Turbo Button | Hacker News Reader

Turbo Button

en.wikipedia.org

195 pointstosh113 comments

Turning off the TURBO mode on my P1 133MHz box helped with misbehaving cdrom drive that refused to read cds. My explanation at the time was slower reading is more reliable.

Oh the disappoitment when I disassembled the case and noticed the turbo button is not connected to anything except for led display showing Hi and Lo on 7-segments. It was just placebo effect.

It's not necessarily placebo: http://catb.org/jargon/html/magic-story.html

that could the premise to a movie. a kid finds the magic button his grandparent's basement and flips it.

Oh man, you just gave me a flashback to the buffer underrun days in the mid 90s. Even moving the mouse in Windows 95 could cause a bad burn and render a (then expensive) CD-R useless. I’m glad that was short lived.

But if you didn't move the mouse, the screensaver would kick in, and then you would definitely get a buffer underrun.

I got more reliable CD burns on my 14MHz Amiga than I did on my ~200MHz Win95.

that certainly wasn't LED display

It almost certainly was -- https://en.wikipedia.org/wiki/Seven-segment_display

I’ve heard from multiple sources that the original goal of the “turbo” button was back-compat for a feature of the original 4.77Mhz PC that depended on clock-rate: using an NTSC TV as a monitor. The original PC could have been 5Mhz, but 4.77 chosen instead so that the CGA graphics adapter could use the bus clock to sync up with the NTSC signal without needing an addition clock (saved a good 25 cents on the hardware that way!). When faster PCs came out, they lost the TV display feature. But, if you really needed it, you could clock down your fancy 6Mhz PC back to 4.77 to regain TV display capability.

“TV back compat mode” is long and boring, so the button was named “turbo” in a brilliant marketing move. It became so popular that PCs added the ability to arbitrarily clock down at the touch of a button long after it made sense to do so.

I was about 6 or 7 years old, and the home computer wasn't much of a thing back then, and the 12 year old kid next door got a 386 DX, which was a huge deal, and he kept it in his bedroom! Lucky guy.

Anyway, his Dad loaded it up with some DOS games, which I would go over and watch him play. He would play Prince of Persia and let me watch. I would occasionally ask him if I could have a go.

He said I could only play if the TURBO button was set to OFF. I asked why and he said that it was to "save electricity". I would play PRINCE.EXE at a snails pace until he kicked me off, me being the determined kid from a non-computer owning household.

The upside is that their family had a cat (which was albino, deaf as a post, and eccentric as heck). They wouldn't let the cat inside (maybe to save on electricity) and the cat adopted us because we let it in. They moved and we kept the cat.

How I played Tetris: 1. start it while turbo is enabled

2. wait for Tetris to adjust to the speed of 8Mhz 8086

3. turn off turbo and play Tetris almost twice as slow - actually managed to reach 32000+ a few times ;) (as it overflows on 32768 you had to stop playing just before or else you won't end on the high score list)

I kinda want one on my desktop machine now. There's a simple beauty to a button that can slow your computer down if it's too fast for you. Maybe one on my life too.

My i7-7700 with 32 GB RAM is already crawling when I want to use Facebook Messenger, slowing it further down would be terrible...

Yeah: software turbo buttons have been a thing for a few years now. I find the Microsoft Teams client works quite well in this regard, also.

Try closing Slack for Desktop!

Renault Sport, Porsche, SAAB and Nissan were doing great things with turbo charged cars in the 1980's, the implication of 'turbo' was clear and you could get 'Turbo' bicycle seats and even 'Turbo' sunglasses that embraced the concept. So it was a bit of brilliant thinking to name the clock speed button 'Turbo', it did what it said on the tin.

In the current era cars have more sophisticated interiors, often featuring a dial that enables the driver to choose between 'eco', 'comfort' and 'sport' settings. If you have got a long drive and you are paying the petrol then 'eco' might make sense, if you haven't got the wife and kids in the car and the boss is paying for the petrol then 'sport' might be the preferred setting.

We have all sat in front of laptops that whirr away even if we are just kicking back and reading Hacker News. We have all sat in front of laptops on some train journey where 'range anxiety' sets in and you fumble for screen brightness and other controls to extend the battery life. Yet we probably specified that super fast CPU with lots of RAM for compiling stuff or, heaven forbid, playing the odd game.

