Can one afford a mainframe computer?

[icett]

Hello,


Can we, the normal people afford a mainframe computer? By mainframe I mean not personal computer but large computer usually used in companies. I dont know what other kind of computers are there but I hope that you may have understood my point. Also would be able to use it as personal computer like running desktop Linux or Windows or other desktop operating systems and use it like personal computer? Could we play PC games and use other PC software in that computer? Please clarify and forgive if what I am asking is stupid. :)

[bsharp]

I highly doubt anyone could afford a mainframe, and they usually had obscure RISC architecture processors which probably don't have a Linux port (although you never can be completely sure in the Linux world). Also, the mainframe has been almost completely rendered obsolete because of advances in off-the-shelf hardware.

[holiday]

Interesting! Why do you want one?

[cherva]

Look here http://idea.uab.es/mcreel/ParallelKnoppix/ You can transform normal pc's into cluster :) But normal programs are not written to be executed on clusters....

[starcannon]

You mean something along the lines of this?
http://www.youtube.com/watch?v=vLujLtgBJC0

Oh and then Japan has this:
http://www.youtube.com/watch?v=hrVS45nuIfs&feature=related

No, I doubt the average bear can afford one, at $5.2 million to $350 million dollars.

[abgemacht]

Licenses for IBM Mainframes can be up to $60,000. Not including hardware. Mainframes are used for keeping track of massive amounts of data. They are not designed for intensive graphics (such as gaming).


Also, the mainframe has been almost completely rendered obsolete because of advances in off-the-shelf hardware.

Not true at all. Mainframes are still the backbone of almost all major IT companies, banks, and other large institutions.

[lisati]

Interesting idea......

A few years ago (back in th 1980s), a colleague and I had the idea of getting hold of an old IMB 360, and setting it up to use, but we never got very far with the idea.

[yabbadabbadont]

Interesting idea......

A few years ago (back in th 1980s), a colleague and I had the idea of getting hold of an old IMB 360, and setting it up to use, but we never got very far with the idea.

IBM 360/370 Assembler was the first assembly language I ever learned. 15 registers and no stack, if I remember correctly.

[lisati]

IBM 360/370 Assembler was the first assembly language I ever learned. 15 registers and no stack, if I remember correctly.

It was one of the first languages I learned and regularly used too. If memory serves correctly, instead of a stack like we'd use on an x86 machine, there was a linked list of "save areas".

It wouldn't surprise me if there's still a copy of some of my badly written IBM 370 assembler code hiding in cyberspace somewhere in Wellington (NZ), even though I haven't used it for over 20 years.

[Old_Grey_Wolf]

You may be interested in a HP Blade System. They are expensive. Links:
- description http://www.hp.com/sbso/busproducts_blades.html
- some prices http://www.google.com/products?hl=en&rlz=1G1GGLQ_ENUS280&q=hp+blade+system+price&um=1&ie=UTF-8&sa=X&oi=product_result_group&resnum=5&ct=title
- and http://geekswithblogs.net/WallabyFan/archive/2007/09/14/HP-BladeSystem-c3000-for-SMBs.aspx

I can't afford one. :lolflag:

[dominiquec]

There was one selling on eBay not a few years back.

That said, it's the maintenance that will kill you ;-)

[yabbadabbadont]

It was one of the first languages I learned and regularly used too. If memory serves correctly, instead of a stack like we'd use on an x86 machine, there was a linked list of "save areas".


