Comment On Confessions: The Soft CPU Upgrade

"Years ago," writes Maxime, "we found ourselves plagued with a brand new, unusably sluggish website. Most of the team blamed the esoteric VMCMWTH-based architecture (i.e. View-Model-Controller-Model-What-The-Huuhhhhh) that was pioneered by the Chief Developer. But the Chief Developer and the CTO (who also happened to be his uncle), blamed the hardware. More specifically, it was the 'inferior, off brand' CPU." [expand full text]
« PrevPage 1 | Page 2 | Page 3Next »

Re: Confessions: The Soft CPU Upgrade

2012-05-21 11:09 • by tchize
I've got to like this :)

Re: Confessions: The Soft CPU Upgrade

2012-05-21 11:10 • by Zac (unregistered)
And... cue the buffer overflow

Re: Confessions: The Soft CPU Upgrade

2012-05-21 11:15 • by PiisAWheeL
Awesome Comment Enabled!

Re: Confessions: The Soft CPU Upgrade

2012-05-21 11:15 • by Larry (unregistered)
Wow! Third! I am so fucking awesome! This never happened to me before in my life. The gods are finally smiling on me. Everyone will have to read my comment.

Hmm.

Now if only I had something to say.

Re: Confessions: The Soft CPU Upgrade

2012-05-21 11:20 • by DonaldK (unregistered)
381582 in reply to 381581
Larry:
Wow! Third!


Third? Your counter must be zero-based.

Re: Confessions: The Soft CPU Upgrade

2012-05-21 11:20 • by the beholder (unregistered)
We all know that if your app is going slow the obvious solution is to throw more hardware at it. We just never knew what to do if you can't do or exhaust that option. Now we know!

Re: Confessions: The Soft CPU Upgrade

2012-05-21 11:24 • by ¯\(°_o)/¯ I DUNNO LOL (unregistered)
static char *tdwtfname = "HyperTurbo 256-bit, AwesomeAkismet enabled";
Oh boy, now we'll be back to five a week!

Re: Confessions: The Soft CPU Upgrade

2012-05-21 11:32 • by Doctor_of_Ineptitude (unregistered)
A WTH (What The Huhh) framework app running off a 256-bit HyperTurbo CPU with AwesomeCache enabled, what possibly could have slowed it down.

On the other hand, the WTH frame does suffer from Relative syndrome. Now if only the CPU honoured the UNCLE flag.

Re: Confessions: The Soft CPU Upgrade

2012-05-21 11:35 • by John Doe (unregistered)
This reminds me of one of my first hacking attempts. It was a multi-user mainframe, and as a student my user ID didn't have what I considered sufficient access rights. But I knew one of the admins, let's say "bill", did. So I wrote a fortran program to deliberately overflow its array boundaries and go tromping through "core" looking for the address where my user ID was stored. When I found it, I tried to overwrite that string with "bill", reasoning I would thereby become bill as far as the system knew. Sorta like a soft CPU upgrade, you see?

Unfortunately when I tried to write to that memory location I discovered a little feature called hardware enforced access control. It was a read-only address for my process. Drat!

My second attempt was much more successful. I went over to the DecWriter (a hardcopy terminal) Bill had recently used, tore off the last several pages, and flipped back to where he had logged in. There was a button you could push to turn off echoing while you entered your password, but nobody bothered, because as soon as you hit Return the system would print asterisks and hashes etc. over your password rendering it unreadable.

Or not.

It wasn't hard to see that his password was "Jeni", his wife's name.

Fast forward a couple decades and attacking the system is still a waste of time when it is so easy to hack the user. How little we've learned!

Re: Confessions: The Soft CPU Upgrade

2012-05-21 11:38 • by Tasty (unregistered)
381587 in reply to 381585
Doctor_of_Ineptitude:
A WTH (What The Huhh) framework app running off a 256-bit HyperTurbo CPU with AwesomeCache enabled, what possibly could have slowed it down.

On the other hand, the WTH frame does suffer from Relative syndrome. Now if only the CPU honoured the UNCLE flag.


The CPU was saying uncle!