Wouldn't it be great if the Turbo button came back as a simple hardware switch that enabled you to easily switch up from 'eco' to 'comfort' to 'sport' modes? If you wanted a break from the fan noise then you could just slip into 'comfort' mode, then, when back to the grindstone with every second counting (due to the realities of project time keeping) you could step it up into 'sport' mode, knowing that fan noise won't disturb anyone in your open office? Back home you could then go back to 'comfort' mode with the fans only kicking in on compile, not because you have a gazillion tabs open.

So yes, I share your sentiment but I think the idea needs a reboot for the 2020's, by which time, as per the original 'Turbo' button the concept will be obsolete as we will all be using Intel 12th generation CPUs.

Said Intel CPUs do have “turbo boost” that increases the clock on certain cores. It would be cool to have a button to influence that.

There were many computer cases where case buttons including Turbo were installed on standard drive bay covers. You can move it to your PC and connect it to the fan turning it on and off or switching a voltage. Without sufficient cooling your system will throttle.

Aaaah our first PC. It had a Intel 80286, a display up front and a turbo button. I pressed it once. My parents became very mad. Because the fans of the PC spun up so much in turbo mode, they thought it would damage the computer. Just like how pressing ctrl+alt+delete did damage to the computer

Sweet memories

Still wish I hadn't taken that computer apart (and never put back together) in my never ending quest to quench my curiosity

But think how much you learned in the process!

Hahaha, I never use reaction gifs but this too perfectly describes how I feel: https://media.tenor.com/images/0ba58bde3909480fd8f1749982df6...

On the one hand, yea I guess I did learn some things from it, but on the other hand I went on to become a software developer not an electrical engineer, so how much did I really learn from it? Wouldn't it have been better off keeping it for the future... Some of my peers have some really cool old hardware that they used to do whatever on and I'm kind of envious of that :)

my friend (we were kids back then) refused to edit his config.sys and autoexec.bat because he didn't want to "unplug the processor" :)

Because the fans of the PC spun up so much in turbo mode,

That's interesting, because 286s and 386s normally didn't even have heatsinks. Did they wire the turbo button to the fan of the PSU?

IIRC, heatsinks were showing up on low-frequency 486s and heat sinks with fans on 486s >= 50MHz. I think (but may misremember) that most 486 fans were also always running at the same speed.

(Not contesting your statement, just wondering what weird beast your first PC was ;).)

It could have been a white box PC. My stepfather’s business computers (in the mid 90s) were from some local mom and pop place that built custom PCs and also supported it if things went wrong. I think they tended to over-engineer things as all the PCs were full size (beige of course) AT towers with noisy fans.

Dating back to the 5150, most PCs had case fans. IIRC sometimes the fans would spin a little faster with Turbo engaged, even if just for effect.

They must have. The computer spend some months at a tech savy uncle when he first acquired some pre release version of windows 95 but it was also bought probably from a local store.

Windows 95 did not run on a 286, the 286 was very limited in running real mode (8086/88) applications concurrently in protected mode (it did not have the virtual 8086 mode that the 386 had).

The minimum requirement of Windows 95 was an 80386, though it was unbearably slow (tried to run it on an 386SX25 or 386DX40). It was usable on a 486 with 8GB of memory.

My first personally bought computer was a 386 with said turbo button. Other crucial performance features were the sram cache and the ability to overclock the isa bus from the standard 8mhz up to 11.

This was particularly important for running OS/2 1.3 with the svga vesa (non-accelerated) driver that was produced by an intern at IBM. That software made many lives including mine better on a daily basis.

A weird thing was how at 11MHz I could audibly hear large screen refreshes. Never quite understood where that sound was from.

> A weird thing was how at 11MHz I could audibly hear large screen refreshes. Never quite understood where that sound was from.

Was it a whining sound? First guess would be spike in power consumption and the inductors making noise in the power supply.

It wasn't like a transformer whine, more of like a white/pink noise 'sheeh' sound that would sound once when there was a large amount of video memory updated e.g. full graphic screen clear especially on high res/colour modes.

I remember that from certain phix modes (a TSR VGA emulator for Hercules/monochrome systems). Other modes would cause a consistent audible whistle to come from the CRT. But hey, at least I could play raster games.

Nowadays, certain games make the GPU in my desktop whine, though I'm pretty sure it's some kind of coil whine in my case. It's equally mysterious however, it's only certain games. And it's not load-based, some relatively lightweight games do it and sometimes it's heavier games. Only a few do it, too.

