I'm not surprised that he said that. ARM was developed by people who were very much connected with the 6502 world and had been involved with the various 8-bit Acorn machines. In fact the ARM 1 was developed using BBC Micros and their Tube adapters. So you could say that the 6502's influence still impacts the computing world to this day, even if you ignore the retro world!
Yes, and I‘ve noticed that at least one of the dev teams I’ve got to know for medical grade hardware use a 6502 core instead of more modern ones in FPGA as they‘re basically feeding it the same software they did decades ago. Can make validation simpler.
The 6502 is sure historically important and influential, but you can't see anything of it in the ARM design really. (Yes, I have studied both instruction sets, along with many others.)
Dang, that interview was strange. The video feels like it is missing the part of why the chip was designed in the first place and why it had such wide proliferation.
My grandfather was the CTO of Zilog and raised me. He helped develop the Z80. I am very proud of him. He made life unrealistic for me.. he raised me and spent all his free time with me building PCs and listening to whatever music I found. I’ve never found anyone who holds a candle to him. Not having him around has been tough. Knowing his work won’t ever die in some capacity makes me happy. Not this IC.. just in general with his work. He also helped GE in the 60s build their first main frame. Dude was the most modest man you’d ever meet. Edit: I’m not going to put my families personal info out there, had no idea this comment would get any traction. Believe me if you want or don’t.
The TRS-80 Model I was my first home computer, and the first platform I learned assembly language on. The Z-80 will always have a special place of honor in my vintage-computing collection. 🙂
interesting my grandpa on my dad's side couldn't even find the on button on a pc and refused to use them .my grandpa on my moms side knew a lot about analogue electronics like crt tv's and vcr's but when it came to computers i didn't have one until 2008 and kept getting malware .
At 4:38 that's me closing the VCR lid in the video "VHS VCRs Revisited" ... Big fan and honored to be mentioned/shown in an 8-bit-guy video! (First programming experience was BASIC on the Apple //e on a 6502 ...)
"6502's, cases, RAM, you want it? It's yours my friend, as long as you have enough money. Sorry fuzzy7644, can't give tickets to the retro PC store. Come back when you're a little, MMMMMMMM richer." -What you probably thought he would've said
A few thoughts... 1. I feel this only scratches the surface. I get that this must be a monster project, but I'd watch a 2 hour version of this going into endless detail. Or a 4 part series even. The 6502 strikes me as one of the most important CPUs in history, it deserves ALL the effort and detail. 2. The segment with Bill Mensch was very short, perhaps there wasn't much to add, or he couldn't share anything new, but I wish there had been more. 3. The Z80 needs a similar video, as does the 6800/68000, as does the 8086
You got some solid points, but The 8-Bit Guy already made several videos of consoles and computers that were powered by the 6502 and he did go into fair details of the chip scattered across them. I hope David covers the Z80. I own a 20mhz variant and I would love to learn more about it.
Thanks so much for the informative videos. Your channel inspired me to transition into IT. Two years later, I’m working full time in the industry and couldn’t be happier! Your channel is the best.
6502 is a canonical piece of computer hardware, similar to classic game consoles. They will keep getting made, emulated and implemented in FPGA forever.
the 8-bit guys is correct about how cpu speed does not compare well against different cpu's because amd's cpu's where slower then the pentium 4's but could seriously match or even outperform faster pentium4's so the 6502 is not the only cpu that proves this issue it works even with modern computers cpu's
I wouldn't be so sure, machines are getting ridiculously complex, and I doubt you're going to be implementing RDNA GPUs with unified memory in an FPGA anytime soon. Let alone accurately. 😅
@@SaraMorgan-ym6ueRight, however on "modern" hardware (so even 20 years ago) the clockspeed was just a tiny bit of the puzzle. The size and usage of the cache was in most cases much more important. CPUs have reached a complexity level that makes it almost impossible to predict the performance ahead of time. They have dynamic pipeline optimisation stuff going on that's just wild. In the good old day you just pick up an instruction set sheet, look up the clock cycles of each instruction, add them up and you get a fairly good estimate how long your code would take to run. This isn't the case anymore.
