Lyle Bickley explains the PDP-1 (and we play the original Spacewar!)
HTML-код
- Опубликовано: 1 май 2017
- Lyle Bickley, of the PDP-1 restoration team, gives us a tour of this amazing, early scientific interactive computer at the Computer History Museum. The first machine built by DEC in 1959, it features a superb graphics screen. DEC gave one to MIT, and some very bright students went wild. Gems such as Spacewar!, Snowflake, 4-voice music programs were all developed by moonlighting MIT students, unencumbered by its measly 12kW memory and pokey 100,000 instructions per second. Along with much more serious debugging and programming languages of course. You can come and see the real machine for yourself at the Computer History Museum in Mountain View, California:
www.computerhistory.org/
Also, Norbert Landsteiner made this incredible simulation of the PDP-1 that can run the original Spacewar! and Minkytron code in your browser:
www.masswerk.at/spacewar/
www.masswerk.at/minskytron/
He also made a gate exact replica with Verilog code on github:
• PDP-1 recreated in FPG...
Our sponsor for PCBs: www.pcbway.com
Support the team on Patreon: / curiousmarc
Merch on Teespring: teespring.com/stores/curiousm...
Learn more on companion site: www.curiousmarc.com
Contact info: ruclips.net/user/curiousmarca... 
Наука
Correction! After reviewing the video, I realized that I misstated the size of the Model 30 display. I said it was 19", but it is actually a 16". The tube (a 16ADP7), was commonly used in radar systems of the period.
P7 phosphor is yellow green and long persistence.
Would be interesting to see the code for that light pen. I wonder how much data can it capture, or in other words, how long of a line can it register.
You are the most valuable resource in that museum. With out, you those machines will just be big pieces of junk. Hope to be there one day.
@alysdexia I´m talking about the person who apears in the video: Lyle Bickley.
U are a genius 😉
My brain is telling me I am looking at the 1980s because I cannot fathom this at 1959, this is beyond cool
Funny you commented that because Atari also got a version of Space War and the engine for it was used in Asteroids too.
I remember the 80ies, especially the Amiga a bit more powerful ;)
I just am blown away how quickly things progressed after the microprocessor was invented. There were some experts before then that thought only governments, schools and large businesses would onlt want or need computers.
@@twistedyogert This was not *due* to the microprocessor though. The key was massive fast memories. That would have led to powerful computers even without the microprocessor.
@@herrbonk3635 Right, but cost went down with microprocessors. Only large corporations or government agencies like NASA could afford a computer. But with microprocessors you can buy one that is just as powerful for your office. I was told that price comes down with fewer numbers of components. So it's more efficient to combine multiple circuits into a single component rather than building them discreetly. Also, power requirements are lower, didn't some of these early machines use several kilowatts?
I played Space War on the PDP-1 when I was 6 years old in 1962. My dad worked on the video monitor. It was difficult for him to bring me into work, because the engineers didn't want kids around. The machine also played Bach-like music through speakers inside burlap faced speaker boxes. In that time the game was played through the main console switch bank.
Wow that is a amazing
wait so you’re 62/63?
You was a lucky kid!
@@samus-rd1channel585 *where (if you weren't using slang)
@@simonzinc-trumpetharris852 whoops haha. Thank you for correcting me (for real).
It's absolutely amazing. Actual head to head combat, full on gravity physics, unique models for each player, projectile tracking, particle effects... I mean it's impressive.
I'm thinking it's not full on gravity physics. I notice the shells don't seem to be affected by gravity. :) It only affects the ships. But then, this is inside of 4KW. Can't have everything. :)
I remember playing a version of this on my Atari STe way back when (with sprite graphics), and it had the shells and particles affected by gravity, as well. That was pretty cool. When you'd blow up, the particles got drawn in, and scattered around the central star! :)
Programming those light effects was incredible for 1959.
It's incredible for 2021.
This Pdp has shader 2.0 😂
It wasnt 1959 but 1962
This was the first computer I ever saw. It was in a warehouse of a company south of market Street in San Francisco and they made their own ribbon cable joysticks and invited a bunch of us in to look at the computer
And allowed all of us to play space war for a little bit.
I was 16 in 1968 And from then on I wanted my own computer one day and finally got an apple two in the 1970s.
I am 68 now.
I will be 70 in 21 days. How computing has changed since 1968 !
apple 2 was great.. it birthed multiple genres
@@rudestbeast4907 I still have my Apple ][ GS, ][E, ][C. apple ][+ & integer based apple & all work fine though I have not used them for awhile. My GS has a bad 3 1/2 inch drive & I need to get a new one.
Most of the demos were indeed written by MIT students, including the music program by Peter Samson. Marvin Minsky was a Professor when he wrote "Minskytron". Steve Russell was doing consulting work at MIT when he wrote "Spacewar!".
Fantastic demonstration, I learned a lot from it. Kudos to everyone involved with keeping this machine in such great condition!
I was under the impression several people contributed to Spacewar!
brilliant !!!
The principal author was Steve Russell, Peter Samson was responsible for the star field, and Dan Edwards and Martin Graetz were contributors to optimization of the code, etc.
I remember Steven Levy wrote about that machine and the culture it created when it was at MIT. In the book HACKERS, Levy mentioned the guy who wrote the star field code made it match a certain part of the sky. I don't know if that's true.