The large screen refreshes could be coil whine. I had a video card in 2010 that would do the same thing.

When I was a kid, we had a BBC Micro that would make audible little hissy noises when writing to the screen. At the time I just assumed it was "the sound of the computer thinking" but in hindsight there was probably some EMF cross-talk between the analog sound lines and the bus to the video memory. Maybe this was something similar?

Even when the carrier frequency is orders of magnitude too high to hear, there are usually harmonics low enough to be audible.

That makes sense. In my case I didn't have a sound card, OS/2 didn't have the best support for peripherals or gaming. Before 1.3 running on something other than a PS/2 was pretty much unheard of.

Other crucial performance features were the sram cache and the ability to overclock the isa bus from the standard 8mhz up to 11.

And among other oddballs, does anyone remember VESA Local Bus?

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

Terribly unstable, but also terribly fast (for the time) for 2D graphics with a Tseng Labs GPU.

VLB was the hotness of the time. Cheap, fast, and widely available. It's success brought about the AGP and other faster pci variants.

Haha memory lane. I remember walking into a little computer store with a good friend years ago and seeing a 386 that displayed some really odd frequency on the front panel led, like 53 Mhz or something at a time when 33 Mhz was the fastest 386 you could buy. I asked the owner and he said "Landmark speed!" (Landmark was a popular benching utility and he had programmed the led to display the benchmark results when the turbo button was pressed).

Programmed? The speed displays I remember were simply wired to show whatever. I've seen turbo buttons swapping the display between '66' and '33'.

That's what I remember as well. You'd set jumpers to light up the segments of the display. No programming involved.

I'm guessing the store owner ran the benchmark once and then hard-coded the benchmark result in the display.

Well yeah, I used "programming" rather loosely. He set it up that way.

It just appeared to me that there are now people in tech who never saw one of these.

> Disengaging turbo mode slows the system down to a state compatible with original 8086/8088 chips. 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.

Now I have no idea if the PC I grew up with was running much slower than it could have been.

I pressed mine and everything went slow-motion, so I guess you could have tested it xD

Surely you tried it both ways!

And if you never noticed either way then it doesn't matter!

My family's first Pc has an LED to show turbo mode was on, and the second had a big green LED display that changed from 16 to 33.

Even without that the difference was very obvious, especially when playing games.

Of course, just because the button was there doesn't mean it was actually connected to anything!

I was rather amused when I read about this years ago. As a child I always asked myself "why should I not press this button?!"

Only the cleverest of people figured out they could run their PCs on turbo full-time.

> In almost all cases the display wasn't a measure of the speed, but the two "fast" and "slow" display options were set by jumpers on the motherboard

Ah yes, I remember changing those jumpers to get that display show the correct speed, after overclocking my "CyrixInstead" 486.

Do you remember seeing jumpers on the motherboard for that? The only ones I recall seeing were on the display in the case itself, like this this: http://www.minuszerodegrees.net/led_speed_display/LED%20disp...

You are exactly right. They were never on the motherboard

Yes these jumpers resided on the display board.

> The turbo button was linked to a turbo LED or MHz display on the system case, although the indicated MHz was not always accurate.

On my 486 system the display said "486" with turbo on and "66" with turbo off, which does not seem to make any sense.

On mine there were jumpers and you could configure the 8 segment LEDs to show any numbers you wanted.

That's how mine is too. When its original owner added an AMD 5x86, they changed it to show "133" in turbo mode.

One case I saw was configured to display "Lo" and "Hi".

That is how I always configured the display jumpers on AT cases. It seemed more cool than meaningless numbers

"...cpu frequency was used for timing"

reminds me how even today how seemingly hard it is to program a simple clock. for example, the countdown in pubg...

> the countdown in pubg

Care to explain?

PlayerUnknown's Battlegrounds or "pubg" is a popular but somewhat shoddy battle royale multiplayer shooter, widely criticised for bugs including several to do with the countdown that takes place as players are loading into a match.

Thanks, I know PUBG. Just curious what exactly the bug is? The timer was counting down waiting for enough players to join, is the issue related to estimating how long it will take to get 100 players?

once enough players join there is a 60 second countdown until the match starts. but it is not realtime. it often does things like, 60..59..58..55..53...

