PDA

View Full Version : What causes screen freeze ?


Donn
11-10-2004, 11:08 PM
What are the mechaniscs of screen freeze, cursor freezes, not freezes, whatever with it?

I don't mean 'what cause scren freeze?' like adware or spyware can 'cause' screen freeze. I mean what actually happens that screen freeze is ocurring?Perhaps I should be asking--are there different kinds of screen freeze? Is the screen freze that spyware causes different from the screen freeze caused by, what?. . . some internal failure(s) within the system?

saphalline
11-11-2004, 04:01 AM
I think I understand your question. Yes, there are different kinds of screen freezes. Lock-ups are another name, but the connotation can be different. Screen freezes are of two types: OS or BIOS/firmware. These can be broken down even more into software-related and hardware-related. Yes, the lines are fuzzy.

As an example, let's say a virus causes your system to freeze. What happened? Well, it most likely caused a software-related freeze in your OS. So, the virus code actually halted all processing in your system. It either got your system stuck in a wait state or invoked a closed infinite loop. Depending on what happened, you can sometimes get out of this using the three-fingered salute (Ctrl-Alt-Del).

Another example: your optical drive suddenly died. You put a CD into it, not knowing this. The CD drive is able to tell Windows that it has a CD, so Windows starts auto-run. Your poor CD drive never gets back to Windows. Bam! Freeze. What happened? Unfortunately, auto-running a CD is one of those tasks in Windows that takes a high priority, so doing a hard restart may be the only way out of this.

Another example: you're playing a game on your computer, but suddenly, you get shot, and the screen freezes! What happened? Well, games use DirectX or OpenGL or another API and are given low-level control of most hardware, including the mouse, keyboard, video card, and others. When a game performs an API error, this strikes at the kernel of the OS and it is not able to recover. This is essentially a software-related BIOS issue but often can be fixed with the three fingered salute (worse cases need a hard reboot).

Another example: your PSU is starting to die. As a result, it's outputting some voltages that are really far off. This causes your hard drive to write incorrect data while you're installing a Windows update. A few 0's become 1's and vise versa. Your system freezes. What happened? Well, I told you what happened, but in the same situation you probably wouldn't know! :p This is a case of a hardware-related OS freeze. Windows was working fine, but because your hard drive was given incorrect voltages by a dying PSU, the data got messed up. So Windows tried to run off bad data, and freezing occured.

There are many, many, many more possibilities and stories to tell, but this at least is a small sampling of problems that can occur. In all of these examples, the system froze, but that's only a result. The real causes are different in all cases. This is exactly why when people come to us with problems, it's best to have as much info as possible in order to determine what went wrong. Simply hearing "my system froze" is not enough to fix it!

Paleo Pete
11-11-2004, 10:24 AM
Good replies so far, and all true. Plus a bunch more causes...

Probably the most common I've seen are memory problems, flaky drivers and IRQ conflicts. When the mouse moves jerky, as I think you tried to describe, that's usually something running in the background that maxes out the CPU and hits the swap file pretty hard at the same time. Move your mouse around during an antivirus system scan, you'll see a mild example, but easily recoverable, just stop moving the mouse.

Probably the strangest one I've seen was on my own machine, 5-6 years ago. I set up Task Scheduler to run scandisk and defrag at 3am once a week. Didn't take into account the fact that it wouldn't be able to also turn off everything accessing the hard drive, like AV and good old Find Fast, (didn't know about that yet) so it went into an endless loop of stopping-restarting, then continued to try to run in the background even after a reboot. The program window wasn't even open, but defrag still tried to run...really strange, but after I opened and ran defrag in Safe Mode so it could finish the problem stopped, and I couldn't get task scheduler turned off fast enough...the mouse not only moved really jerky, it blinked...like a christmas light, on and off, about one second on, one second off...

More common things, bad memory, corrupt swap file, bad drivers, bad software installation (quite often happens because AV software was not turned off during installation), out of hard drive space, corrupt or missing system files, virus and spyware of course, and the worst of all types to troubleshoot...several things at once combining to cause system lockups.

Example: My last use of McAffee AV didn't like win98SE, computer developed bad RAM, and video drivers for Matrox Mystique didn't agree with win98SE either, all at the same time. System lockups, invalid page faults, illegal operation errors, terrible performance...took me several months trying to figure it out, since 3 things were causing both similar and different symptoms. Some of you may remember that one, I had only been moderator here 6-8 months to a year when I "upgraded" from a P-233MMX to K6-2/450 and that finally solved it. Different memory and video card, decided to try Norton. Never would have found out but I set up the old board to check out some parts, the old RAM turned out to be flaky. Fixed that and swapped for the Matrox video card, trouble was right back. Didn't care much for Norton so I went back to McAffee on main machine, complete windows reinstall, invalid page faults started again soon as I installed McAffee...2 months later they still hadn't responded to my emails so I threw away their CD, formatted and reinstalled, used Norton until they started to try and get a yearly "subscription" for previously free definition files. Grabbed AVG, the rest is history...I'm not about to pay them per year to use their software after already paying for the progrem. That's about like Ford trying to tell me I have to pay them a dollar every time I fill the gas tank...

But the constant system lockups, illegal operation errors and invalid page faults were finally a thing of the past...and mjc and I must have gone half nutzoid for 6 months trying to figure it out on IRC...