I think Levy also told a story about a student who wanted to add another instruction to the processor, so they went in one night and had wired it, but it caused some problems of some sort.
1024p display on 1959. awesome
And in 2018 most laptops are 768p and most desktop displays are 1080p.
Really makes you think... 🤔
my old tv is 480p and it isn't touchscreen xD
Old CRT televisions have better black levels and contrast levels than modern LCD tvs. Plasma has a far superier image to LCD's.
A plasma will have better quality imagine than a 4k HDR LCD today that cost $10,000.
@@alexnemeth3680 Except this one was for like 10 people and actual laptops are build in millions and have a battery
Technically a vector display didn't have a fixed resolution. 1024x1024 is simply the number of individual coordinates for the electron beam to travel between. Vector displays don't have pixels or dots and technically resolution is infinite.
My first ‘personal’ computer experience. My friends and I would go into the lab at midnight, and load and play Space Wars. 1967. Still looks thrilling 50+ years later!
I was born in 2004, and grew up playing video games and really only experienced modern computing
I got into computers when I terrorized by friends and coded up my own roblox scripts lol. Now im invested into the whole hobby
But after breadboarding up a computer out of discrete logic IC's, ima be honest, its damn near magical lmfao. You get a whole new perspective on computers like the PDP-1.
It's absolutely amazing someone was able to create the PDP-1 in just over 3 months back in 1959, the complexity of this machine boggles the mind.
'
Wonderful video.
I am proud to have worked for this company (Digital Equipment). I learned pretty much everything there, and the spirit and work ethics of DEC are something that sticks with you for life. Even though I joined late, you met so many amazing people there whom you could learn from, it was incredible! We had folks who would read a memory printout in Hex as if it was a children's book, people who could diagnose network or hardware problems with the precision of a laser beam... I'm forever grateful to have met each and every one of them, and I owe what success I had in my career chiefly to them and DEC as a whole. It is so sad that this company ceased to exist because of poor management and decision making...
Sad that so many technology companies with skilled engineers went the same way because of horrible management. You and I both surely know the survivors haven't always been the ones with the best hardware or software!
The music is so beautiful! 60 year old chiptunes!
stillstyle Nope, not chiptunes. Pure wave output without DMA buffers. And with serious speakers too.
@@MegaUpstairs i find the piece interesting simply because of its presentation on the PDP-1, i would be interested to know what the piece is, its routine baroque music and could be from anyone of the thousands of composers of the period or a MIT student.
@@MegaUpstairs someone suggests "Bach BWM 592, movement 3" @4:00 , thats a concerto bach transcribed from an amateur prince composer.
@@MegaUpstairs I'll bet you know the first part as its quite memorable in its simple ways, but i had certainly forgot how the rest sounded : ).
Donating the computer, err, sorry "Programmed-data-processor" to MIT was a genius move from DEC. I bet they were not fully aware of what their creation would be capable of doing on "the right hands"! It's a shame that DEC no longer exists. My father's job was to repair them, beginning from the PDP-8 I believe. Thanks for the amazing video!
Your father had an awesome job!
Only partially true. I started with DEC back in 1976. Back then, DEC made every part of the computer. They made the boards, disk drives, floppy's, terminals both printers and video along with their own CPU's. By the time Compaq came along, DEC had sold off many of these business to other computer company's. For the most part, DEC sold off their PC business to Compaq. What was coveted back in the late 90's was their field service business and of course their CPU manufacturing. The Alpha chip rocked the world at the time and even Intel stole part of its architecture which showed up in their Pentium class of chips. Yes, they were sued and lost. Sadly, once the Board kicked Ken Olsen out as CEO, DEC died. My job was lost in 1993 as our business unit went out to Colorado Springs and after that, I have no idea what company acquired that division. Just a fantastic place to work. The micro VAX was on the Shuttle and there are still VAX running systems out there.
I worked for DEC in the 1980's. Donating computers to universities was a part of their business strategy. When those students went on to become engineers and scientists they would buy the same equipment for their businesses they had used in school.
The founders of DEC came out of MIT. The reason they created the company was they wanted to do more with building interactive computing, which MIT, other universities, and others in the industry were not so keen on funding at the time. The PDP-1 was a commercialized version of the TX-0, which was built at MIT, and was one of the first computers to run on transistors.
It took a little more prompting from John McCarthy, with his concept of "utility computing" in 1961, to get MIT on board with interactive computing.
What I'm remembering from my history is that the reason DEC named their first series of machines "Programmed Data Processors" had something to do with investors. They felt that there were too many computer companies at the time (DEC would be entering a crowded field, they thought), and so the company came up with the idea of not calling their computer a "computer." :)
Jaw dropping.1959, discreet transistors, 4k core memory - and look what it can do! Just amaizing. One deeply humbled programmer here.
4 "kilo"words and it's 18 bit words. So it's 9 KiB in modern terms.
And it didn't have a hard disk! That's what's amazing. 👍👍
At our university, they had us write up something similar in Assembly on a small CPU driving 2 DA converters connected to an XY scope. At the time, tech had advanced quite a bit further of course, but it was instructive and a lot of fun to go back to the basics
@@kaasmeester5903 is that recent?
or 20+ yrs ago?
I'm not seeing/hearing about a lot of low level projects like that anymore.
