Setting up partitions

[EricB]

Look, I don't need to read pages out of some 'tweaker' book. I have written virtual memory/paging algorithms for classes many times. I already know the deal with it. I know all about paging, page faults, and all the schemes that various OS's use to decide which pages get paged out, blah blah blah. I read your pages there, and its not telling me anything I don't already know.

The compelling reason to always have a page file is that you just *might* need to exceed your physical memory size, and therefore you have that room. There is no reason that you are absolutely required to do so, its just a good idea. The point was, whatever the case, he would have spent his money MUCH better by just buying more RAM, rather than monkeying around with buying another hard drive just to put the page file on. Talk about all time biggest wastes of time and money just to gain a few percent performance increase.

You still haven't answered my main point. Why tell this guy to get another hard drive just for swap space when he would be better served just getting more RAM and using a smaller pagefile? And for about the same price, since about the cheapest commercially available hard drives you find in stores right now run around $80. I can get a gig of cheapo PC3200 for that.

well, because the last time I check you can get a 10gig for like 10 buck or less at any used computer store. they usually will give it to you for free, since they don't have much use for it.
the prices of one gig of ram just got to less than $100 in the past month or so. they were 200 previously. you keep missing points

one more time

It is not that serious.

I got better things to do than argue a sensless point. I got 2 gig of ram on my computer (which I paid $180 for last month. the month before I paid $190 for just 1 gig of ram the cheaper ram was better too. samsung ram). guess what? I have a 2 gig page file at the beginning of my 160 gig sata drive. you can talk until you are blue and you still haven't said anything to make me change what I am doing or my opinion.

Because I know for a fact that it works better like that

 

[TheHeadFL]

You keep telling me that its not that serious, then keep going on and on to try to prove me wrong. Dude, whatever floats your boat.

Just don't act like you speak for the experts next time, they can speak for themselves. You can know for a fact that it works that better that way, I simply don't care. My point of contention was your (in my opinion) outrageous assumption that similar performance to your setup would be achieved with a junker old hard drive. I couldn't disagree more. But hey, I'm new to this forum so I'll just bow out of this contest right now.

 

[EricB]

You keep telling me that its not that serious, then keep going on and on to try to prove me wrong. Dude, whatever floats your boat.

Just don't act like you speak for the experts next time, they can speak for themselves. You can know for a fact that it works that better that way, I simply don't care. My point of contention was your (in my opinion) outrageous assumption that similar performance to your setup would be achieved with a junker old hard drive. I couldn't disagree more. But hey, I'm new to this forum so I'll just bow out of this contest right now.

Do your chest feel better now? It seem like some machoism is going on.

.

Originally posted by TheHeadFL
Look, I don't need to read pages out of some 'tweaker' book. I have written virtual memory/paging algorithms for classes many times. I already know the deal with it.

I suspect that this contributed to the problem. you think that since you taught it, you must know it all and nobody's else input seemed to count


you know my way works. at one point, you admit as much. your point is kinda right too. but your answer (adding more ram) had nothing to do with his question. it was like putting a band aid on a bruise.

I was simply telling him that he could improve his performance by adding the swap file to the first part of the next hard drive would be a better solution in his setup, because despite what you think, the way window's implemented the page file was wrong

 

[EricB]

in a simple world we should be able to partition the hard drive so the swap file is 1st in line on a hard drive then we can install an OS. however, Window's like the rest of the world is not simple

 

[TheHeadFL]

How many different ways can I say that I never said that putting the swap file at the beginning of a drive wouldn't help? I think I've said that plenty of times. My problem was with your acting like it was so incredibly beneficial that it would offset the performance hit from using an old dinosaur drive for the job.

Look, you telling me to go read some JPGs copied out of a tweaker book *is* an insult to my intelligence, knowledge, and ability, whether or not you see it. I didn't learn about virtual memory from reading a book on "How Windows XP works". I'm not a 'noob' to this stuff. I am fully trained and formally educated in Computer Science, and that includes having written virtual memory implementations before for simple OSs.

You're trying to act like I don't know what I'm talking about because I don't agree with what you think those pages are saying. I read those pages and they don't contradict anything I've said.

How about I be more explicit about my point:
Yes, in a perfect scenario, putting the swap file at the beginning of the disc would help decrease virtual memory performance losses due to the increased seek time neccesary when the file is fragmented or at other locations on the drive.

That being said, the PURPOSE of accomplishing this is to decrease latency for virtual memory page faulting operations. There is some performance to be gained here, but probably only a few percent.

Now, consider the fact that you suggest he might use an old hard drive for the job, which was my one and only original point of contention. Old hard drives suffer from having higher latency and from having less cache, which also tends to contribute to higher latency since paging operations have an inherent repetitiveness to them.

So, you are trying to solve the problem of high latency by introducing a piece of old hardware with its own latency problems? How does this make any sense? Think about it.

The only way it makes sense to use a separate dedicated hard drive is if it is of equal speed and cache to your original drive. In this case, you have the best of both worlds. You are solving the problem by introducing lower latency solutions all around.

Using an old hard drive for this job is replacing one latency problem with another.

I don't know how many other ways I can say this.

 

[EricB]

Using an old hard drive for this job is replacing one latency problem with another.

I don't know how many other ways I can say this.

you seem to be intelligent, but you keep missing points. pay partitcular attention to the first and last sentence.

wow....

it wasn't that serious

we weren't talking about a 10,000 rpm drive. a guy asked a question. I answered. somehow you missed the part about windows having to find the swap file within windows will slow it up. but who cares really?

I know that maxtor use to have an 7 or 10 gig 7200rpm drive, because I use to have it.

Let me rephrase the answer. dario03, go get the cheapest 7,200 rpm drive that you can find and use it.