Computer help forum: Using usb flash drive as extra ram in XP

Quite a few folks showing how to do this including, as I mentioned, executive editor at cnet - http://cnettv.cnet.com/8301-13415_53-10121584-11.html. Have you experienced these wearing out? A few posts elsewhere here that they last forever, can even make it through a washing machine, etc.
This would undoubtedly be cheaper, faster and easier. Plus the original configuration could be retained if needed later for numerous operations.

That's not the same thing as ReadyBoost, and it's far from some kind of "insider secret". People have been using tricks like this since before XP ever came along.

Put simply, it's not really going to improve things all that much. Even if your average flash drive is faster than a hard drive (assuming no other USB devices are active at the time) it's by a fairly insignificant amount compared to RAM. RAM is many THOUSANDS of time faster, while a flash drive might be a few times faster.

There is simply NO substitute for additional RAM, period. You can waste time, effort, and money on flash drives, or you can just add more RAM and get a benefit far greater for only a little more up front and a lot less overall effort.

Let's just cut to it in general. There are dozens, probably HUNDREDS of little tips and tricks that are supposed to speed up Windows. Most of them do nothing, others have nasty side effects if you don't know what you're doing, and the rest will have a cumulative <5% net gain. Windows is already pretty optimized, and there really aren't any secret registry hacks or whatever else people try and tell you. Do you really think Microsoft would ship a product with some secret turbo mode turned off? The answer to that is they wouldn't. At least not without a VERY good reason. It would simply be bad business to do so. Think of all the fodder it would provide competitors.

There's a whole lot of misinformation floating around out there by so called "experts". People who don't fully understand the full implications of something they read. One of my favorite examples is from a few years back. Back when systems only had 512K of L2 cache RAM at most, people started to realize that you can only cache a certain number of memory addresses with this. So there was a lot of hysteria about how if you add more than some amount of RAM (I think it was 1GB, but it really doesn't matter) your system would take a performance hit, and so you should never add more RAM than that. This was only partially true. It was true that 512K of L2 cache RAM could only properly handle a set number of memory addresses, and it was also true that you would take a performance hit on any memory above that amount. Where the lack of a full understanding came in, was in that even though you would take a performance hit, that performance hit was relative to cached RAM. Compared to the alternative of using the hard drive for swap space, uncached RAM was still HUNDREDS of times faster. Many people failed to grasp that final, crucial, piece of the puzzle. There are still many such people out there today, and some of them work for large organizations.

CNet is not alone in having a few of these people on its payroll. They have plenty of company in the world of tech journalism. Another quick example. The police chief for the city of San Jose in the San Francisco Bay Area recently (over the summer) gave a presentation talking about how great some Internet filter program was that they wanted to implement in public libraries and what not. All of the reporters covering this event, save one, didn't even bother to bring a laptop so they could actively verify the veracity of the police chief's claims. The one reporter who DID bring a laptop, found that the presentation was very tightly controlled to basically cherry pick results. But all the rest were just simply taking everything said at face value. I'm not certain whether or not Cnet sent anyone to cover this event or not, I just know they were not the ones who had the reporter with a laptop.

Long story short, there are basically three things that affect system performance, and they are universal across all operating systems.

1: CPU
2: RAM
3: Number of programs running

You want to see any kind of significant performance gain in a system, focus on those three things. Everything else is either a scam, a waste of time, or both.

I regularly run Indesign, photoshop, an FTP client, MS Word and 8 tabs in Firefox. Ran fine a year ago on this machine. Now when I open Indesign, something comes up like 'memory low'.

I've kept this clean, running av, anitspyware, etc., which I hear can munch on memory like there's no tamarah.

So I figure putting some extra memory in this think might help. I'm still wonderin why this was fine a year ago and not now.

The problem with that, is that you fail to take into account that for all intents and purposes, the registry is simply a database. It's a very simple, very crude, and very ugly, flat file database. I'm sure it was only intended to be a temporary solution, and like so many temporary solutions, ended up being extended well beyond its original scope and intent. Unfortunately, now we're all stuck with this rather ugly and ineffective solution.

Flat file databases do have a couple of small things going for them however. One, is that they are very easy to construct. Someone with only a minimal knowledge of programming can make these things. They're also generally very fast on account of their low overhead. As they get bigger, it does get a bit harder to search them, BUT much of that is negated by the entire thing being loaded into RAM. Part of being essentially a specially formatted plain text file, is that you can cram a huge amount of info into a very tiny amount of space. While the registry has certainly ballooned beyond what it was when it took center stage in Windows95 (it existed at least as early as Windows 3.1, it just was almost never used) the amount of RAM in computer systems has more than kept pace.

I'd be surprised if the registry on the average computer was over 64MB. Which, granted, is a lot of text, but when we're talking systems with 4-8GB of RAM being pretty standard, it's nothing. And if the entire thing is loaded into RAM, you can search through every single entry in a minute or two. But you generally don't have to do that, because if you look a little closer at the registry the designers clearly anticipated this. There's a path system that programs can use to quickly locate the bit of info they need. It's one of the few good things you can say about the design of the registry, though why you want to store that much sensitive info into a single location, providing an excellent single point of failure... Well, only Microsoft can answer that one. I much preferred the old DOS system, where every program was it's own self-contained ecosystem. I rather like the way Apple puts kind of a modern spin on this. Under the hood, every .app file on OS X is really a directory tree for that app. You can probe it with the command line if so inclined. Everything the program needs to run is contained within, and the Finder shell automatically looks for the main executable in a specific location within that directory tree. Installing an app is as easy as dragging the icon to /Applications, and uninstalling is as easy as dragging it to the Trash. None of this wizard crap Microsoft has had a hard on for since Windows95.

But, long story short, it's a myth that leftover bits of programs in the registry have any non-negligible impact on system performance. It's simply not true. System performance really doesn't degrade over time, it's that our perceptions change. The expectation bar is raised as we become accustomed to our computer's performance levels, and as we are aware of faster computers being made. If you format, the process of reinstalling is usually slow slow and arduous, that by the time you're done, the system seems very quick and responsive. It's all in your head.

it seems to have a built in 4G limit

jonah

,.

eboostr4 has a feature where it can use your unmanaged additional RAM on your system if you are running a 32 bit windows OS & have more than 3.25 GB. Although it is in beta stages it works pretty well & its free to test