In the '90s we did real time controls with 68HC11 MCUs and assembly, but the equivalent class was PIC based using python when my son went in 2010s.
@@TEDodd Early 90s.
That's absolutely awesome. It has practically everything a modern computer has: HD video... stereo audio... gaming... etc... all in a cool-looking gigantic space-opera/science-fiction setup with lots of blinking lights, too. ;-)
Mega-kudos to Lyle and other folks at CHM for getting that rig up and running beautifully.
I love how you turn this huge machine one with the tiniest button haha
Anticlimax at its finest.
That's why the old radar screens had a cone over it, so it would make the afterglow easier to see. That's why the original Star Trek show had the same rig, as a callback to those radar screens.
525Lines i worked with old CRT radars, (1970-80 systems) on ships, those had a cone that was needed for glare, not afterglow.
Man. This is a beautiful machine. I love the sound, the mechanics of it all. It's one of a kind.
Ex-DEC employee. I think that the first DEC system I used was a PDP-15 when I was a teenager and it was dedicated to running chess. I never knew that there was a PDP-1 though I did wonder about it. I used lots of PDP-8s and 11s, DECSystems, VAXen and Alpha systems. Great to see this video and nice to see the gentleman maintaining this hardware.
This has got to be the best quality video of Space War on RUclips today! I've always wanted to play it. Thanks for capturing and sharing here!
This is mind blowing. The people who created this ... I am just in awe of...
I wonder if they knew the impact on the future they were making
Of all the things that I consider amazing about this machine and the programs that were written for it, I am drawn to Spacewar! It had absolutely nothing to base itself off of. Nothing like it existed prior. The programmer had to create so much of it from scratch. A way to control two drawn objects with separate controls. A way for them to interact (shoot), the physics of how the ships should move in the space provided, the gravity star in the middle, the scoring system, collision detection, EVERYTHING. I think back to when powered flight was invented. Prior to that, thousands of people came up with some very wacky ideas for how to get an object to fly in the air. But 1 of those designs became the basis for everything afterward. Incredible.
Lyle does a fantastic job with this presentation. His enthusiasm is infectious!
This is soooooooooooooo ahead of its time! So cool! I cant beleive they could do this in the 50s! AMAZING!!!!
Fun fact; the National Socialist party of Germany had 3D videos of their leader's speeches nearly 70 years before 3D movies became a fad.
@@kana22693 3D films were also a fad in the 1980s and 1950s, while early 3D films were made in the 1920s.
They were so ahead of their time... or are we now behind the times! We take for granted how technology got to now.
@@rudolfrieder186 Catain Disillusion has a good mention of that. "Arrival of a train at La Ciotat" as a very famous film of the era, "And it was remade in 3D, 20 years later". Which is actually true. And as one can see, nothing changes.
A lot of the trouble was just the scale of the hardware, a lot of stuff was possible before it got miniaturized but took up entire rooms and thousands of dollars, now a laptop the size of a book can run VR games or crunch gigabytes of image data to make high detail images of galaxies thanks to stuffing orders of magnitude more processing hardware and memory.
Meanwhile a roomful of computing hardware now can handle more information processing and storage than could have been imagined in 1950, and we still haven't reached the level of processing power, storage density, or power efficiency of the human brain, though we are starting to apply lessons from it's architecture to improve AI and processing difficult to parse information
Spacewar was a really great game for it's time. I think it still holds up now.
Thats the first game I made on scratch when my 6th grade teacher had the whole coding class lmfao.
I swear theres something instinctual, borderline BIOLOGICAL behind our species love of computers.
Something about coding up something out of just information, and seeing it come to life, is just wild.
Most species have the natural perogative to eat and reproduce.
Humans add another one, which is build.
Eat, Build, Reproduce
surprised Nintendo didn't buy it
To think: This was the machine, and the game, that George Lucas saw at MIT in the day that inspired him to create THX 1128 and Star Wars, respectively.
this is so freaking amazing! I'm just spechless! It's so ahead of it's time and it is so interesting to see this machine working after nearly 60 Years!
I love when the guy interviewing the expert actually knows what he's talking about. It saves the expert having to dumb it down for the interviewer. Then you really get no useful information.
Wow, just wow! That old phosphor screen looks absolutely beautiful , I want to play Space War! 4k , quite amazing.
A real pleasure to see that one working.
DEC were so .... nice ... even on VMS & TRU64 generations :)
It is amazing to see the vision and implementation of this system so early in computer history
I was born in 1959, and I've seen a lot of videos, but this one is by far the most fascinating. The first term for a 'computer' that I remember learning was that of an 'electronic brain', which is what I would have called the PDP-1, had I seen it in, say 1966. Jaw-dropping stuff. Thank you.
(I wonder--what would it cost to build a perfect replica?)
There would be no reason to build a replica that uses 2000 watts of power and can't store anything on a hard disk. You could emulate the instruction set and display in software and it could be run on a smartphone.
@@acmefixer1 it can run on your calculator
@@acmefixer1 Tell that to the countless people 1000x smarter than you or me who have built exact replicas of the manchester baby, colossus ww2 code breaking computers, etc. There are PLENTY of reasons to build replicas, and thankfully, there are people not like you who are satisfied with emulators running on a phone...
In all honesty, those were some of the most interesting 22 minutes i´ve ever had on youtube. Thank you both for explaining all of this.
