Hacker News Reader

Comment by _old_dude_ | Hacker News Reader

I will show my age, but is there a turbo button [1] ?

[1] https://en.wikipedia.org/wiki/Turbo_button

> Contrary to what it suggests, the "turbo" button was generally intended to let a computer run slower than the speed for which it had been designed when pressed.

Damn these trickster designers and their shitty design decisions. That's exactly the opposite of what I thought it did.

Like 10 years after my 486-DX2 went into the trash after a long long life, I was reminiscing about the turbo button with a coworker, and he mentioned that he always kept his turbo button activated and that there was never a single repercussion from it!

Given that Turbo made it slower, why would that cause repercussions?

Well, slower was the repercussion, but my coworker thought turbo meant faster, and was thinking along the lines of "my CPU didn't catch fire from all the extra speed I was getting"

You misuderstood. When the turbo button is pressed, computer runs on nominal speed. When turbo button is un-pressed, computer would run on half of nominal speed.

So, unpressed button makes computer slower, not pressed one.

Source: actually used the button on my 386 SX 20/40 MHz

Yep. I always had the turbo button pressed, except for occasionally deciding to play Test Drive 3 at speeds lower than "light speed". It was the one game I had that didn't calibrate the delay loop properly and my 486 was waaay too fast for normal playing - something that I didn't know so I learnt to play a very twitchy version...

Sorry, but I think you're confused in the same manner my coworker was. Activating the turbo button decreases the speed of your computer.

Maybe yours operated differently, but everything I can find about turbo buttons on the internet indicates that they operated as I remember them - your computer gets slower when you turn the button on. I specifically remember playing certain DOS games where, if I didn't activate the turbo button, the entire game would speed through at what felt like double or triple speed.

From wikipedia on Turbo button:

> On IBM PC compatible computers, the turbo button selects one of two run states: *the default normal speed or a reduced "turbo" speed*. It was relatively common on computers using the Intel 80286,[1] Intel 80386 and Intel 80486 processors,[2] from the mid 1980s to mid 1990s. The name is inspired by turbocharger, a device which increases an engine's power and efficiency. *Contrary to what it suggests, the "turbo" button was generally intended to let a computer run slower than the speed for which it had been designed when pressed.*[3]

https://en.wikipedia.org/wiki/Turbo_button

sigh

Look, you read about it in wikipedia, but I actually lived with this computer for my whole childhood. I know for sure how the turbo button worked, because I actually used it to cheat in F1 and Lotus racing games.

Best I can find in 10m I dedicated to this is this video showing switch between 22 and 40 MHz modes with a turbo button (around 8:50): https://youtu.be/TdenDH7S8LM

I've literally told you multiple times that I did too, and my computer behaved in the exact manner that all of the results googling the concept turns up.

You both may be right and wrong… further down in wikipedia page:

“ On most systems, turbo mode was with the button pushed in, but since the button could often be wired either way, on some systems it was the opposite.”

i dont know why the summary paragraph is so authoritative about what pressing the button means when a couple paragraphs down basically says “sometimes it’s wired backwards, so really it could be either faster or slower”

I don't know if you already knew why the Turbo button was there in the first place, but back at a time where there is no dedicated monotonic timer, applications (mostly games) rely on the fixed processor speed to determine sub-second time, which while worked on the original IBM PC... made it too fast for new generations of hardware.

Assuming you can detect what clock speed the device is running at, such an approach should still be able to work on different devices. Although with newer systems able to process multiple instructions per clock, as well as having clock speeds change rapidly, it probably wouldn't work as well.

The original applications that were effected were written without knowing there would be IBM PCs (no less clones) that ran at a different clock speed!

See sibling comment. Also, there was no CPUID in those ages!

Vide "macro photography"

Nikon got that one right.

As far as I remember, the button itself was supposed to let the computers run at their normal, faster, speed. Meaning the computer was supposed to be in Turbo mode, all the time.

You turned the Turbo mode off when you needed to run it slower, say for playing games.

So, technically the button did what it said, run faster in Turbo mode.

There were also clock speed selectors that really did let the CPU run faster. They would let you switch to a CPU clock speed that was higher than the speed the CPU was originally rated for. The cost was a greater possibility of crashes. Depending on what you were doing and how stable the CPU was it was a noticeable improvement.

Some selectors would let you choose between 2 speeds, others had multiple speed selections. Often the higher ones would crash a lot more. There were also variable speed controls that acted like a rheostat with continuous speed changes. Those worked but were even more unstable.

One of my father's computers had one of these.

It literally did nothing except change an LED readout on the case from saying "16" to saying "166".

I remember setting those 7-segment LEDs up on my first PC build.

On the motherboard there was a jumper which took the input from the turbo switch.

To provide feedback to the user, some cases ran a wire from the turbo switch to a bi-color LED on the front of the case as well as to the motherboard to control the turbo state.

On fancier cases they had multiple seven-segment LEDs which had a multi-state jumper for each segment, which decided to light that segment depending if the switch was on or off (example: http://www.minuszerodegrees.net/led_speed_display/LED%20disp...), so you could display either the two different clock frequencies, or a "HI"/"LO" message, or anything you wanted. And connecting that switch input to the motherboard wasn't necessary for the LED to change, so I guess that is how your father's was setup :)

Funnily enough, this is how this is done on intel+linux:

`echo 0 > /sys/devices/system/cpu/intel_pstate/no_turbo`

For just a second there I thought Intel provides the kernel some kind of backwards compatibility mode to clock down to 4.77MHz like the original turbo button.

There is a pin PROCHOT on all Intel processors that can be used to drop the CPU (and IGP, if equipped) into its lowest-power but still running state. It's designed to allow external voltage regulators on the board to assert that they are overloaded, but I don't see why you couldn't also wire in a turbo button there :=)

On my Lenovo Yoga it's always on when waking from sleep. You have to connect charger, even for a short time, to drop out of PROCHOT.

Reply on news.ycombinator.com