@@Bunny99s well give it credit for surviving this long since no other cpu's have lasted as long as the 65c02 and 65c816 they are in it for the long haul plain and simple (for being manufactured this long) I mean lasting this long for manufacturing life all other cpu's are discontinued
My first computer was in college in 1978: the MEK6800D2 development kit. I programmed a prototype for an automotive MPG display and a Times Square marquee, each in 512 bytes. Programming was all in assembly and hexadecimal. I would measure the frequency of the odometer and fuel sensor, divide distance pulses per second by fuel pulses per second and converting that to MPG units all using bit shifting. Storage was Kansas City standard audio tapes. My productivity improved immensely when I learned how to use the 6800 assembler on the university mainframe. I was unaware of the history of the 6502 and its connection to the 6800 at the time. It is great to hear from the people like Bill Mensch, who were hands on with some these things that are now household words. Thousands of pacemaker users are happy to know they are using some of the most tried and true processors there is.
My Brother owns an Original ATARI 2600, and what is left of the ACER 486 S/X PC we had for Years back in the 1980's and some of the Early 90's. I own The NES - and Many NES games! one for sure is quite rare and I can not even find any info on the game. I think it was a prototype/Unmissed game!
I first programmed the 6502 on a synthesizer in 1979. A small music company in Oklahoma City called PAiA Electronics (still in business!) used a 6502 for computer control of a modular synthesizer in 1970's. John Simonton (en.wikipedia.org/wiki/John_Simonton ) was the designer. A fun fact is Larry Fast of Synergy used the Apple II to play some of Simonton's software for the PAiA platform. Fast played with Peter Gabriel and relied on Simonton for tour support on the first tour through Texas in 1977. So many electronic music performers are very familiar with the Rockwell 6502.
@davidryle Do you remember a guy named Hal Chamberlin? He wrote a killer book called Musical Applications of Microprocessors. It was my “Bible” in 1980.
@@GodmanchesterGoblin I met Synth Legend Roger Powell about that time. He was working with an Apple 2 with a Mountain Hardware 16 channel A/D-D/A converter board in it. He used to to handle control voltages and triggers for his Moog and Serge modules. Cool stuff! I think he used it on his second solo album Air Pocket.
Fascinating David I learnt a lot and I had no idea that the 6502 was still being made and it makes me happy that it is. Back in the day I had a Commodore 64 and Commodore Plus 4 at home and used BBC Computers at school, all 6502 based and used to write programs in basic for them all so I understand about memory access and some of the CPU instructions. But what I never really understood is how it all fitted together. You turned on and there was a BASIC prompt and a flashing cursor but it all seemed a bit magic as to how that all happened. I never knew about how memory was addressed and stuff like chip select lines. Between yours and Adrian's channel I have learnt so much over the last few years and now understand much better how the computers of my youth actually worked. Really enjoyed the video, thanks!
Thank you! This was one of the most interesting 8-Bit Guy videos I watched in a long time. The only thing is that it kind of left me longing for more. Also, I expected a bit more of the Bill Mensch interview. I appreciate the effort that goes into making a video like this, but if it was up to me it could be at least twice as long without being boring. Would love to see more of this!
Great video! Fun fact: The 6502 is essentially a simplified Motorola 6800 implemented using NMOS with a different pinout and a slightly modified instruction set. That's because most of its engineers came from the Motorola 6800 team!
Those simplifications were super important though: 6800 was eye-meltingly expensive at the time and the 6502 was cheap enough to practically sell by the barrel!
@@talideoni think it was $25 vs $125. I have the feeling that production cost was not the only reason why the motorola was so expensive. Misunderstanding the market costed motorola to lose the industry
Fun fact:(s) Similar to the 6800, the Data General Nova (1969) had two accumulators and two index registers and a zero page! But historians and Chuck Peddle himself (who worked on both the 6800 and 6502) said the 6800 was inspired by the PDP-11 although from a modern perspective the PDP-11 is a lot closer to the 68000. However when you are coding the Motorola processors (6800, 6809 or 68000) vs the MOS Technology 6502 it is apparent that the C style structs are prefered by the addressing mode design and that does have PDP-11 roots. On the 6502 you end up wanting structure of arrays.