Re: Confessions: The Soft CPU Upgrade

2012-05-21 11:46 • by Nagesh (unregistered)
This ain't being good solution. If unemploymented, modole is ain't making sence to casual observer, yet is having fer reaching efects for entire system.

Re: Confessions: The Soft CPU Upgrade

2012-05-21 11:50 • by Se (unregistered)
381589 in reply to 381579
Zac:
And... cue the buffer overflow

From arch/x86/include/asm/processor.h:
char x86_model_id[64];

So there's no buffer overflow here :)

Re: Confessions: The Soft CPU Upgrade

2012-05-21 11:51 • by squidfood (unregistered)
It's good practice. Every optimization should include a placebo.

Re: Confessions: The Soft CPU Upgrade

2012-05-21 12:12 • by chooks

It's been a while since my C days, so...the real WTF is not using strncpy?

Re: Confessions: The Soft CPU Upgrade

2012-05-21 12:50 • by PiisAWheeL
381592 in reply to 381591
chooks:

It's been a while since my C days, so...the real WTF is not using strncpy?
Objection: Asked and answered your honor.
squidfood:
It's good practice. Every optimization should include a placebo.

Re: Confessions: The Soft CPU Upgrade

2012-05-21 12:51 • by Dave-Sir (unregistered)
381593 in reply to 381591
chooks:

It's been a while since my C days, so...the real WTF is not using strncpy?
TRWTF is strncpy. Hint: It doesn't terminate the string when you would most want it to.

Re: Confessions: The Soft CPU Upgrade

2012-05-21 13:03 • by Gazzonyx
381594 in reply to 381586
john doe:

This reminds me of one of my first hacking attempts. It was a multi-user mainframe, and as a student my user ID didn't have what I considered sufficient access rights. But I knew one of the admins, let's say "bill", did. So I wrote a fortran program to deliberately overflow its array boundaries and go tromping through "core" looking for the address where my user ID was stored. When I found it, I tried to overwrite that string with "bill", reasoning I would thereby become bill as far as the system knew. Sorta like a soft CPU upgrade, you see?

Unfortunately when I tried to write to that memory location I discovered a little feature called hardware enforced access control. It was a read-only address for my process. Drat!


What OS was this? You'd think the OS would segfault when you tried to write to memory outside of your process' allocated memory. Isn't that the usual thing to do?

Re: Confessions: The Soft CPU Upgrade

2012-05-21 13:13 • by JimboJones (unregistered)
I gotta ask - was that really the CPU ID string used in the solution, or is it an anonymized string used to avoid vendor unhappiness with this site? I just find it a little hard to believe that the Chief Developer could identify an "inferior, off-brand" CPU, but let a completely fictitious CPU slide.

I think I would have been tempted to request documentary evidence of the performance increase from the Chief Developer, and then revealed the CPU ID switch with an aim to discredit him.

Re: Confessions: The Soft CPU Upgrade

2012-05-21 13:26 • by TheCPUWizard
381596 in reply to 381586
...multi-user mainframe....DecWriter (sic)


It is DECwriter...made by...wait for it...DEC [Digital Equipment Corp]. The vast majority of these were used for PDP-11 and VAX computers, both of which were considered Mini computers and not mainframes...

Given that nearly evry mainframe of that time period was EBCDIC based, and the DECwriter was ASCII, it would be extremely suprising to see that combination.

Since I am involved in historical computing (and a member of a number of Musuems and Rescue organizations), I would be very interested in getting more information about this extremely suprising configuration.

Re: Confessions: The Soft CPU Upgrade

2012-05-21 13:26 • by Nagesh
This sound like made up fake story to me.

Re: Confessions: The Soft CPU Upgrade

2012-05-21 13:43 • by F (unregistered)
381599 in reply to 381596
TheCPUWizard:
...multi-user mainframe....DecWriter (sic)


It is DECwriter...made by...wait for it...DEC [Digital Equipment Corp]. The vast majority of these were used for PDP-11 and VAX computers, both of which were considered Mini computers and not mainframes...

Given that nearly evry mainframe of that time period was EBCDIC based, and the DECwriter was ASCII, it would be extremely suprising to see that combination.