DEC. I still miss them. I just remember VMS (a little) and Digital Unix (True 64) on Alpha. They really made a great job. I remember having, flying on the wings of decadence, switched the shell for root on a Digital Unix Alpha )to /usr/local/bash. This worked quiet nice to the day, /usr/local could not be mounted - in the CRM i activated the singleusermode and the system defaulted to a statical linked shell in /sbin, IIRC.
THANK YOU, DEC!
You saved my newly begun job.
They had a great philosophy. A simple one: Perfectionism.
the song is Bach BWM 592, movement 3 @4:00
Awesome! I was trying to figure this out via Google searches.
Just what I wanted to know.
thanks
This should be in a public shrine where all us geeks can go and worship it. It is absolutely incredible to see the influence this machine had on Nolan Bushnell first hand. That's Asteroids/Gravitar's grandparent running right there. And that code on paper is the dead sea scrolls of gaming.
Thank you so much for this video. Seeing a machine in action that I read about in computer history books at the library while growing up was an amazing experience.
The sheer size of the shoulders of giants that ANYONE who uses a computer, smartphone or server stands on... The mind melts if you attempt to comprehend it. I am deeply humbled at the intellectual prowess of those before me who built the machines that propelled the human race into the information age, and am extremely fortunate to be living in such a time. This was in 1959! Oh my, This is amazing...
This was amazing to watch, thank you for taking me back in time!
brilliant video.
If I ever had the chance to visit the States, I have to visit the Computer History Museum in Mountain View, California. I will be at the door at 10AM, and I will be the last one to leave at 8PM.
And come back the next day :D
Ing. Max Koschuh One day won't be enough!
yes, I guess I could stay there a whole week
should not travel with a wife though,.... except a nerdy one
I read this will be made into a new compute institute? is there still stuff open for public?
And it will make your GAS even worse...
GAS = Gear Aqcuisition Syndrome.
I had no idea the PDP-1 display was bitmap. I always assumed it was vector based. Thanks Marc! Another awesome video!
I think it is more like a vector-pixel display. The PDP-1 sends out a stream of data which I believe is in fact vector like, but the points map to discrete locations (with a gap).
The display is actually a "point plot" display. You send it an x-coordinate, a y-coordinate and an intensify level and you get a single dot. 50us later you get to do it again.
That is correct: you give it two 10 bits numbers and it lights up one point of 1024 by 1024. If you light up two neighboring points it is possible to see a slight gap between them depending on the focus of the beam. You draw a line by lighting its individual points, so it looks like a bitmap unlike on a Vectrex videogame which has analog circuits to smoothly draw a line between two points. Despite this difference, it is correct to call both "vector displays".
I was wondering about that. Clearly there couldn't have been a frame buffer because that would represent more memory than the entire rest of the machine by a factor of ten or more.
I'm confused, is it raster or vector?
Excellent video. The efforts of the restoration team are impressive.
Such a beauty! Every wire connected by hand... it's an amazing piece of work.
I used to work for DEC, and we had one under maintenance at Chalk River. I got a week long training with one of the designers. Played Space War on it too. :) This was in 1977.
I loved this!!!! Thanks for sharing, I had no idea about this machine and its capabilities. You brought it to life for me... and thanks to Lyle.
This is awesome. Thanks for sharing and thanks to the Computer History Museum for keeping these machine 'alive' :)
Awesome that the PDP-1 is beautifully restored. Amazing work and machine.
I started watching this video because I wanted to know the shapes of the ships is Spacewar. Kept watching because the rest of the content was so FASCINATING!!
Thank you very much for the video. My very first contact with a computer was a DEC PDP-8 running SpaceWar in 1971 in the Chemistry department at Cornell.
My happiest programming was done on a PDP-11/45 in the next room. What I wouldn't give for another hour at that console. Many happy memories.
Sweet. What kind of programs did you write?
Marc, thanks for bringing this to us, and a shout out to Lyle: Thank-you for your tireless effort and expertise in keeping this beast and valuable part of computing history alive and well. Today, we can see this machine. 50 years from now, videos like this may be our only archive. As a prof in comp sci, it helps to see where we have come from to help guide where we are headed.
Incredible! Thank you Marc and Lyle! A real treat to watch this.
Incredible they could do that in the 1959 and 62. I would love to know how they knew or came up with how to program a game.
It is described in the book "Hackers: Heroes of the Computer Revolution" by Steven Levy from 1984. The book got updated and republished in 1994 and 2010 respectivly. There's also an audiobook version of the 2010 edition on audible.
5:55 they made THIS almost over 60 years ago!!
that machine is AMAZING! thanks for making this video tour!
Way cool! Thank you for showing the PDP-1 to all of us! I've only just seen it referred to one or two times in books. That system was way ahead of it's time.
Absolutely amazing, and Lyle seems like a really cool guy too :)
Smart guys with pony tails are always interesting.
Hard to imagine that Spacewar! would have been over 20 years old by the time 3d games like elite were being created. It boggles my mind.
awesome :) thanks a million for the tour
I like the passion the man talks about the computer
This would have blown my head off in 1959
The best about this amazing piece of great Slug Russell is that it is free for any addition to the code.The final version of the game contained features from great hackers like Peter Samson, Kotak, etc. Just Great Porgrammers they are
The spacewar graphics are actually nice!