Each subroutine (whatever they were called) had to provide it's own stack area. The actual use of the stack was pretty much the same as how a C compiler uses it on an x86 machine, in that the caller had to clean it. However, I seem to remember that the called subroutine's address went in a specific register, and the return address went into another, and then you had to call BALR. My textbook from the time is across the room on the bottom shelf of a bookcase. I guess I could look it up. :D

Edit: and the address of the stack area also went into a predefined register... I think. Darn! I'm really going to have to go get that book now. :lol:

[bobbob1016]

My dad actually programs AS-400's, in RPG. For a few days we had one in the house, one of the "new, smaller" ones, only 3 feet tall, 3 feet or so wide, vs the 6 or 7 feet ones. They had to take it to fix it though, this is years ago, haven't seen it since... All I can remember about it is wanting to use it as a file server. Since it had a few terabytes. My dad said "You can't run stuff on it, or off of it, just a green screen." I said "Oh" at the time, I wasn't as good with computers at the time, so I said ok. Now I'd have said "Yeah, the AS-400 can't run any of the files but my PC can."

That computer was about $300,000-400,000 at the time, this was under a different president, so you know how long ago I mean.

[lisati]

Each subroutine (whatever they were called) had to provide it's own stack area. The actual use of the stack was pretty much the same as how a C compiler uses it on an x86 machine, in that the caller had to clean it. However, I seem to remember that the called subroutine's address went in a specific register, and the return address went into another, and then you had to call BALR. My textbook from the time is across the room on the bottom shelf of a bookcase. I guess I could look it up. :D

Edit: and the address of the stack area also went into a predefined register... I think. Darn! I'm really going to have to go get that book now. :lol:

That sounds right: I have memories of BALR, R14 being used for the return address, R13 being used for the address of the stack area, R1 pointing to a parameter list. There was, if memory serves, a "do nothing" program called IEFBR14, which was commonly used when people wanted to create or delete files using JCL only. It has been a long time....

[yabbadabbadont]

That sounds right: I have memories of BALR, R14 being used for the return address, R13 being used for the address of the stack area, R1 pointing to a parameter list. There was, if memory serves, a "do nothing" program called IEFBR14, which was commonly used when people wanted to create or delete files using JCL only. It has been a long time....

OK, I just had to pull out the textbook and check. Looks like there were 16 registers, R0 through R15. R0 and R1 never seem to be used in any of the example code. R2 is the address of the stack space. R15 is the address of the main program, and R14 is the return address.

I don't think that we did too badly with our recollections, given how long ago it was. ;)

[t0p]

I highly doubt anyone could afford a mainframe, and they usually had obscure RISC architecture processors which probably don't have a Linux port (although you never can be completely sure in the Linux world)

Linux a a *nix. And Unix certainly runs on minicomputers, so I'd imagine it'll run on a mainframe. I suspect the OP isn't differentiating between minicomputers and mainframes anyway.

EDIT: Thanks to Sporkman's post (below), I found this link (http://www-03.ibm.com/systems/z/os/) that shows Linux indeed does run on mainframe servers.

[Sporkman]

http://www-03.ibm.com/systems/z/news/announcement/20080226_annc.html

[Daishiman]

For those interested in classical mainframe usage model (IBM sells it to new customers now for Linux server consolidation), you can download the Hercules z/Architecture emulator (http://http://www.hercules-390.org/) and get one of the free public-domain OS/360 operating systems to run on it http://www.ibiblio.org/jmaynard, or check in your favorite P2P app for the latest copy of z/VM or z/OS.

It's not that much fun actually, since you're not running the stuff that makes it interesting (CICS, DB2, Oracle), but it does scratch the itch.

[lisati]

OK, I just had to pull out the textbook and check. Looks like there were 16 registers, R0 through R15. R0 and R1 never seem to be used in any of the example code. R2 is the address of the stack space. R15 is the address of the main program, and R14 is the return address.

I don't think that we did too badly with our recollections, given how long ago it was. ;)

In some of the programs I did, R1 was used to point at the equivalent of the command line parameters or arguments to functions/subroutines. I think R15 was sometimes used to set a return code. I don't think I ever used R0.

[yabbadabbadont]

It could just be that the textbook consistently used one convention while the actual hardware didn't care. I never got to play with the real hardware, just an emulator back in the 80's. My dad taught me to read punched cards in the early 70's using assembler programs he had written for the 360. (I was doomed from an early age...)