The original IBM PC lacked an RTC, clones (as with everything else) had occasional portability glitches with them, and this was a decade before the TSC arrived. If you wanted a game to have a delay loop, you counted instruction cycles. If you wanted those delays to be portable, you manually detected the CPU and ran a pre-calibrated custom loop.

Yeah, that seems hard to me.

The "correct" way to do this was the 1/18th sec timer interrupt, even when the RTC worked it didn't have sub-second resolution.

http://www.pldworld.com/_hdl/1/VHDL_courses/www.ee.duke.edu/...

I thought the PIT was slaved to the bus clock on early systems and wasn't real time, no? Might be wrong.

I didn't have a turbo button on my first PC (mid 1990s), but the BIOS menu had a "deturbo" option. I had one game (some version of The Incredible Machine, or possibly a knock-off) that could only really be played in deturbo mode. I envied my friends who could easily switch between turbo and deturbo; I'd usually have to wait several extra minutes for the system to start up any time I wanted to play the game.

Also https://news.ycombinator.com/item?id=17345227.

This is such a PC solution. When the Apple IIGS came out, there was a software control panel to adjust the speed:

http://apple2history.org/wp-content/uploads/2010/06/Apple-II...

I remember when noobs in my middle school computer class got their first sight of an IBM PC and were mystified as to the function of the 3.5" floppy eject button.

"You have to use your hands? That's like a baby's toy!"

UK Apple ][e's came with a switch under the keyboard to change $ to £, so Apple was not immune from this approach.

Nope it changed # to £ by switching the character ROM instead. I think it also swapped keyboard layouts (i.e. AZERTY to QWERTY) but that wasn't an issue for UK folks.

Anyone remember cutting themselves really badly on AT cases? I still have scars from raiding a bunch of ex-office machines

Yes! The biggest scar on my body is a 3cm long cut in the shape of the eye of Sauron from the edge of an AT case. Dangerous stuff fiddling with computers back in the day.

I still remember when I made my dad to open the tower case of our 486 and reconfigure jumpers located beyond the 7-segment display which showed the CPU frequency. After 10 minutes we were done and I was so proud that, when turbo is pressed, display showed 42 MHz. Oh, the memories...

I was a kid at the time when we had a 486 and I was always scared of the turbo button. I had no idea what it did. I remember pressing it once and fooled around for a bit...nothing seemed to be different so I turned it off. I was scared for a little while after that someone might have found out that I pressed the turbo button...

For a long time after though, until we got a new computer, I would try turning it on randomly playing games or doing different things. I remember coming to the conclusion that it didn't do anything and we probably just didn't have whatever the turbo mode was supposed to be for.

I forgot about all that until I read this. It's pretty cool to find out after all these years it actually did do something.

My first PC, bought back in the summer of 1986, was an Opus (now defunct UK clone builder) with a 12MHz 286 processor, 1MB of RAM and a 37.5MB (RLL not MFM) 3.5" hard disk. It cost me an eye watering GBP 1200 (I was still a student at the time). However it did pay for itself quite quickly doing homers in knocking out dBase II and dBase III "apps" (and eventually Clipper stuff).

This thing had a turbo button on the front which would switch the CPU between 4.77MHz and 12MHz. I never had a need to run in 4.77 mode except for nostalgia purposes (I bumped into IBM PC's and compatibles quite early on).

My 386sx/25MHz also had a turbo button. It was rather useless, as I wanted the machine to go faster, not slower...

Here's another turbo-nostalgia memory from further back:

The Commodore C64 tape drive had a small hole which let you adjust the tape head azimuth to get a better signal. Sometimes, after borrowing tapes with games from a friend, one had to do adjustments as the turbo loader did not load the game correctly.

Hence the adjustment screw was called "the turbo screw".

I remember those games that were too fast on "newer" CPUs... unfortunately I don't have this problem anymore and the games are not that much better...

Same, I recall upgrading from a 286 to a 486 in the 90s, and on the new computer, Lemmings was way too fast to play. It didn't have a turbo button and I wasn't allowed to tinker with the guts, so there was no way to slow it down.

My 486DX had one. It actually connected to the MB, and controlled the CPU speed. Some games like Prehistorik 1/2 had slowed down considerably when I toggled the button.

The button was also wired handsomely. When untoggled (out position), the Turbo LED was lit and system was fast. When pressed, LED is turned off and system was slow. Since all buttons (Power, Turbo and Reset) were flush, it was aesthetically pleasing too.