Yeah, as an electrical engineering student, I cut my teeth on assembly on the 6800. I guess the college must have gotten a deal going with Motorola so they could “hook” us early. We soon moved to the TI TMS9900. And holy cow that thing was 16 bit! Wow. …lol
You got tantalising close to giving away a key reason why the 6502 was so popular, and that's that it leaves the bus available for video half the time in a predictable way. By luck (or judgement?) one clock cycle equalled one character time on the CRT, so one half of the cycle the CPU guzzled data, and the other half was left to video. Essentially it gave way to an early form of DMA. This wouldn't work on a Z80 as it doesn't have a predictable bus, so instead you have to implement other segregated ways for video.
It was just as easy to implement memory mapped graphics in Z80 machines. All clones of Sinclair Spectrum implemented it with a handful of ttl logic chips. Overall the complexity and performance were very similar.
In 1979 a I bought 6502 based kit computer called the Tangerine Microtan 65 . Every single component had to be soldered by hand onto the circuit-board etc.. I had never soldered before and was learning as I went. It worked the first time I switched it on; to my utter astonishment. Fondly remembered.
1:22 I wouldn't be surprised if there's cars running around with 6502s under their dashboard. My 85 F150 is one of the last vehicles Ford ever sold with a carb, but it had an Intel 8061 under the dash in the primitive EEC-IV system. It would be shocking if there ISNT a ton of cars rolling around with 6502s running their primitive EFI setups/last gen carb setups.
Ditto, I drive an '84 F150, also with the same 8061 EEC-IV feedback carb (300 I6). They only offered feedback carbs for 3 years (300 and 351 84-86, 302 84-85). A bit of a weird technology, but for an eccentric like me, it fits!
You wouldn't be wrong... GM actually commissioned a 6800 variant for their earliest engine control modules in the early '80s! (the 6502 began as a low-cost version of the 6800 after all)
There are a lot of cars using 8051-compatible processors for various tasks. In recent times, most of those have moved to ARM, but, possibly some simple tasks like seat adjustment might be done with a specialized motor controller running 8051 code.
10:50 The TurboGrafx/PC Engine's HuC6280 (A variant of the 65SC02) runs at 7.16 MHz & in 1987 may have been one of the fastest 6502 variants. You mentioned the SNES and Apple IIgs use the 65c816 a "16 bit" variation which may make them a bad benchmark. Where the TG16/PCE CPU is closer to the 8 bit 6502. 11:35 Although the TG16/PCE had a few RAM expansion HuCards used by the CD-ROM attachment, there were no cartridge based enhancement chips ( 11:11 ) such as those used in SNES games. The TG16/PCE 65SC02 does make use of a video controller (HuC6270) and a color encoder (HuC6260) which certainly do some heavy lifting, but I still think the TG16/PCE would be a very good showcase of what the 6502 can do. Oh, and there is a mouse available if someone decided to port some games over.
This statement is a bit deceptive. The HuC6280 had an instruction set extension for copying memory to / from its video controller, which was far faster than doing it in software with the regular LDA/STA instructions.
That video co-processor did a lot of the heavy lifting on the PC Engine, much like most consoles of the time. The same even went for the C64. Things such as scrolling the screen, throwing sprites around the screen and collision detection, all of which could be very CPU intensive, were handled by these chips allowing the CPU to handle other things. That's why if you look at C64 versions of games vs. a Spectrum or an Amstrad version, no matter the quality of the gameplay itself(I can tell you from experience some Spectrum versions play better than their C64 versions), the C64 would typically have a smoother framerate and better scrolling. The only games I can think of where that isn't true is games that use 3D rendering.
I got my game dev career start on the Apple II, back in 1983 and coded the 6502 on the Apple, Vic-20, C64, Atari 8's, etc. until moving to 68000 on the ST, Amiga and Sega Genesis. 40 years later and I kind of miss the 6502 and regardless of the struggle and limitations, there was something really "fun" about getting things working back then vs. now.
@@Okurka.oh here we go... Well, I've coded several machines over my 40 year career. I guess I could get into the minutia of every specific variant, but I was being general. But I guess I should be more pedantic, considering the video and likely audience. 🙄
I love the way you do nitty technical deep dives. Your ability to take incredibly complicated subjects and explain them in such a way that they are understandable withought omitting important information is truly awe inspiring.
