From what hd partition Knoppix core work faster???

[alex52]

I use Win-xp boot.ini to boot Windows, DOS, several small Linux versions, and Knoppix without installation, which core is placed on separated small hdd partition, as Knoppix consideres all partition as read-only cdrom.
But I’m curious; from what hd partition (vfat, ext2, ext3, reiserfs) Knoppix core should work faster????

Thanks for any information. In this question I concern only about speed, not security and so on.

Best, Alex

[chris-harry]

well... the one closest in the middle of the drive (dead center), is the fastest...

but as file thingy... i really dont know... i would guess ext2... but thats only a guess...

[ZeroKun]

I'm not quite sure what your exact question is, but ext2 or 3 should be ok to install. They should down your harddrive's preformance. Ext3 might be the best choice since it has journalling, which keeps everything from fragmenting.

[Cuddles]

I use Win-xp boot.ini to boot Windows, DOS, several small Linux versions, and Knoppix without installation, which core is placed on separated small hdd partition, as Knoppix consideres all partition as read-only cdrom.
But I’m curious; from what hd partition (vfat, ext2, ext3, reiserfs) Knoppix core should work faster????

Thanks for any information. In this question I concern only about speed, not security and so on.

Best, Alex

I tend to agree with ZeroKun...

The biggest thing that "plagues" Windows' systems is fragmentation, DOS's idea of "filling all the gaps" over "keeping the file together" causes the drive to jump around to get what information it needs... With small files, this isnt an issue, but, with bigger files, the heads of the drive must jump around to pick up all the pieces of the file. Faster drives have made up for a file system not maintaining continuous files, but, this "fragmentation" issue hits both data and program information. If, lets say, both your program and data files, are fragmented, the system must not only wait for the complete program to load, but also its data...

What becomes more of a "double-edged" sword, is the "cost" of having something running and keeping the file system un-fragmented. Microsofts answer to this, at least in the VFAT arena, is to have the user run, frequently, a defragmentor, preferably on your down-time, by use of there scheduler.

Linux's ideas, are to build that defragmentor into the file system itself. Ext3, and reiser, are perfect examples of this. The file system itself, monitors the fragmentation, and resolves as it goes, thus, no "de-fragmentor" program is needed for the user to worry about doing on there own time.

Not to "bash" any OS, but, M$ has been adding "Band-Aide" 's to there OS "file system" for a while, and Linux, just started from scratch. The file systems in Linux were "designed" to be "self-monitoring", and "self-correcting" on the fragmentation issue...

Speed is relative... Most everything on a computer is based on the "speed" of something else. Your processor is based on the speed of your memory, of your hard drive, of the bus, etc... You can have a 6 GHz processor, but if your RAM can only run at 433 MHz, your processor is going to be waiting, if your bus is not as fast as your processor, it will be waiting, if your hard drive isnt as fast, it will be waiting, etc...

I like to use a good ( old ) example: I had a TRS-80 "Color Computer", ran at some insanely slow speed, used your TV for a monitor, and a cassette tape for storage... You could "poke" a value at an address, and "force" the processor into a "over clock" speed. Problem was, it lost all access to any I/O channels after you did this. So, a program would poke this "over clock" when it started, and when it needed input, poke the "normal" speed to regain its composure. The biggest problem was that the program could lock-up, or crash, and you had no way of knowing it, or doing anything about it.

You can find, that almost anything on your computer, is dictated, by speed, by something else. As chris-harry pointed out, the "center-most" area on a drive is probably the fastest, but, the cylinder sizes get smaller nearer the center. Imagine a record ( old ), songs are placed on the record, from outside to inside, the Revolutions Per Minute dont change, but the songs look longer, as you get towards the center. It takes more revolutions in the center to store the same information as on the outer. Same holds true with your hard drive. You would think the "speed" would be faster in the center, but, it must read more, in essence, than on the outer cylnders. Time and Speed are factors in the same equation.

How fast can your hard drives "heads" move from on place to another? How fast does the "platters" spin? How much, and how fast, is the "burst rate" of your hard drive? How fast is the bus that it must transmit this data on? How fast is the "temporary" storage of where this data is stored, till the processor needs it? How fast is the processor itself? If this "information" must have "Input / Output", how fast are those devices? ( video, seriel, keyboard, mouse, USB, etc... )

I, honestly, think, even if you had the "fastest" computer, in EVERY way, from the hard drive, all the way up to your video, that having a fragmented drive, would slow it down... Even at the cost of having something else running in the "background" that would defragment it... But, then again, if you had a drive that wasnt fragmented, and everything else was slow, around it, the outcome would still be slow.

The best you can hope for, is, a fast system, a fast file system, and a un-fragmented drive... ( and lets not add into this, internet access... Having a seriously fast system, as mentioned above, would be nothing compared to having to "squeeze" full-frame video, through a garden hose, like 56K - it would be like comparing drinking ice through a straw )

So, alex52, if I was going to suggest the "fastest", or "speed" related, as far as "file systems" are concerned - go with any of the "journaled" ones; ext3 or reiser... What little is "given up" in the fact that they process fragmentation, you will gain in not having to de-fragment. As far as location, on the drive, I think that one is negligable and insignificant; inside, outside, or center, your drive uses an "average" speed of all three areas of the drive surface... Not sure if its still out there, but Spin-Rite used to have a drive tester. It tested center access speeds, outer, and inner, and based them on time. Then used an average to determine "total access time", because not "everything" can be in one specific place on a drive. Some defragmentors, I think it was Symantic could "sort" the defragmentation by Programs first, data last. And, as far as I recall, most programs need data, usually, so its a give and take thing. I think the "newest" thinking on defragmentors is that: place most used first, least used last, but, I think its all a sense of milliseconds, or even, nanoseconds, at that level...

Maybe a engineer, of the "old computer" school, can give more information on the "fastest place to put data, or programs", to get you the most "speed" from your hard drive, but, again, not everything can be in the exact spot of your hard drive...

Just my two schillings worth ( and with the cost of living, I gave a little more, for the value )
Ms. Cuddles

[alex52]

P.S. I’ve found answers for my questions. Extended explanations may be fount here:

http://www.knoppix.net/forum/viewtopic.php?t=17024

Best, Alex