Since I am involved in historical computing (and a member of a number of Musuems and Rescue organizations), I would be very interested in getting more information about this extremely suprising configuration.


He didn't say the DECWriter was connected to the mainframe, only that Bill had just been using it. The likelihood of someone using the same password for both his VAX account and his mainframe account, back then, would have been close to 100%. Especially someone who didn't turn off echo when typing his password.

Re: Confessions: The Soft CPU Upgrade

2012-05-21 13:44 • by F (unregistered)
381600 in reply to 381597
Nagesh:
This sound like made up fake story to me.


If you're going to have a fake story, it might as well be a made-up one.

Re: Confessions: The Soft CPU Upgrade

2012-05-21 13:47 • by Matt (unregistered)
Kinda like the time I installed Firefox on a user's PC, put a shortcut on the desktop, renamed the shortcut "Internet Explorer," and give it IE's icon. Whatever gets the job done!

Re: Confessions: The Soft CPU Upgrade

2012-05-21 13:58 • by n_slash_a (unregistered)
TRWTF is using loff_t i; instead of int i;

Re: Confessions: The Soft CPU Upgrade

2012-05-21 14:08 • by synp (unregistered)
381603 in reply to 381586
John Doe:
Fast forward a couple decades and attacking the system is still a waste of time when it is so easy to hack the user. How little we've learned!


That's because we're getting new and better systems all the time, but we've been using the same model user for millenia

Re: Confessions: The Soft CPU Upgrade

2012-05-21 14:31 • by Jim Rees (unregistered)
381604 in reply to 381596
We had large numbers of ascii terminals connected to our System 360 running MTS at University of Michigan, through a device called a Data Concentrator, which was a pdp-11 with a bunch of rs232 line cards. The ascii terminals included ASR33 teletypes, DECwriters, Tektronix 4014s, Silent 700s, etc. This was a common configuration at the time.

Also, the DEC-10 was often considered a mainframe, although it was tiny compared to the 360.

Re: Confessions: The Soft CPU Upgrade

2012-05-21 14:33 • by fizzbuzz (unregistered)
381605 in reply to 381588
Nagesh:
This ain't being good solution.


"Ain't?"

Nagesh is from South Carolina now? Mississippi, perhaps?

Re: Confessions: The Soft CPU Upgrade

2012-05-21 14:57 • by Anon (unregistered)
381607 in reply to 381595
JimboJones:
I gotta ask - was that really the CPU ID string used in the solution, or is it an anonymized string used to avoid vendor unhappiness with this site? I just find it a little hard to believe that the Chief Developer could identify an "inferior, off-brand" CPU, but let a completely fictitious CPU slide.

I think I would have been tempted to request documentary evidence of the performance increase from the Chief Developer, and then revealed the CPU ID switch with an aim to discredit him.


I think you'll find the replacement id was annoymized for comic effect. The original was probably AMD (which must of had a 40% market share at some point in the past).

Re: Confessions: The Soft CPU Upgrade

2012-05-21 14:58 • by Nickster (unregistered)
381608 in reply to 381596
It is DECwriter...made by...wait for it...DEC [Digital Equipment Corp]. The vast majority of these were used for PDP-11 and VAX computers, both of which were considered Mini computers and not mainframes...


I used to use a DECWriter attached to a Harris 400 on campus, which is evidently classified as a "supermini." My guess is the OP was referring to a large, installed, non-portable computer with a beefy OS as the "mainframe" even if it was technically a mini or supermini. It didn't take long for these small (read: refrigerator-sized) computers to surpass the mainframes of yesteryear in computing power. The operating systems they used worked pretty much like mainframe OSes anyway.

Re: Confessions: The Soft CPU Upgrade

2012-05-21 15:02 • by Jack (unregistered)
381609 in reply to 381605
fizzbuzz:
Nagesh:
This ain't being good solution.


"Ain't?"

Nagesh is from South Carolina now? Mississippi, perhaps?

Last week he "learned" the phrase "If it ain't broke, don't fix it." Ever since then, he's been using "ain't" in every post. It's getting pretty old. But then, the fake Nagesh has been getting old for a long time.