I skipped that and wrote an 8086 emulator for MIPS and ARM 25 years ago so you could run PC programs on those processors for Windows CE. Found an undocumented feature of the 8086 on the way that MS-DOS was dependent on too.
David, this video was AWESOME I learnd soo much, as always thanks a ton 🙏 from the guns to the keys and cpus the sheer qualify of you videos is admirable! Thanks!
I have been a subscriber for a couple of years and come here every so often when I'm on RUclips . I have many other channels that I watch for different topics but I always enjoy spending time here as in this video. I remember all the hardware that used the 6502 back then so it's like a trip down memory lane. You put your videos together in such a way that makes them very enjoyable to watch and this one is no exception. Thanks a lot for this and please keep giving us the great content that you have been providing.
Another awesome video! Thank you! I'd like to hear more of the interview with Bill Mensch and what he had to say about the 6502 architecture and the direction it might have gone.
Wait. The damn furby is running on a 6502?! That's the kind of random factoid that I live for. The 6502 and Z80 are immortal in a way. These ancient chips from the 70s keep finding uses. You've listed many of the uses of the 6502. The Z80 also has some fun ones. Maybe you all remember those old MP3 players, the ones that were basically USB sticks with a battery and headphone jack. Those things use a Z80. It's fast enough to decode an MP3 as long as it doesn't need to do anything else.
There might be a Z80 core in some mp3 players for controlling them, but no way that thing does the mp3 decoding. That is done in a hardware decoder on the same chip.
I'm glad we are all on this journey together 8bit guy has taken us a long way from, vintage PCs to now ancient classic computer architecture In a way, I feel like we finally arrived. If this was the intro to vintage computing I had at the start of your channel, I can say I would be able to follow you. But, here we are! I feel like we are getting closer, if not already, at the heart of vintage computing. I think, the future is bright. People can now take the first steps in the right direction, of where we need to go if we are to grasp the future of computing while still connecting to the real legacy of computer technology and architecture!!!!
1977 - the year I moved to Colombia. 1980 - the year I wrote my first computer program. It was written in 6502 assembly on an Apple II. Great memories!
This was informative and approachable. Nice work, 8BG! Do you have any plans to release the full interview with Bill Mensch? That segment was shorter than I was hoping it would be.
Love your videos!! This is a great video I never realized that modern technology still uses the 6502 It really powers the whole world! And all because some guys wanted to make a copy of the Motorola 6800
Well I'm a big 6502 fan but the world is probably powered by ARM. Which is not bad because at the end of the day ARM was designed by Acorn which previously used a 6502 but then needed something better which still doesn't use much power, and the result was ARM. There is an excellent video by the channel www.youtube.com/@LowSpecGamer about the history of ARM; he also has very good videos on the history of the 6502 and the Z80.
Wow, I knew the NES used a 6502 but I didn't know about the furby and Tama gotcha, that chip powered my '90s childhood. Perhaps the 6502 was also in the "digivice" that had a pedometer and you walked or shook the thing to get your Digimon to walk and play the game.
I had a 8080 at home, self build, with 2MHz, at school we have a CBM3032 with a 6502, 1MHz. I created two almost identical programs in assembler, calculating the 92 (or were it 72?) ways to put 8 draughts on a checker-board. One on my 8080, the other on the 6502. And the calculation-time for this was the same on both machines. 4 Seconds for all solutions, a Pascal-Program on the CBM needed 30 minutes. The 6502 was almost a RISC-processor, all commands were done much faster.
I grew up with the 6502. Nostalgia makes me love it. Its abysmal registers brings me back to reality. I'd like to see a deeper dive. It was a transformative chip.
First time I got to use a PC in real life was a OSI C1p and 41P at 12 years old in the late seventies. I still remember the first words OSI 6502 BASIC VERSION 3.2 (c) MS etc. The MS part doomed me for life lol
@@maxxdahl6062 Similar. The Commodore also had a more advanced color display as well as the advanced sound mapping, but even the base code had subtle differences. Apple II, Amstrad etc had subtle differences. There is a page somewhere in wiki that has a list of the different 6502 BASIC variants. From the OSI BASIC start I can code in any form OF BASIC ever created. I Often use FreeBASIC beside C and Python.