Heh I had many computers with turbo buttons, including a few Pentium II slot processor machines. Oh man, slot processors. Slot 1 was so weird.

My first PC had two turbo buttons; one on the case for the CPU, and another on the modem to cap at 300 baud or allow blazing 1200 baud.

Police Quest was simply impossible without the turbo button back in the day. Or that is, until I discovered the menu bar by accident.

Despite being released around the beginning of the 486 era, Space Quest IV still had CPU-speed-sensitive timing. It was fine on a 386, but on a Pentium-60, the chances that the patrol droid would appear and kill you instantly at the beginning bit of the game got close to unity.

Slowing down a computer is useful when programming, but isn't easy to do. I had made my own tool to slow down network connexions, maybe useful to someone: https://github.com/webdev23/kamtar

That sounds pretty handy. Not sure if it's the same thing, but Charles Proxy offers throttling too. Either way, I'll take a look at your when I'm not on a phone

The Devtools in Firefox have this built in (a slow network simulator).

Yes, but can't switch off.

We had an IBM PC XT with an expansion board that upgraded the 8088 to a 80286. To drop speeds back to approximately what they used to be, or toggle back, you would hold both shift keys at the same time, and it would respond with either a high- or low-frequency beep.

If you don't mind me asking, is it possible you're misremembering? I believe the XT/286 was actually a different computer than the XT, and ran a 286 on the motherboard; it did not upgrade to 286 via expansion board.

I have fond memories of my XT/286.

The 80286 went into the IBM PC AT. The XT was the one with the 8088 chip. However after the AT 80286 came out, it was possible to get an upgrade card for the XT 8088 that was essentially an 80286 and motherboard on a card. It was like brain transplant for your computer. It wasn't as fast as running a real AT but it was cheaper and a lot faster than the 8088.

This doesn't sound right to me. We had an XT/286 and I don't recall any expansion card involved. Wikipedia seems to agree with me that it was a totally separate model. That being said, it doesn't include a citation for the info, so it's possible it's incorrect.

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

>We had an XT/286 and I don't recall any expansion card involved...

And other people had an XT that they installed an expansion card into -- like the Pfaster-286[1]

[1]https://books.google.com/books?id=xsMx9D2s6y0C&pg=PA43&lpg=P...

The '286 expansion cards were not IBM products. they were produced by third parties to give the older XT performance approaching the AT '286 computers.

286 expansion cards were indeed a thing. Here's some discussion of them:

http://www.vcfed.org/forum/showthread.php?43553-286-processo...

From time to time I still miss turbo button and a (functioning) pause key. On DOS you could pause execution to inspect output... I guess with multiprocessing it became more difficult to implement this...

  pkill -STOP "$(xprop _NET_WM_PID)"
  pkill -CONT "$(xprop _NET_WM_PID)"

This might have some unintended consequences though (file locks, clock drift between running program and your quartz...).

Regarding turbo, I'm pretty sure a shell script could set the CPU's speed governor to "conservative" at the press of a button.

Thanks, it's not quite the same though. Pause key was much more convenient, and worked on every system. Thanks nevertheless!

Also, for terminals that support it: `Ctrl+S` to pause the display only (the command continues to run) and `Ctrl+Q` to unpause. Still not perfect... but it could help sometimes.

You can also use the cpulimit program to do that.

https://github.com/opsengine/cpulimit

Ctrl+S does this in most terminal emulators (use Ctrl+Q to resume).

I just built a PC today and the motherboard has a switch for "slow mode". This is a modern system.

What motherboard?

Bunch of high-end gaming motherboards (at least some models from ASUS and MSI) have that, for extreme overclockers as a "safety switch" to clock down the system while they are changing settings (where otherwise the system would crash in the meantime due to overheating)

That's a gaming/enthusiast motherboard that supports clocking down the CPU to help with thermal issues, meant for extreme overclocking/cooling situations.

Slightly different use case, the "slow mode" switch on your motherboard is for use when overclocking with liquid nitrogen cooling.

actually now curious what this is.

was thinking about undervolting my home system this sweltering summer.

My first every PC, in high school. I changed the jumpers on the turbo button to read "69".

Love reminders of how old I am :)

Yep I got one on my brand new 386SX 40 MHz back in the day.

Turbo would switch it between 20 and 40 MHz.

Reply on news.ycombinator.com