovered this was purely by accident. While I was tweaking the BIOS I input some settings which required a CMOS clear. After I rebooted I loaded a bank of settings from the CMOS ReLoaded and booted into windows and I started another run of Prime95. About 10-15 minutes into the Prime run I remembered that I hadn't set the date back after the CMOS clear. I changed the date from its default of 1/1/06 to whatever the date was when I did this. At the end of the 8 hour run i hit stop and the display said it had run for 8000+ hours without an error. I was like holy crap what happened? I did a bit of math and figured out that there are roughly 8700 hours in a year.

 

Apparently when I started Prime95 on the CMOS's default date after the CMOS clear, it logged all the hours between the date changes as actual hours run. I ran a few more tests to verify this and everything came back as expected. Anytime you change the date while Prime95 is running you can cheat the display time. I even changed the date back a year and made Prime 95 run negative hours.

 

I ran a few more runs of Prime95 and also used the Task Manager to log CPU time for Prime95. The "time warp" bug did not follow into Task Manager. It's a good thing you can't cheat the CPU time in Task Manager or this could potentially invalidate a number of prime runs. I don't know if anyone else has experienced this interesting bug in Prime95 but I figured I would post it hear since the core of the OCBD is running Prime95.

 

I have a few screen shots from my discoveries I will post here later this evening because I'm at work and they are on my system at home.

 

Cuss and Discuss Time.

Share this post


Link to post
Share on other sites

I'm not quite sure. I was preparing for an OCBD entry when I discovered all this, so I hadn't really tested with those apps. Now that you bring it up it would be interesting to see if it affects them as well. If it does we may need a revision of the first OCBD entry screen shot to include a shot of Windows Task Manager with CPU Process time listed. I'm going to download SP2004/Orthos and test it now. (the boss is in Knoxville he he he)

Share this post


Link to post
Share on other sites

SP2004 is also affected in the end time displayed, but not in the duration window at the top. Uh oh, this doesn't look good.

 

Screen shot of SP2004 with the bug showing. CPU time in the task manager doesn't follow the duration time like it does in Prime95. Interesting.

Share this post


Link to post
Share on other sites

Almost wish you wouldn't have posted this.

But yes, it is a bug, so you could cheat on a Prime run.

Really cheating no one but yourself (and it's awful low) but folks have been caught before.

 

Angry's latest rules for the OCDB clearly spell out SP2004 or Orthos so it shouldn't be an issue.

 

Devil_07 hit the nail on the head and Orthos will show it.

 

ExRoadie also pointed out to me that "Duration/Time change" could be caused if you are "syncing with a time clock" and your PC "syncs" while the program is running.

(Swear to god I wasn't trying anything trick here)

 

ocdb24hrwp0.th.jpg

Share this post


Link to post
Share on other sites
Almost wish you wouldn't have posted this.

But yes, it is a bug, so you could cheat on a Prime run.

Really cheating no one but yourself (and it's awful low) but folks have been caught before.

 

Angry's latest rules for the OCDB clearly spell out SP2004 or Orthos so it shouldn't be an issue.

 

Devil_07 hit the nail on the head and Orthos will show it.

 

ExRoadie also pointed out to me that "Duration/Time change" could be caused if you are "syncing with a time clock" and your PC "syncs" while the program is running.

(Swear to god I wasn't trying anything trick here)

 

ocdb24hrwp0.th.jpg

 

I agree with you 100%, you are only cheating yourself, but I'm a tech so I feel like I need to share info I find good or bad.

 

I didn't realize A_G had changed the requirements for socket 939 OCDB entries. Last I had checked was actually the day DFI-Street went down and Prime95 was the only accepted one at the time. Looks like I will be redoing my screen shoots for my OCDB entry.

Share this post


Link to post
Share on other sites

I am going to put it like this...

 

Circumventing the time/duration is cheating as you all know that...And as SoundX puts it you are the lowest of the low if you do this to make yourself look good...

 

But it's like this..If you are caught posting a entry with any time/duration circumvention...your entry will be deleted and you will get a vacation...period...

 

we will not take lightly to cheaters...it does no one any good...

 

Let this be a reminder...

Share this post


Link to post
Share on other sites
I am going to put it like this...

 

Circumventing the time/duration is cheating as you all know that...And as SoundX puts it you are the lowest of the low if you do this to make yourself look good...

 

But it's like this..If you are caught posting a entry with any time/duration circumvention...your entry will be deleted and you will get a vaction...period...

 

we will not take lightly to cheaters...it does no one any good...

 

Let this be a reminder...

 

Well said.

Share this post


Link to post
Share on other sites

Here is the screen shot that initially freaked me out. Notice it was in the wee hours of the morning,and as I'd been up all night, you can magnify my feeling of dread x10 when I saw the discrepancy in the times.

 

Pic is a bit small I'll try and re-size it.

 

Never mind. I have to pay to upload bigger pics.

 

Anyway, I hope I didn't piss on the proverbial "hornets nest" with this one and cast a shadow of doubt around any of my potential OCDB entries, as I have none yet. I just didn't feel comfortable entering data into a system without first knowing how the system would react to false data.

Share this post


Link to post
Share on other sites

I was certainly glad to see Orthos become required for dual cores. Prime95 is far too easy to cheat with a dual core..

Share this post


Link to post
Share on other sites

Ya hit the nail on the head Cliff,

Any new entries into the OCDB using Prime 95 will be deleted.

 

(The Devil made me do it :))

Share this post


Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Sign in to follow this  

×
×
  • Create New...