Once you realise all the zero page instructions are there to provide loads of “registers” that can be used in lots of flexible ways, the 6502 is actually replete with registers ;)
The only CPU I learned to program in Assembly language!!!!!! I’ve got a soft spot in my heart for the 6502, and I’m amazed it is still alive and kicking after all this time!
Hello David. Since I am so broke, I can't really afford Patreon, so this is the only place I can say it, but I wanted to express my deep thankfulness for you, LGR, and Techmoan changing my life. I have really wanted to meet you all, and get to know you a little better, and maybe send you something from the goodness of my heart, just to say thanks for getting me to where I am now. I am a sixteen-year-old nerd who from you three got really interested in old tech and began collecting old software, computers, typewriters, hi-fi equipment, etc. At this time I am now thinking about making my first video, but I haven't thought about it. I wanted to come on here and tell you guys I love all of you. I love watching your videos and learning new things. Always learning new things and weirding out teachers at school because they dont expect someone like me talking about these. I am talking about things they grew up with (lol)! And it makes me happy that I am now so interested in something I have spent literal years to dive into. Sounds weird to say but I enjoy growing up with you and the other two's videos. It makes me proud. Anyway, that's all I have right now. Thank you! ❤📼💾
I appreciate these guys’ work too. But if you want real, eternal fulfilment I highly recommend you turn to the Lord Jesus. Read one of the Gospels today, perhaps. God bless. 🙏
1:58 Thank you for including the OSI 4P. Warms my heart to see that :) 7:40 Thank your for the excellent illustrations so far. I am hoping to replicate some of this with the Atmel mega328 some time soon. The bus CS I recognized instantly, but I was curious about how they were getting the 16-bit pointers and integers. 25:16 thanks for the presentation. That was awesome :)
Great Video! Last year I actually did a project where I took a W65C02S, and connected it to a microcontroller, and monitored its inputs and outputs. I used the microcontroller's memory space as the addressable memory that the 65C02S used, and included peek and poke commands to monitor the memory addresses, as well as a clock step functionality and breakpoint functionality. I even wrote a disassembler for 6502 Assembly to go along with it. This was a fun project that i taught myself 6502 assembly with. It was a very cool experience. I'm now thinking of designing my own side project device around operating on the W65C02S.
I remember reading a book about the history of MOS, in one chapter which oddly you can hardly find any reference to online, it talks about the toxic chemicals used in the creation of the masks, the fumes from which would turn paper yellow. In 1974 the MOS plant in Audubon was listed as a hazardous waste site due to a leak of trichloroethylene (TCE) which caused local groundwater contamination. It stood out to me as I knew people who worked in the area who ended up with a profound life changing chemical sensitivity. While the 6502 may have powered a generation, it's creation may also have damaged one. I just think it's important that those things are not forgotten too.
Wikipedia does mention the leak, and the factory becoming an EPA Superfund site just as Commodore was in the process of taking over. Sounds like MOS hadn't informed the new owners and left them holding the bag. It doesn't say what happened after that; if I were in Jack Tramiel's shoes, I'd have sued the former owners into the middle ages!
That's life in the 70's! I grew up where a chrome plant dumped their plating chemicals into the town aquifer to help boost testicular cancer rates. Probably why the EPA came about in the 70's, too.
Did they ever want to dispose of TCE properly, or just pay the shareholders and then go out of business? I don’t understand why everyone wants humans in a manufacturing or chemical plant. I also kinda hate the ISS.
I did electrical and electronic engineering at college in the late 80's/ early 90's, and we were taught about computer architecture and assembly language using a 6502 dev board. Really interesting and loved it. Although I cant remember much of the assembly language now.
My first experience with programming 8 bit CPUs was with the Signetics 2650. I built a system based on an Electronics Australia project based on the S100 bus. It had 4k of RAM and was clocked at 2Mhz. The system could be expanded as the S100 was already a standard for expansion cards. I never hear anyone discussing this CPU. I later switch to the 6502 (6510) when the C64 was released. A much easier instruction set to use.
