RAM DISK

[Paul]

It's no use placing the page file on a RAM disk.
And yo my opinion is better to put the page file on a different physical
disk then the system disk. And as in most cases the system disk will be
C:, that won't be the C: drive.

Fokke

There is, if:

1) You're using Windows x32, with memory license limit.
2) You have more than 4GB of memory installed.
3) You want to get some usage from that memory.
4) Windows has PAE enabled, and the RAMDisk driver method,
means you get to use the memory. (A driver can access PAE
accessible memory, when a regular application is not allowed to.)
5) Now, placing the pagefile.sys on the RAMDisk, is "free".

Under lots of other conditions, it makes no sense at all.

I'm only suggesting this as a *test*, if you meet the conditions
above. It wasn't stable enough for my liking, but that
could change tomorrow, as more bugs are found and eradicated
in the product. Because it was reasonably well behaved, it's
pretty close to "useful" right now. But as long as I see the
odd glitch, all I can suggest to others, is *test carefully*
and make note of what happens.

Starting a program, seeing the program in Task Manager,
yet no window appears, is one of the glitches I saw.
That's with pagefile.sys on RAMDisk.

Paul

 

[Yousuf Khan]

I've got 16GB of RAM. Don't have much activity on the Pagefile.

I used to think the same thing until early yesterday, when my 8GB of RAM
was suddenly 96% full! I had no reason to think it would ever get that
full, because usually after a fresh reboot, it sits at less than 25%
full, and hardly ever grows to more than 50% full. Obviously, some
application must've overflowed its memory banks, and not released the
memory after exiting. But the point is that however rare, these things
do happen. I was seeing a lot of hard fault activity on my disk,
averaging about 16 faults/sec in some cases. But the system remained
perfectly usable and responsive, as I had my page file spread out over
multiple hard disks, none of which were the system disk. I was able to
reboot the system in a normal fashion and get it back to sanity.

Yousuf Khan

 

[Paul]

I used to think the same thing until early yesterday, when my 8GB of RAM
was suddenly 96% full! I had no reason to think it would ever get that
full, because usually after a fresh reboot, it sits at less than 25%
full, and hardly ever grows to more than 50% full. Obviously, some
application must've overflowed its memory banks, and not released the
memory after exiting. But the point is that however rare, these things
do happen. I was seeing a lot of hard fault activity on my disk,
averaging about 16 faults/sec in some cases. But the system remained
perfectly usable and responsive, as I had my page file spread out over
multiple hard disks, none of which were the system disk. I was able to
reboot the system in a normal fashion and get it back to sanity.

Yousuf Khan

The 64 bit version of CHKDSK, will do that, use all the RAM.

No other program I've got here, is quite that greedy.

Paul

 

[Fokke Nauta]

I tried it. I installed Dataram's RAMDisk; set up a 2GB disk (I only
have 6GB RAM installed), loaded ConvertXtoDVD, changed the work folder
to be on the RAMdisk, started it on converting a DivX file; and got an
immediate error, "not enough space - you need at least 4.8GB".

Hhhhmmm! The rest of the computer seems to be working ok as normally.

That's the end of my trial. I'm not buying more RAM just for the dubious
pleasure of whatever minute benefit I might get.

I'll be uninstalling it shortly.

Ed

Well, 2 GB is surely not enough.
I use 8 GB and even that was not enough in a situation where I tried to
denoise a lot of pictures with Neat image. It was hungry for scratch space.

Fokke

 

[Paul]

I tried it. I installed Dataram's RAMDisk; set up a 2GB disk (I only
have 6GB RAM installed), loaded ConvertXtoDVD, changed the work folder
to be on the RAMdisk, started it on converting a DivX file; and got an
immediate error, "not enough space - you need at least 4.8GB".

Hhhhmmm! The rest of the computer seems to be working ok as normally.

That's the end of my trial. I'm not buying more RAM just for the dubious
pleasure of whatever minute benefit I might get.

I'll be uninstalling it shortly.

Ed

4.7GB = a single sided DVD.

It was trying to write the entire single sided DVD output in there ?

When compressing video, writing to a destination disk is
not likely to be the rate-limiting step. The compressing
is the hard part. Even with hardware acceleration, the
write rate still probably doesn't need more than a hard
drive. Even an SSD would be overkill.

Paul

 

[Ed Cryer]

4.7GB = a single sided DVD.

It was trying to write the entire single sided DVD output in there ?

When compressing video, writing to a destination disk is
not likely to be the rate-limiting step. The compressing
is the hard part. Even with hardware acceleration, the
write rate still probably doesn't need more than a hard
drive. Even an SSD would be overkill.

