This kind of situation can happen a lot; particularly when running Windows XP I’ve noted. It can even get to a point where the entire machine seems to freeze-up, and nothing appears to respond: Not the display, nor any functions. – Yet the processor-monitoring display appears to show that the processor-load never goes above a couple or three percent while this is going on.

There may be a number of reasons why this is happening, and several things that are going on here: -

Screen-requests

Windows XP queues screen requests and executes them one at a time: Therefore if the queue becomes too long you may find that the display may become sluggish at times until the queue of requests have all been processed. During this time the processor’s waiting for the queue to clear and is idling and/or getting on with any other minor processing tasks.

Disk-access

The major reason why the system appears to jam-up is due to disk-access, possibly combined with the above screen-request queuing. You may notice that at the time your computer becomes temporarily unresponsive the hard-disk activity appears to be constant. – That means that queued data is being written to, and other data is being read from, the hard-disk. While this is happening the processor is just twiddling it’s thumbs while waiting.

Disk-access is a well-known bottleneck in computers. While the processor can interact with RAM whilst only having to wait a couple or three cycles for a response, it’s a totally different ball-game when it comes to accessing data on the hard-disk: The processor requests the information and waits. The request quickly gets to the SATA controller, and either joins the queue or is passed straight to the hard-disk. This takes a minute fraction of a second thusfar, as it’s all electronic and happening extremely fast.

When the hard-drive receives the instructions, however, the slow part begins: First the hard-drive’s electronics process the request and instruct the read/write arms to retrieve X data from Y location on Z platter. The read/write arms then have to physically move to do so, and when they find the data they may have to read it a number of times to ensure that the data they send back to the electronics is a 100% perfect copy of what’s stored on the magnetic disk.

In the meantime the processor has been waiting several thousand cycles or more already, and the reply’s not even been despatched yet. After another hundred or so cycles the processor receives the long-awaited reply. – Meanwhile other operations dependant upon the outcome of that request are backing up in the processor queue. – Many of these could be pieces of data that when processed will result in another screen request, upon which the graphics-processor on the graphics card is waiting in order to complete a screen-draw while it holds other pending requests in a queue.

At times this situation can escalate to a point where the request-queue for the hard-disk, as well as every other queue, becomes so long that the entire machine has to wait a while for the multiple-queues to clear and their respective caches to empty, and thus the entire machine becomes temporarily unresponsive for a while. – All because hard-disk access times are so slow compared to other request execution times.

Is there a way to avoid these excessive queues building up?

Not completely, no: But the more RAM that’s installed and the less the number of processes running at the same time, the less frequently this situation will occur. Why?

Each process requires a separate process-thread in the processor. That in itself isn’t a problem with today’s processors, especially multi-cored processors with hyperthreading. – But the processor still has to issue various requests of other components related to each individual process; many of which will be for access to data on the hard-disk, resulting in a queue of a length dependant upon the number of requests for hard-disk data issued. – Therefore the less threads running at once, the smaller the queue.

RAM and 64-bit

Also the more RAM installed the more can be used by the processor to store data in. The more data stored in RAM rather than on disk the quicker the processor can access that data when it needs to refer to it… And at this point I return to one of my favourite subjects: The advantage of a 64-bit operating system over a 32-bit equivalent.

A 32-bit operating system can only see an absolute maximum of 4 gigabytes of RAM. When the system caches and shared-graphics-memory if any have all been taken into account that usually leaves about 3.5MBs – Which isn’t always enough, depending upon how many processes are being run. Eventually the available RAM is all used up and the processor relies upon the paging file on the hard-drive for storage. – The consequence of this is that over time the queues build up and – see above.

A 64-bit operating-system, however, can theoretically see over 3 exabytes of RAM: About a small football-field’s area of 1GB RAM sticks. In reality a 64-bit Windows operating system can see well over 10GBs RAM; and the most RAM that can commonly be installed on today’s motherboards is usually 16GBs anyway. Is that enough RAM to prevent massive queues building up. Normally, yes. (I have a 64-bit Windows 7 box running 8GBs of DDR2 800MHz RAM and it’s never “frozen” even once in 8 months, no matter how many process are being run on it.(Yes I’ve had the beta 64-bit, the RC 64-bit, and now the RTM 64-bit of Windows 7 on it.))

…In short then; the more RAM the better, in which case a 64-bit operating system is going to be superior to a 32-bit operating system in that regard.

Solid-State disks

Also, solid-state disks help one heck of a lot too; as the disk-access-time for SSDs is much quicker than for their spinning-platter equivalents – because there are no moving-parts in SSDs, and therefore the seek-time is vastly reduced.

There is one drawback or more, though, with SSDs at the time of writing: Firstly they cost a small fortune in comparison to spinning-platter disks. The price is coming down, it’s already more than halved, but there is a long way to go yet. Secondly they aren’t available at the same storage capacities as spinning-platter drives: the largest commercially-widely-available SSD drive at time of writing is 320GB, although there are 500GB and larger SSD prototypes in use in secret in the military and in big industry to a lesser extent. (I can’t confirm that as a fact for obvious reasons; but you can believe it or not as you like.)

Do I have to change my operating system to a 64-bit (Windows 7) OS, buy an SSD, and install more RAM to avoid my comp “freezing”?

No; but those three things would be a complete solution.

If you’re running 32-bit XP I would advise an upgrade to 64-bit Windows 7, provided that your machine is 64-bit capable. If you’re running 32-bit XP and intend to continue doing so, then if you have with less than a single gigabyte of RAM installed I’d advise you to add a gigabyte to whatever you currently have installed. If you want to buy and can afford to buy an SSD than go for it if you like; but upgrading to a 64-bit operating system with 4 gigabytes of RAM or more should solve most if not all of your worries. (Personally I’m going to wait until SSD storage becomes much bigger in capacity and also until the price is less than twice the price of standard HDDs before I invest in that technology.)

Do I have to stick to Windows?

Not at all: I do because I’m me, but you can use any operating system you want to use. I don’t advise trying to run Mac OSX on a PC though; unless your name’s Psystar. I’ve tried most operating systems; and I’ve found that Windows works best for me. You’re not me though, and maybe Linux or something might be preferable to you for your needs. – Try a dual-boot of Ubuntu and Windows. Ubuntu’ s free after all – so you have nothing to lose apart from a bit of time.

… So if your computer (running Windows) becomes unresponsive temporarily at some point, you’ll now know why it’s doing so. – Does anyone have anything to add to that? If so please use the comment box below. – It seems a shame to waste it.