Re: Confessions: The Soft CPU Upgrade

2012-05-21 15:46 • by Nagesh (unregistered)
381611 in reply to 381609
Jack:

Last week he "learned" the phrase "If it ain't broke, don't fix it." Ever since then, he's been using "ain't" in every post. It's getting pretty old. But then, the fake Nagesh has been getting old for a long time.

We're all getting old.

Granted, some of us have been at it for longer than others.

Re: Confessions: The Soft CPU Upgrade

2012-05-21 15:56 • by titanofold (unregistered)
381612 in reply to 381591
Nope, it's that it didn't follow the GNU style guidelines.

acsi: I'm no' gonna acsi agun...pick up your room!

Re: Confessions: The Soft CPU Upgrade

2012-05-21 15:59 • by Nagesh (unregistered)
381613 in reply to 381605
fizzbuzz:
Nagesh:
This ain't being good solution.


"Ain't?"

Nagesh is from South Carolina now? Mississippi, perhaps?

I am Indian only from Hyderabad.

Re: Confessions: The Soft CPU Upgrade

2012-05-21 16:01 • by Gibbon1 (unregistered)
381614 in reply to 381605
fizzbuzz:

"Ain't?"

Nagesh is from South Carolina now? Mississippi, perhaps?


Nagesh learned English at the Baptist run Gen Lee Missionary School in Hyderabad.

Re: Confessions: The Soft CPU Upgrade

2012-05-21 16:08 • by vt_mruhlin (unregistered)
Call me crazy, but shouldn't you have just bought 1 HyperTurbo CPU and installed it in a test machine, then ran your load tests to prove that performance still sucked with the new CPU?

That would be a lot easier and more ethical than lying to your boss, and would carry the added benefit of maybe actually getting the problem fixed.

Re: Confessions: The Soft CPU Upgrade

2012-05-21 16:16 • by Hexadecima (unregistered)
381616 in reply to 381596
PDP-10s, PDP-11s, and VAXen were all considered mainframes by the late eighties, when the term no longer meant "the CPU cabinet is the size of a refrigerator" and meant instead "that big box that everyone dials into." This was especially prominent once the microcomputer era began.

Also, there were still lots of big non-IBM mainframes around when the DECwriter was introduced in the early seventies. :)

Re: Confessions: The Soft CPU Upgrade

2012-05-21 16:21 • by herby
381617 in reply to 381585
Doctor_of_Ineptitude:
A WTH (What The Huhh) framework app running off a 256-bit HyperTurbo CPU with AwesomeCache enabled, what possibly could have slowed it down.

On the other hand, the WTH frame does suffer from Relative syndrome. Now if only the CPU honoured the UNCLE flag.

Maybe using the UNCLE flag opened up channel D. You never know.

Re: Confessions: The Soft CPU Upgrade

2012-05-21 16:31 • by Ken B. (unregistered)
381618 in reply to 381596
TheCPUWizard:
...multi-user mainframe....DecWriter (sic)
It is DECwriter...made by...wait for it...DEC [Digital Equipment Corp]. The vast majority of these were used for PDP-11 and VAX computers, both of which were considered Mini computers and not mainframes...

Given that nearly evry mainframe of that time period was EBCDIC based, and the DECwriter was ASCII, it would be extremely suprising to see that combination.

Since I am involved in historical computing (and a member of a number of Musuems and Rescue organizations), I would be very interested in getting more information about this extremely suprising configuration.
What would you consider a KL-10 to be? That's what the DecWriters at my college were connected to back in 19(mumble-mumble). Given that it took up numerous cabinets in the computer room (behind the locked sets of doors, but visible through the pane-glass windows), I think "mini" doesn't fit. Sounds like "mainframe" to me.

Re: Confessions: The Soft CPU Upgrade

2012-05-21 16:59 • by The Great Lobachevsky
381619 in reply to 381618
KL-10 was a variety of PDP-10

http://en.wikipedia.org/wiki/PDP-10#KI10_and_KL10

Re: Confessions: The Soft CPU Upgrade

2012-05-21 17:35 • by AGray (unregistered)
So, lemme get this straight.