The PCEngine/TG16 ran a HuC6280 CPU, which is an upgraded CMOS version of the 6502. It ran at 7.18mhz and was released in 1987. I believe that was the fastest clocked 6502 based CPU of the time period you listed in the chart that has the SNES and Apple IIc Plus.
I think that it is weird that the cartridge also ran at that clock. Hudson added an instruction to burst copy data to VRAM ( and RAM). So level loading was fast and you can spend the splash screen on decompression. Or stream a level. Only problem: all operands are immediate. So jump indirect or self modify code.
Did it really run at 7.18 MHz, or did the system insert waitstates? The NES and SNES have a system clock at 21.47{72} MHz (6 times the NTSC color subcarrier frequency), but that doesn't mean that anything else in the system runs that fast. The graphics chips were the fastest at 5.3693{18} MHz. Also, the SEGA Genesis had a system clock of 53.693{18} MHz (15 x color subcarrier), with various dividers for the components. This is where MIPS benchmarks become more useful...
@@shinyhappyrem8728 Fanboys assured me that it does run at this clock. Remember: that was in 1987. Even the first draft/prototype of any circuit made in a NEC fab would run at that clock.
@@shinyhappyrem8728 You can easily tell the way games look. Most SNES games depend heavily on "eye candy": rotating, scaling, particle effects. But underneath, not much happens, as the CPU is weak. A shmup on the PC Engine easily has twice the bullets and enemies, and all the calculations for the positions are CPU work. Weird enough, even the NES is able to handle more complex patterns of objects than the SNES can. Many games that came out on the NES and SNES, are "emptier" on the SNES. 2 or 3 less enemies, fewer bullits on screen, stuff like that.
The 6502 was a legendary CPU is so many ways. In a way I think it's use just about everywhere well into the 80s hurt the 65C816's chances of success since it seemed like there was no urgency in designing a successor IMO. The 65C816 it wasn't available until 1985 and it's chief competitor was the 68000 was introduced 6 years earlier. The 68000 was comparatively much more expensive than a 6502 but once it's price started to decline in the early 80s interest in using it went up dramatically! If MOS or WDC had started working on what became the 65C816 years earlier maybe it could have scored wins in devices that ended up using the 68000. Also, I can only imagine what the SNES could have been capable of if it would have come standard with the version of the 65C816 in the SA1 chips that was used in a few dozen games. The SA1 ran at 10.74 MHz which is 3x the 3.58MHz the CPU in the stock SNES ran at.
CPU's back then: "I was built for an arcarde maschine, but can also run your car, home control system and lunar rovers as well!" CPU's today: "I am sorry, but you cant swap me from an identical Maschine, since our serial numbers dont match. Please go to an authorized store that forces you to buy an entirely new motherboard assembly thats like twice as expensive than a new computer"
Thank you for the sidebar into multiplication with rotate and/or lookup tables, it was really elegantly explained. I gained a new appreciation for how the 6502 instruction set allows for a clever programmer (or compiler) to do multiplication in just a handful of clock cycles. I had previously assumed you were just stuck with "a series of additions" (with large-number multiplications just taking an absurd number of cycles) or "paying for die space with extra-fancy chip circuitry", with no middle ground.
Wow. I am impressed to hear about Ben. His videos are great and an excellent supplement to Davids about older computer. Bens videos explain in detail how things in an 8 bit computer work, because he is building on in front of our eyes.
The big idea behind ARM was to exploit the memory bandwidth to the max. When they evaluated all the current 16 bits CPUS (80285, 16032, 68000, 65816) they were appalled by their inefficient use of the memory bandwidth.
As did the 6502, with the rising / falling edge CLK logic almost doubling memory bandwidth (as i believe a video somewhere on RUclips mentions). Acorn's MEMC 1 / 1A may have been more efficient, with a near single clock cycle memory accesses possible, when using the read / store multiple instructions (to push / pull fifteen of the twenty seven, 32 bit registers, as a block, taking 18 clk cycles = 3.33 bytes per 8MHz clock tick, rather than the standard 2 clk ticks per single ARM read / store op = 2 bytes per clock tick, let alone the 4 clk ticks needed on the like of a 68000 to read a 32bit word = one bytes per tick. Memory access was 2 to 3.33 times faster on an Acorn Archimedes, at the same clock speed, than a 68000 based Amiga, ST, Mac, Sun One, ..., the Archimedes were also marginally faster clocked than the PAL variants of the above), making for some quick stack accessing wizardry on a branch or return, along with some weird, 7-bit, reverse page logic, initially eliminating the need for much of the slow blocking logic found in the CISC toys. Though only 22 bits (4 MB) of the address bus went anywhere (without a multi MEMC and motherboard fudge), and the VLSI logic / DRAM couldn't be clocked above 8MHz, for the first six years of ARM chippery. Till the ARM 610 appeared, you couldn't source a true 32bit ARM SoC.