Thank you and Lyle for the awesome demonstration of this beautiful piece!
That was amazing and eye-opening. Thanks for making this!
This piece of ancient computing is amazing.
A video game AND touchscreen in the 50s/60s - this thing was way ahead of its time!!!
That's alot of wait for loading! Lol wow how cool is that? 1952! I can't imagine how genius these people are, from scratch and now we're all enjoying the modern video games because of them, Thank You!
I visited MIT on an interview trip in 1971. I knew a few people there and one of them had a computer geek friend who was one of those obsessed with that PDP-1. He was a true hacker and I stayed up all night with him while he wrote and tested programs. It was too cool.
Simply amazing, thanks for give this knowledge to the new generations :)
Seeing assembly code gives me pleasure. There is nothing more beautiful in this world than ASM. It gives you a more intimate relationship with the machine and unlimited power.
Conversation is always better without an interpreter ;-)
This video will be valuable for the 1000 years. Great Job Lyle and Marc!
Seeing this for the first time. Mind absolutely blown.
Thank you for this video. It was the perfect lunch break.
I wonder, is the three-dots program deterministic or does it evolve differently each time?
This machine is amazing! I can't believe how sharp and clean that CRT still is too. Though I assume it's been replaced at some point?
The inside is also very clean, considering the amount of airflow combined with age.
To the best of our knowledge, the CRT is original. We did not replace it. The display, while relatively simple to program, was difficult to restore. It's complicated - analog stuff of this period often is ;)
BTW: We clean the entire system on a preventative maintenance schedule...
Thats not true I replaced the CRT in april of 1984 with a new old stock screen from DEC. Also did you replace that caps in that? Every armchair internet expert knows that and nothing else
Dave B how do we know you’re telling the truth?
Because I am lying.
Dave B ah! Fantastic...
What an incredible job those scientists and engineers did back in the day. Thanks for sharing.
I learned on a PDP-8 in 1974, so it's a rare treat to see its great^3 grandfather in action. Thanks!
P.S. The machine is only one year older than me.
Pretty cool how this was done in 1959. Ironic how little things has changed far as the basics go.
19:50 Odyssee 2001
"I'm Afraid Dave"
Wow. just enjoyed so much to be in this museum and actually see how PDP-1 works. The best part was to see SpaceWar getting loaded 13:00 and of course the Gameplay. Thank you so much for this detailed video. (thumbs up)
Thank you, Marc, for the look at DEC's first PDP. Everyone heard of the PDP-11 and VAX, but this is their ancestor. 👍👍
With any modern OS, it takes a minimum of 100 lines of code and 5 days of research on the internet to change the colour of a single pixel on screen. While back then in assembly it took only a single line of code.
100 lines of code? Here:
#include
int main(){ SetPixel(GetDC(0), 1000,1000, RGB(255,0,0)); }
That obviously didn't take me 5 days to figure out. I suggest you to check out the 4K PC demoscene, they challenge themselves to produce the most impressive audio/visual demonstrations while keeping the program size below 4096 bytes (including music, graphics, everything). That is even smaller than the 4 KW spacewar shown here if I understand correctly. There are a lot of reasons why we don't program like that today, but it can still be done and it is amazing what you can do with only 4 kilobyte.
@@Spillerrec
Most people today are software engineers and not programers.
Spillerec - If you hide everything in headers, every program only needs one line.
How many machine instructions do you think that SetPixel call takes?
@@manuell3505 it is a OS function, so as little as possible on Windows. It is not much, actually it is too little to be useful. Try running the code and observe what happens. What is the issue? If you actually is interested in why an OS makes it more complicated I can go more in depth of why you would appreciate that.
Spillerec - "OS functions" that consist of predefined software routines are technically not OS functions, but side-applications, as is SetPixel(), part of the Win API.
Only user-, RAM- and storage I/O need to be managed to make a computer useable.
The legendary machine.
Hats off to the engineer(s) who designed Space War.
They saw a giant hardwired computer system with an oscilloscope looking display and 4kb of memory and thought, “I could make a head-to-head space combat game centered in the gravity well of a star.”
No, this is not 4 KB. It is 9 KB.
Amazing video, thanks for sharing.
This is pretty amazing. As a software engineer, but a young one i've never seen one of these in person. I dream to one day write a program, punch it into tape and run it one of of these. That would be amazing to me.
Thanks for this.
Max Danielsson Likewise for me. I hope to see one of these giants in person one day. Perhaps even buy one in retirement as a /very/ expensive hobby ;)
I love this idea
Max Danielsson my first computer was the pdp-8e in my high school! It had mag tape and 4 teletype machines
@@swiftfox3461 There are only 3 PDP-1's known to exist currently.
I would highly recommend "structure and interpretation of computer programs" for any aspiring software engineer.
World first pre-digital camera, world first game one of the , world first music ozillo graphs, world first digital diagnostic debug-tool. High res bitmap screen, smart design for easy fix , lpu´s that would fit todays standards of handling. World first 8 bit speakers
omfg give a break. Makes me wonder what those guys all defined what we still see today as standard just by doing it and what holds the rest of humanity down from achieving such. We think we could never do so.We may have to think again
Are you having a stroke?
What a fantastic video. Thank you for making this.
Excellent job CuriousMarc. Thank you so much man
This thing can do more than Windows 10 S wow!