WTF #1: Writing a fictional hardware type to the operating system (provides no performance impact, but prevents insidious ID-10T Runtime errors.) I can buy that.

WTF #2: The 'improvement' in performance. Was it entirely vapor, or did some refactoring take place behind the scenes?

CAPTCHA: Transverbero - A word that becomes movement...

Re: Confessions: The Soft CPU Upgrade

2012-05-21 17:44 • by Mcoder
381621 in reply to 381620
AGray:
So, lemme get this straight.

WTF #2: The 'improvement' in performance. Was it entirely vapor, or did some refactoring take place behind the scenes?


It is not vapour. Placebo effect is quite real.

Re: Confessions: The Soft CPU Upgrade

2012-05-21 18:10 • by Anonymous (unregistered)
381622 in reply to 381601
Matt:
Kinda like the time I installed Firefox on a user's PC, put a shortcut on the desktop, renamed the shortcut "Internet Explorer," and give it IE's icon. Whatever gets the job done!
I usually replace iexplore.exe with firefox.exe and get the IE theme just in case.

Re: Confessions: The Soft CPU Upgrade

2012-05-21 18:11 • by Anonymous (unregistered)
381623 in reply to 381622
Anonymous:
Matt:
Kinda like the time I installed Firefox on a user's PC, put a shortcut on the desktop, renamed the shortcut "Internet Explorer," and give it IE's icon. Whatever gets the job done!
I usually replace iexplore.exe with firefox.exe and get the IE theme just in case.
I also edit hosts file to redirect bing.com to google.com

Re: Confessions: The Soft CPU Upgrade

2012-05-21 18:17 • by facilisis (unregistered)
381624 in reply to 381609
Jack:
fizzbuzz:
Nagesh:
This ain't being good solution.


"Ain't?"

Nagesh is from South Carolina now? Mississippi, perhaps?

Last week he "learned" the phrase "If it ain't broke, don't fix it." Ever since then, he's been using "ain't" in every post. It's getting pretty old. But then, the fake Nagesh has been getting old for a long time.

Yes, he's quite old now. Unfortunately even if he dies of old age, he'll probably be reincarnated as an even bigger wtf than he is already.

Re: Confessions: The Soft CPU Upgrade

2012-05-21 18:20 • by Vanders
381625 in reply to 381605
fizzbuzz:
Nagesh:
This ain't being good solution.


"Ain't?"

Nagesh is from South Carolina now? Mississippi, perhaps?


Nagesh almost becomes funny again if you read every comment in a hick Southern accent.

Re: Confessions: The Soft CPU Upgrade

2012-05-21 18:51 • by Norman Diamond (unregistered)
381628 in reply to 381596
TheCPUWizard:
...multi-user mainframe....DecWriter (sic)
It is DECwriter...made by...wait for it...DEC [Digital Equipment Corp]. The vast majority of these were used for PDP-11 and VAX computers, both of which were considered Mini computers and not mainframes...

Given that nearly evry mainframe of that time period was EBCDIC based, and the DECwriter was ASCII, it would be extremely suprising to see that combination.

Since I am involved in historical computing (and a member of a number of Musuems and Rescue organizations), I would be very interested in getting more information about this extremely suprising configuration.
OK, since you're involved in historical computing, you need to acquire some historic IBM manuals. For the 360 and 370, which must be considered mainframes even by yourself, an IBM manual explained the absurd manner in which ASCII values would be transfered from ASCII terminals to the computer's core memory. The parity bit was preserved, affecting the value visible by the user. Each byte had its 8 bits reversed, including 7 bits of ASCII value and 1 bit of parity. The TR instruction didn't care how absurd the translation table was, so it happily converted the input to EBCDIC or converted back to IBMSCII for output.

Terminals made by IBM used other weird character sets. The 2741 came in two varieties which needed different translation tables. However, the 2741 was good for security. I removed the typeball before typing my password and then put the typeball back on afterwards.

I don't remember what I did on ASR33 teletypes. Maybe I ripped out the part of the paper where my password had been partly covered up.