Paul

I guess the program allocates an output file of fixed length; big enough
for the standard DVD±R. Or at least it insists on that space being
available before it starts.
Seems a fair check to me. I'd hate it if it ever got half way through
and then told me that there wasn't enough space.

I think the average sizes involved with the ones I work on are 700MB
(DivX) >>> 2.5 GB (DVD).

Ed

 

[Paul]

I guess the program allocates an output file of fixed length; big enough
for the standard DVD±R. Or at least it insists on that space being
available before it starts.
Seems a fair check to me. I'd hate it if it ever got half way through
and then told me that there wasn't enough space.

I think the average sizes involved with the ones I work on are 700MB
(DivX) >>> 2.5 GB (DVD).

Ed

I can give you an example of something I used a RAMDisk for.

I downloaded the source for Firefox. There are a lot of file in
there. Perhaps the number was 70000 or so (~3600 folders, 65000 files).
Now, if I decompressed and worked on that on the hard drive, it was
quite slow decompressing the files. The size of the archive wasn't
the challenge. It was the seek speed of the hard drive. The heads
took a pounding during the decompression, and later, when I was
running scripts (to do things like change the line endings so
that the source files would look good in Notepad).

So I tried the RAMDisk for that, a 2GB file system just to hold
Firefox source while I was reading it. I could search the source text,
looking for subroutine names. That sort of thing. And that was very
smooth and fast. Zero seek time on a RAMDisk. An SSD would have been
perfectly good for the job as well, but I still don't own one.

Paul

 

[Ed Cryer]

I can give you an example of something I used a RAMDisk for.

I downloaded the source for Firefox. There are a lot of file in
there. Perhaps the number was 70000 or so (~3600 folders, 65000 files).
Now, if I decompressed and worked on that on the hard drive, it was
quite slow decompressing the files. The size of the archive wasn't
the challenge. It was the seek speed of the hard drive. The heads
took a pounding during the decompression, and later, when I was
running scripts (to do things like change the line endings so
that the source files would look good in Notepad).

So I tried the RAMDisk for that, a 2GB file system just to hold
Firefox source while I was reading it. I could search the source text,
looking for subroutine names. That sort of thing. And that was very
smooth and fast. Zero seek time on a RAMDisk. An SSD would have been
perfectly good for the job as well, but I still don't own one.

Paul

We're all going to own an SSD in the near future. I've seen SSDs of 1TB
size. And the largest HDD I've seen is 3TB.
To my way of thinking that's the death-knell for HDDs.

Just sit back and watch the price tumble and tumble. That occurs at the
crossroads of old/new technology. Pretty soon new PCs will have a SSD as
standard; and then prices tumble faster.

That will include all spinning discs; CDs, DVDs, retro vinyl et al.
They'll become like mag tape and VCRs. Maybe good "vintage" stuff for
the wealthy, but useless for the majority.

I wonder what might happen with the "Voyager" probes. Maybe some day
some civilisation light-years away will find one with its spinning discs
and instructions to rotate at 33 1/3 RPM, while bouncing a needle along
the spiral groove and amplifying the generated squeaks and squarks. And
they'll build a vast spaceship to come and find the origin. And they'll
find Earth, and we'll dismiss them as primitive life-forms.

Anyway, back to the future!
Your usage for editing vast text files might be worthwhile, but I ask in
all seriousness, how many such large text files are there?

Ed

 

[Gene E. Bloch]

The heads
took a pounding during the decompression, and later, when I was
running scripts (to do things like change the line endings so
that the source files would look good in Notepad).

Some editors (I use Notepad++ and have used NoteTab Lite) are happy with
non-Windows line endings, so there's no need to convert if you're
willing to use a new interface.

I often will open up a text-type file with Notepad by accident, and then
I grumble until I remember to go back to Notepad++ (I'm sloppy about my
file associations).

http://notepad-plus-plus.org/

I don't have a link for NoteTab Lite.

 

[Paul]

Some editors (I use Notepad++ and have used NoteTab Lite) are happy with
non-Windows line endings, so there's no need to convert if you're
willing to use a new interface.

I often will open up a text-type file with Notepad by accident, and then
I grumble until I remember to go back to Notepad++ (I'm sloppy about my
file associations).

http://notepad-plus-plus.org/

I don't have a link for NoteTab Lite.

I think I ran something like "unix2dos" against the files,
and "fixed" them. Not a big deal. I used "file.exe" to
identify which ones were text, then called "unix2dos".

Paul

 

[Paul]

Anyway, back to the future!
Your usage for editing vast text files might be worthwhile, but I ask in
all seriousness, how many such large text files are there?

Ed

I think I have three archives of that type (porkers).

I know, because the Kaspersky scanner dies when it hits them
They have to be moved to another (unscanned) partition because of that.
You would think a good scanner would have schemes for protecting
itself against decompression bombs - the archives don't cause
7ZIP a problem at all.