Has anyone found the Magic / More Magic switch yet?
Incredible, thanks for posting
This is so awesome, Thank you for sharing this
awesome to see this! thanks for making these.
Also, @CuriousMarc .. if you want to see what modern programmers do with 4k today, see this: ruclips.net/video/rML-KvlWk5s/видео.html
let me know if you want the exe file for this (which is 4096 bytes in size)
Wow - thanks! After a bit of sleuthing, it seems that this is an entry of the "4K Intro Compo" type for Windows, as described at en.wikipedia.org/wiki/Demoscene_compo and en.wikipedia.org/wiki/Demo_(computer_programming)
I have to protest, since 4k intros on Windows make full use of available frameworks, which are hundreds of Megabytes in size, and the millions of times faster CPUs 58 years later. In other words, in no way are modern programmers 250 million times better at programming. ;) Insiders will know why this Windows 4k is a good one over the other Windows 4ks.
4KB binary, hundreds of Megs of OS and graphics DLLs from Windows, not comparable with a machine that has at most 16KB.
Its only been on 3 years and two months? I would expect more considering I had a computer that was never shut off for four years. That video game is way better then atari seems like atari had more memory.
Nope. Atari 2600 had 128 bytes. This has about 12000 "words", and each of those words was more than twice the size of a byte. So the PDP-1 had roughly 200 times as much memory as the Atari 2600, despite being almost 20 years earlier!
I am completely blown away!
Waoo! really impressive ! That's really a piece of art, the hardware and software too... Thank you so much to keep this memory alive for future générations... Greetings from France
The pen is like 'GRRRRRRUUURURUURR!'
+kakureu Because we forgot to turn off the audio amp that is connected to the program flag lights used in the earlier music demo! The unintended effect is quite interesting though...
And for that it was like all 'GRRRRRRRURRRUUURR' ;) but still I was figuring that case :P thanks for confirming my suspicion. That place is one of my Todo lists if I ever find myself able to travel.
Just to help me out, is this less or more powerful than, let's say a Commodore 64? I'm a total noob regarding these REALLY old computers.
A bit hard to compare. The C64 had 64K bytes of 8 bit RAM plus 20K of ROM, and a ~1 MHz 8 bit CPU. The PDP1 used 12 K of 18 bit words, and the logic was capable of running at about 5 MHz, but I don't know how the instruction set compared with the C64 processor.
The PDP-1 required many kilowatts of power, while the C64 could run off a wall plug, and had floppy discs for data storage. The PDP-1 had about a 20 word program loader to bootstrap, but then you would have to load the operating system from paper tape, if it wasn't already in its non-volatile magnetic core memory.
Biggest difference is that only 53 PDP-1 units were ever made at a cost of nearly $1 million each in today's money, while 10 - 17 million C64s were sold at a much more affordable price.
Aside from the price and the difference in I/O capability, the PDP-1 was probably a more powerful computer, but only a handful of people could have afforded to buy or operate one.
less, much less,
Keep in mind the C64 had its custom chipset coprocessors for video (VIC2) and sound (SID). This is why the C64 is light years more powerful in graphics and sound. The Vic2 ran at 14mhz. The 6510 processor also was no slouch esp when unsaddled from the burden of driving the graphics system, able to run an instruction every 4 cycles vs 8 on the common Z80 processor of its time. The main limitation was the lack of memory.
Let's go point by point...
Memory (not counting ROM, as we don't know what, if any, the PDP has) -
PDP-1: 12K x 18 bits = 221,184 bits (non volatile)
C64: 64K x 8 bits = 524,288 bits (volatile)
The Commodore has about twice (to 2.5x) as much, at least once you subtract the ~8k needed for a typical screen bitmap, though you do have to sit through loading programs into it every single time you power on, unless you're running a cartridge of course. However, if you boot into the built-in BASIC, the load screen claims "38911 bytes free", which is equal to only 311,288 bits (because some of the address space is shared with ROM, and only some of the missing space can be shadowed by video RAM). A closer run thing, but in either case, the C64 still wins this round. Other contemporary machines like the Spectrum 48k and BBC Micro (32k, typically) would be even closer or actually lose. So we can say the PDP's memory is about equal to that of an early 80s home computer, albeit divided up somewhat more coarsely.
Bear in mind however this is just how much memory THIS PDP-1 has; the base model sported only 4Kw (4096 x 18 bits), and expansions were available up to 64Kw (65536 x 18 bits), which obviously changes the equation quite severely in both directions.
Processor power -
PDP-1: 18-bit architecture, parts rated for upto 5mhz switch rate, but only used at that speed for executing microcode at a level rather lower than the actual "instructions". Memory speed, a much more important thing for the operation speed of both this system and a typical 80s machine, was about 187khz, which at an average of 2 memory ops per instruction gives an overall speed of approx 93,500 18-bit instructions per second. As computer and specifically CPU architecture was in its infancy, each instruction was more primitive, less refined and took up far more microcode steps than in later microchip processors, hence the discrepancy in speed; also, with fewer opcodes available overall, especially for complex instructions, you'd have to use more of them to get the same result, making for an equivalent speed probably far below 90kips. Total transistor count about 2700, with a similar number of diodes. Able to access up to 64k x 18 bits of memory without paging, so 18/8ths that of the C64.