Donn
11-12-2004, 07:06 PM
Thanks for that infor guys, but, ah, it's not exactly what I was looking for, really. What I mean is what actually physically happens that the cursor is tuck, or if it moves nothing responds to it...screen totally frozen. . .what I mean is what is the physical thing that has happened-- or is that even explainable.

For instance...the spark plugs in you car are blacking out. Physically that could be. . .over rich mixture, plug is too coo of a pspark for the engine, radiator running hot. . .see what I mean?

What is physically happening in the box that the screen no longer responds? Is it a power loss to the screen or the software that runs it? Something like that? Saphalline, your responses were closer to what I was looking for. It's just a moon walk, but I'm thinking that if there is a symptom liek screen freeze, could it be that hard to write a program that reverse engineers it, and tells you were the knot is, and undoes it (when it can be undone) or just identifies it for you so you know where to start to undo it. . .?

Now, right, I know NOTHING about writing programs and engineering power systms to form up screen phenomena, but, I'm just wondering . . . .

gwallen4
11-12-2004, 08:10 PM
When the computer freezes, the CPU has either halted operation altogether or is stuck in an endless loop. Most freezes are caused by hardware problems.

Mick_D
11-12-2004, 10:15 PM
Well Donn I'll take a shot and hope I hit what you're looking for.

In very basic terms the computer operates on a very rigid and sequential set of commands or instructions. When you encounter a screen freeze the computer, through either software or hardware reasons, cannot execute the next step in it's logical sequence, and does not know why or how to bypass the problem, and the process running stops in mid sequence. The result is a screen freeze, as you call it.

This could be a power , heat , memory, or any other type of failing hardware issue. It can also be a software related issue that a file is missing, or corrupt, or other software conflicts exist, and a meriad of other reasons. But basically a computer follows a one way path to it's destination. If somethimg gets in the path along the way it just stops. The computer can't think for itself ....yet :) and it doesn't know where to go next and it stops.

To address the second part of your question; yes, software does address where the "knot" is as you put it. That is when you get error messages or error codes telling you why the program cannot execute. Some programs may have built in error checks and recovery routines built in so the user never knows there is a problem. But there are so many causes of system lock ups, it unrealistic to think a recovery routine can be written for every one of them.

Donn
11-12-2004, 10:35 PM
Mick D said:

"But basically a computer follows a one way path to it's destination. If somethimg gets in the path along the way it just stops. The computer can't think for itself ....yet and it doesn't know where to go next and it stops."


So the error boxes are somewhat of an "intervention program" that tells it to stop at the last thing it "knew" what it was doing before it does something to make itself into a paper weight. . .? So the next step is self-diagnosis and planned out alternate routes, or is that just too complicated to contemplate. . .? Edit--So the next step in computer 'evolution'--would that be something like a parallel or alternate hard drive that would take over on signal and pick up where the first one left off, while the diagnostic figured out what was wrong, ran the fix (turns on spybot, AV, software patch. . .) to get the program back to normal or alert the user that it is a hardware problem the diagnostic cannot handle.

Now, bear with me, I've been watching Flash Gordon moives tonight, so I'm thinking kind of like there is a 'magic tool' for everything . . . . :)

Mick_D
11-12-2004, 10:48 PM
So the error boxes are somewhat of an "intervention program" that tells it to stop at the last thing it "knew" what it was doing before it does something to make itself into a paper weight. . .? So the next step is self-diagnosis and planned out alternate routes, or is that just too complicated to contemplate. . .?[/QUOTE]


What do you mean by "self diagnosis and planned out alternate routes"? Do you expect programmed diagnostics and automatic recovery for every software problem? Of course hardware problems cannot be solved unless the hardware problem is fixed, so error reporting is about as best that can be expected.

Donn
11-13-2004, 11:16 AM
I don't know if I expect it, but I am wondering if that is the next stage--self diagnosing, self repairing. . . except, of course, for hardware... :eek:

saphalline
11-14-2004, 06:16 PM
That indeed could be the next "evolution" in computing, but as it stands right now, there are just waaay too many problems. In a Windows-based evironment, there are so many programs available that they can't all keep up with eachother. That's why some problems exist when using a certain AV program with a certain CD burning program (for instance) and so on. There's simply no way to effectively test all programs with eachother in a relatively short period of time. Add to that the complications that arise with Windows updates and file corruption that occurs in software (for instance dropped packets over the internet) and the problem quickly goes out of control. And let's face it, there are bugs in every program! Bugs that do bad things like divide by zero (no real way to recover from that one).

Now, WindowsXP is smart, smarter than anything we've seen from MS, but it's not perfect. Most of the time, it can recover from a program crash, even a game, but sometimes, the OS itself freezes. When the OS freezes, how are you supposed to recover? The only way to do that would be to have a completely separate system (separate OS and hardware) that would kick in and figure out what went wrong. Maybe the future of desktops is to have one system to use and another mini-system to fix freezes and crashes, but right now that's just a pipe dream.

Hardware is certainly a problem we're all familiar with, but you should really do some software research in regards to this. The way computers work right now, with their binary system and sequential execution style, software errors can halt all operation. And most of the time, the system can't recover. That's why it would take another independent system to get the first one out of trouble.