Re: Confessions: The Soft CPU Upgrade

2012-05-21 19:35 • by Friedrice the Great (unregistered)
381630 in reply to 381609
Jack:
fizzbuzz:
Nagesh:
This ain't being good solution.


"Ain't?"

Nagesh is from South Carolina now? Mississippi, perhaps?

Last week he "learned" the phrase "If it ain't broke, don't fix it." Ever since then, he's been using "ain't" in every post. It's getting pretty old. But then, the fake Nagesh has been getting old for a long time.


Old enough to retire yet? Please?

Re: Confessions: The Soft CPU Upgrade

2012-05-21 19:36 • by Coyne
381631 in reply to 381628
Norman Diamond:
For the 360 and 370, which must be considered mainframes even by yourself, an IBM manual explained the absurd manner in which ASCII values would be transfered from ASCII terminals to the computer's core memory. The parity bit was preserved, affecting the value visible by the user. Each byte had its 8 bits reversed, including 7 bits of ASCII value and 1 bit of parity. The TR instruction didn't care how absurd the translation table was, so it happily converted the input to EBCDIC or converted back to IBMSCII for output.


The reversal of bits was a requirement of RS-232, which transmits bits in LSB-first order, by standard. So a "C" character, normally considered 0100 0011 in ASCII was sent with a start bit of 0 and a stop bit of 1, as 0 1100 0010 1. (An idle line was normally mark or 1 and so if you could consider the idle line as a series of ones, a single C would be sent as ...1111101100001011111111...)

For ASCII, the parity bit was normally in the MSB position of the 8 bits, because the ASCII code assignments in that era actually used only the lower 7 bits (of 8). Parity would not typically be visible on the terminal (though it might indicate an error if the parity was wrong). But, yes, one would be able to TR ASCII-with-parity into normal EBCDIC characters or vice-versa, since both were 8 bits.

Rumor had it that the exact translation table from ASCII to EBCDIC was an IBM trade secret that was actually kept on paper in a vault for safekeeping. And one would guess it still is since I have seen (and been frustrated) by multiple, but slightly different, translation tables.

Re: Confessions: The Soft CPU Upgrade

2012-05-21 20:31 • by Computer Clown (unregistered)
381632 in reply to 381594
It smells suspiciously like Honeywell GCOS, which allowed such tomfoolery. Back in the day, you could do some very (nasty|cool) things, like deliberately cause a system program to fail, and specify the file that the coredump went to. Ok, it involved rumaging through the object code, but you could patch it, and send it on it's way, altered to your specification.

Re: Confessions: The Soft CPU Upgrade

2012-05-21 20:35 • by Computer Clown (unregistered)
381633 in reply to 381586
John Doe:
This reminds me of one of my first hacking attempts. It was a multi-user mainframe, and as a student my user ID didn't have what I considered sufficient access rights. But I knew one of the admins, let's say "bill", did. So I wrote a fortran program to deliberately overflow its array boundaries and go tromping through "core" looking for the address where my user ID was stored. When I found it, I tried to overwrite that string with "bill", reasoning I would thereby become bill as far as the system knew. Sorta like a soft CPU upgrade, you see?

Unfortunately when I tried to write to that memory location I discovered a little feature called hardware enforced access control. It was a read-only address for my process. Drat!

My second attempt was much more successful. I went over to the DecWriter (a hardcopy terminal) Bill had recently used, tore off the last several pages, and flipped back to where he had logged in. There was a button you could push to turn off echoing while you entered your password, but nobody bothered, because as soon as you hit Return the system would print asterisks and hashes etc. over your password rendering it unreadable.

Or not.

It wasn't hard to see that his password was "Jeni", his wife's name.

Fast forward a couple decades and attacking the system is still a waste of time when it is so easy to hack the user. How little we've learned!


It smells suspiciously like Honeywell GCOS, which allowed such tomfoolery. Back in the day, you could do some very (nasty|cool) things, like deliberately cause a system program to fail, and specify the file that the coredump went to. Ok, it involved rumaging through the object code, but you could patch it, and send it on it's way, altered to your specification.
« PrevPage 1 | Page 2 | Page 3Next »

Add Comment