C64: 8-bit architecture 6510 running at approx 1mhz (a little higher for NTSC, a little lower for PAL), with somewhere north of 3500 transistors on the die. If in line with general 6502 system design, memory may run at the equivalent of about 500ns (=2mhz), more than fast enough to perform read-write cycles at up to 250khz equivalent (the CPU can only access memory on every other cycle, and reading then writing are of course separate acts), or about 2.67x that of the PDP-1 (give or take a little depending on region). In the C64, the CPU has to share memory access cycles with the video chip, and the full performance is only available when video output is disabled, but as this is basically the same as the PDP (which outputs to the teletype by default and uses the scope as an additional rather than main device) it's not worth working out the speed hit at this point. Obviously, only capable of reading a 64k x 8 bit block of memory without use of paging (cheated in by use of the GPIO pins added on to the 6510 vs the 6502, which selects between ROM and RAM, but that's still only about 96k to 128k x 8 bits of total volatile storage, still not as much as the PDP).
Advantage: Commodore, again by a factor of about 2 to 2.5x
Video output -
PDP-1: 1024x1024, 1bpp quasi-vector display; can display 20,000 points per second (1 per 50us), or the approx equivalent of 400 points per frame of a 50hz raster display (equivalent of about 8400 bits, just over 1kb, or 467 words of storage, given 20 bits for X/Y location and 1 for on/off). Long-persistence phosphor of course helps to cut down on flicker when showing more than 400 discrete points/vector nodes, but there'll still be some obvious pulsing, and a full 1024x1024 raster image would be impractical as it would scan barely any faster than one frame per *minute* (and would need 128kb / 57kw of storage anyway). Readable character output therefore rather limited to maybe a few lines of 16 to 20 chars apiece, and large amounts of text are better output to the teletype with the scope reserved for graphics instead.
C64: 320x200 pixels or equivalent sprite resolution at 1bpp (1 colour + transparent for sprites, 2 colours per 8x8 block for background), and 40x25 text with 2 colours per character space from a limited set (as there are only 4096 bits of colour RAM available), or 160x200/equiv at 2bpp (3 colour + transparent for sprites, 3 ink colours + shared BG colour per text char or 8x8 pixel block, or 2 colours per half character/4x8 block). Effectively about 8kb of video memory in bitmap mode, plus a little extra for colour RAM and a small number of 8x8 sprites.
Advantage very strongly Commodore in this case, even though it can't match the same peak resolution as the PDP's scope, it can make much better use of the lower total resolution available to the VIC chip (limited essentially by its output device, ie a normal TV, which in NTSC regions can't be relied on to provide much more than 200 visible lines in progressive scan mode, or a visible width of more than about 320 pixels at the VIC's higher, NTSC colour crystal based clock rate of 7.14mhz - which is why that same resolution, merely with much better colour depth, was retained for Commodore's own Amiga (with a whole-system speed of 7.14mhz), and the Atari ST that was designed by some of the same team responsible for the C64, as well as various other computers and consoles), showing much more complex figures without flickering, a full 1000-character screen of text, and of course upto 16 different colours instead of just on/off.
Sound output -
PDP-1: 4 channels of square/pulsewave using a somewhat edgy hack running four of the processor status lines out through an amplifier. Somewhat heavy on the CPU, can't be easily used with other programs, no waveform or per-channel volume control, no built in white noise, enveloping, filtering etc.
C64: Legendary SID chip... only 3 channels, so ultimate harmonic sophistication slightly less, but each can be one of four waveforms (square, sine, triangle, noise) and combinations of them, play at different volumes, use fancy ADSR enveloping, filtering and modulation effects, etc, uses very little CPU time to play simple tunes (a few dozen register updates per second, with the chip running automatically in-between), and can be easily used as an accessory to all manner of programs without suffering interference from the CPU status.
Commodore wins that one with no trouble at all, unless you have a specific need to play four different basic notes without any fancy processor trickery to split the job between three.
Offline storage -
PDP-1: original storage by teletype grade paper tape, base format uses 8 bits at a time (ie 8 holes across, not including the timing holes) but the spacewar tape looks like it's only using 6 of them; so, each word is 3 rows on the tape... stated data rate for it is "60 characters per second", which I'm guessing means 60 groups of 3 - so 60 words, 180 rows, and 1080 bits per second, not particularly fast but enough to read a 4k program in a little over a minute - given that pictures of it look to have about 80 rows per fold (same as punch cards), and the reader is blowing through it at more than 2 folds per second (so 4k = a folded stack of paper equivalent to about 150 pages, like a book of raffle tickets). Impossible to edit in any way other than blanking out entire characters by punching all their holes - if a paper tape needs edited, you read it into the computer, edit it in memory, than output a whole new one, same as making a new working copy when the original starts to get worn out (after about 25 reads...); commercial software that was expected to be read in hundreds rather than dozens of times might come on Mylar plastic punch tape instead, but that was merely more durable rather than faster or higher capacity.