Modern code is just so monstrously large. Even when I have
access to the source, there are many, similar looking routines.
Makes it almost impossible to figure anything out. Even with a
good IDE, I'd be lost in there.

Paul

 

[Ken Blake]

We're all going to own an SSD in the near future. I've seen SSDs of 1TB
size. And the largest HDD I've seen is 3TB.
To my way of thinking that's the death-knell for HDDs.

I'm not sure when it will happen, but I agree. Not too far down the
road, SSDs will become much cheaper, spinning hard drives will vanish,
and all drives will be SSDs.

But I've been saying this for the last ten years or more--well before
there were any SSDs on the market. I haven't been right yet, but I'm
getting a lot closer.

 

[Ed Cryer]

I think I have three archives of that type (porkers).

I know, because the Kaspersky scanner dies when it hits them
They have to be moved to another (unscanned) partition because of that.
You would think a good scanner would have schemes for protecting
itself against decompression bombs - the archives don't cause
7ZIP a problem at all.

Modern code is just so monstrously large. Even when I have
access to the source, there are many, similar looking routines.
Makes it almost impossible to figure anything out. Even with a
good IDE, I'd be lost in there.

Paul

I have text files 30MB in size. Dictionaries. When I try to open them in
Notepad it takes for ever; but Editpad and Notepad++ open them
immediately. I've just done a search and found a 300MB text file
supplied in a language package. And it opens immediately with Editpad.
This doesn't come anywhere near 2GB, of course, so I can't comment on that.
The thing with a good text-editor is to read the file into RAM and do
the editing there; and that simply makes a RAMdisk redundant.

Modular programming! Including good internal comments. That keeps source
files down to a manageable level. 2GB for a source file just sounds like
spaghetti to me; archaic, monstrous and unmanageable.

Ed

 

[Iceman]

Some editors (I use Notepad++ and have used NoteTab Lite) are happy with
non-Windows line endings, so there's no need to convert if you're
willing to use a new interface.

I often will open up a text-type file with Notepad by accident, and then
I grumble until I remember to go back to Notepad++ (I'm sloppy about my
file associations).

http://notepad-plus-plus.org/

I don't have a link for NoteTab Lite.

http://www.notetab.com/notetab-light.php

And file associations are easily changed through My Computer > Properties.

 

[Gene E. Bloch]

http://www.notetab.com/notetab-light.php

And file associations are easily changed through My Computer > Properties.

I'm glad you are aware of that.

There are easier ways, BTW.

 

[Iceman]

I'm glad you are aware of that.

There are easier ways, BTW.

Really? Which are they? (And please don't recommend any software )

 

[Gene E. Bloch]

Really? Which are they? (And please don't recommend any software )

Right click on a file of a given type and work it out from what you see.

 

[Yousuf Khan]

The 64 bit version of CHKDSK, will do that, use all the RAM.

No other program I've got here, is quite that greedy.

Thunderbird and Firefox in my case.

 

[Yousuf Khan]

I have text files 30MB in size. Dictionaries. When I try to open them in
Notepad it takes for ever; but Editpad and Notepad++ open them
immediately. I've just done a search and found a 300MB text file
supplied in a language package. And it opens immediately with Editpad.
This doesn't come anywhere near 2GB, of course, so I can't comment on that.
The thing with a good text-editor is to read the file into RAM and do
the editing there; and that simply makes a RAMdisk redundant.

Actually, what those text editors are doing is the exact opposite of
loading it all into RAM. The reason they can start up so quickly is that
they *aren't* reading it all into RAM right away. They're using
demand-paging, meaning they only read the parts of the file they need at
the time that they need it. So when they start, they may only have
loaded a small fraction of the whole file into memory, so that they way
they are starting much more quickly. Notepad on the other hand tries to
load the whole file into memory ahead of time, that's why it takes so
long to load at the beginning.

Yousuf Khan

 

[Yousuf Khan]

I'm not sure when it will happen, but I agree. Not too far down the
road, SSDs will become much cheaper, spinning hard drives will vanish,
and all drives will be SSDs.

But I've been saying this for the last ten years or more--well before
there were any SSDs on the market. I haven't been right yet, but I'm
getting a lot closer.

The next challenge will be to come up with a standard interface faster
than the SATA interface, and allow an SSD to plug in almost directly
into the PCI-E bus. The only reason we're using the SATA interface at
all is because we want to maintain compatibility with hard disks, but
when hard disks go away, the need to maintain compatibility with them
will have passed too. They do have some SSD's that currently plug into
PCI-E x4 or higher interface, but they are by no means standardized
devices.

Yousuf Khan