Later (late 60s, vs PDP range going into series production early 60s) joined by DECtape, storing about 2700 kbits per tape (330 kbytes, 146 kwords) and able to read a full tape in about 30 seconds; effective data rate of ~80 kbit = 10 kbyte or 4.5 kword per second, not much by modern standards but blazing fast compared to papertape (would read Spacewar in less than a second), and the amount of memory in the computer. Access speed about 125ms from a dead stop to running at full read speed (system able to read data asynchronously at whatever speed the tape is actually moving at, so no need to back up and resync even though each data block is around 256 words, and so takes less than 60ms at full speed, or probably 120ms when pulsing the motors to read a single block), although random access obviously averages more like 15,000ms... Accessible like a floppy or hard disc, however, as far as the OS is concerned, with a subprocessor looking after block seek etc, and a secondary drive with a blank tape can even be used as a small block of virtual memory for a multi-user timeshare system.
Hard drives, drum memory, punchcard readers/writers and ultimately floppies also available for the system, but were mostly later and far more expensive developments (justifiable by higher capacity and speed, or compatibility with e.g. IBM card stacks), whilst the DECtape was rather more affordable, easily expandable and ultimately more or less ubiquitous.
Of course, initial bootloaders tended to have to be toggled in through the front panel switches, which might take a couple of minutes of manual operation to input a few dozen one-word opcodes and addresses. ROM modules (which may just be circuit cards with a particular pattern of diodes) were presumably an option but I haven't seen any information anywhere...
(out of space, ironically)
C64: Comes with a healthy 20kb (=100kbit) of autobooting ROM including BASIC for messing about, and a proper keyboard of course for using the BASIC... simple mass storage achieved using the tape deck and regular audio cassettes, at a rather slow speed (about 600 bit/sec, or 60 byte/sec with regular 10 bit/byte tape encoding = 3.5k per minute, or maybe 100k per side on a 60 minute tape with allowance for leader/sync tones and gaps between data blocks... so, slower than the PDP paper tape) using the default routines, or higher speeds (1200 to as much as 6000 bit/sec, so loading a 32kb program in 1 to 5 minutes instead of 9+ minutes) using custom fastloader routines. Automatic controllability limited to turning the drive motor on and off, any fast forwards or backwards seeking (using the analogue counter as an index) having to be effected manually. Writing only possible sequentially, and to a completely blank section of tape unless you want to overwrite the previous contents.
Upgrade mostly in the form of the 5.25" floppy disc drives available for the system, typically Commodore's own 1541 or similar - themselves rather slow and low-density compared to competing systems, but giving a definite payoff in terms of random access, particularly for writing, as they're formatted into discrete blocks. Standard spec of 170kb per disc (or 340k if modified to be flipped over for sequential double-sided use, but still only 170 at once), ie about 1400 kbits including filesystem area, and a peak speed of only 400 bytes (=3200 bits) per second. However, as with the tapes, this could be reprogrammed and a little extra capacity squeezed out, and again upto a 10x speed boost, so around 4kbyte / 32kbit per second.
Also there were ROM cartridges, which operated essentially at the same speed as main memory and whose programs were available directly after power-on... capacity was limited however (16kbyte) and of course they couldn't be written to.
Essentially, in this case, the PDP takes it quite comfortably, as the C64's basic cassette RW speed is even slower than that of the paper tape reader and is about as limited in terms of random access and rewritability, and even when the disc drive is juiced up to the max it can't match the speed of DECtape, the random access speed advantage is completely lost if the disc drive runs at its default speed (even though the tape may be seeking far more often than it's reading or writing, it does the latter so fast that it still catches up even to a disc that's running with entirely sequential RW), and the total useful capacity of a tape is about equal to that of a disc that's been modified for flip-side use, but the whole 2.7mbits can be accessed in one go and with barely more time needed for the entire operation than it would take to park the floppy heads, eject, flip, reinsert, and start loading side B.
Other things -
I think we've gone on quite long enough, don't you? The C64 wins in most cases, but the PDP is clearly the better option for processing large amounts of data programmatically... ;)
what a nice guy demoing this remarkable machine,very interesting video,Thank you.
Many years later, I was learning about debugging on the DecSystem 10. The program was called DDT. On the first page, the origin of the name was explained with a footnote stating that it should not be confused with the insecticide of the same name. The note went on to dryly note that both were used for elimination of bugs, though of, it was hoped, mutually exclusive classes.
Unfortunately there has been some historical crossover - roaches and vacuum tubes do not mix well.
I wrote DDT for the PDP-6 in 1964 (instead of going to my MIT classes). The note you refer to was created by the contract tech writer, Bill English, who wrote the assembly language programming manual. The note was kept in the manual over the nearly-dead bodies of the in-house tech writers who regarded it as "unprofessional".
@@thomasw.eggers4303 Thank you for that slice of history. I've always treasured that comment - it still tickles me.
@@nicolek4076 Thank you both for reminding me of the giggling joy I had back in 1973 while tracing for a suspected bug in the assembly language code of a GE Datanet-30 Front-End Processor for a GE 635 Mainframe computer. As I scanned the source code listing of the parsing logic that determined what type of remote system was to be interfaced, I found these hilarious comments: "Hippity hoppity, here come Big Blue and the Seven Dwarfs". Then followed code sections for IBM, GE, CDC, DEC, RCA, NCR, Burroughs, and Univac. Happily, no one at GE Pheonix had removed the comments as "unprofessional".
Spacewar is 100 times more fun than Pong!
Actually, real Pong is a lot of fun, but what many people think of as Pong was an imitation by General Instruments. Real Pong had four upward angles and four downward angles, but the GI chip only does two of each. The difference between having two angles and four might not seem like much visually, but it makes a huge difference to playability.