@@galier2I remember reading somewhere around the time ARM 1 came out, that they tested the other CPUs using the TUBE on the BBC micro - because that allowed them to separate specific aspects of performance which exposed the issues. I held on to that idea that the BBC micro has that important role in the dev of ARM which is now in billions of devices - but I heard something more recently that cast doubt on that. I’d love to have confirmation either way :)
@@sputukgmail The BBC micro was developed by Acorn the same people who created the ARM cpu. (ARM == Acorn Risc Machine). And yes they used BBC Micros to help develop the first ARM CPU and the software for it.
@@andrewgrant788 yes, I’m aware of of the origins of ARM - used one of the first Archimedes, followed the dev very closely- but it’s the conflicting stories of how important using the Tube with second processors to inform the team realising how to focus on Arm being the right path that I am hoping someone can clarify. As I say, I recall an article/interview were someone said how important it was, but I also watched a documentary about the dev of Arm that suggested they had already decided RISC was the way to go before they got the other processors to experiment with and test using the Tube.
Lol, that fired up some neurons that were dormant for a while. The only thing i can remember from my assembly days is 1 accumulator, 2 registors? My year end project in university was an ATM machine... i think i still have a printout of that project... got to look for it.
Great documentary video. 2 comments. I’ve been a DJ for 36 years. Disco peaked in 1978, not 1975 as you mentioned as a notable moment of 1975. I thought the interview of the inventor of the 5602 could have been longer. It feels like it was cut short but that’s just my opinion. Overall A+ doc-video.
One little but not insignificant mistake in the video is that the WDC 65C816 would be from 1985, but it actually is from 1983 - which means a backwards compatible 16 bit successor to the C64 could have been done early enough to become the ancestor of our current PCs instead of the IBM PC, with Commodore still ruling the world now.
@@NuntiusLegis The design of the WDC65C816 started in 1983, Atari and Apple received prototype samples in the second half of 1984 and the official release of the CPU was early 1985.
@@EddieSheffieldI read elsewhere here that the full interview didn't exactly... go so well. Bill went on a lot of tangents, apparently, as an example.
Fun fact, Wolfenstine for the SNES had tech in it that made it into Doom. Carmack Used it to prototype stuff like BSP-trees (which is what made Doom at the time sooo fast)
Your explanations of these chips are so inspiring. I started with a Vic-20, but only as a user with some stabs at assembly programming. And to see how such humble chips like the 6502 were the basis for so much is amazing. I personally could never be able to mentally handle all the logic you know about these chips. But I enjoy seeing how it all comes together. Seeing how things work a the chip level is astounding to me. Then knowing the current chips are 'simply' expanding on these humble beginnings.
I am so happy I made the time to sit down and watch this video. There was a ton of great tech heavy information here that was extremely easy to understand, and I feel like such a casual "retro" fan having a murky understanding of the fundamental functionalities of these computers.
Still used on TI-84 graphing calculators which are still in current production. Actually, was used for most TI graphing calculators before the TI92 (and 89 series) moved to the 68K.
@@ZenithMusicNet The MSX was sold worldwide (but not in England or the United States). That's why many people today don't know it. They don't know the rest of the world.
For your demonstration of 8 bit multiplication, I personally prefer to show it in terms of long multiplication, just with a different base unit, but your explanation digs a little deeper into how the logic actually works, which I really like as well. I've been playing the video game Turing Complete (on steam) and I liked taking each of the 8 different bit shift components and adding them.
There is no way a 7 MHz 68000 was only 3 times faster than a 1 MHz 6502, in practical use it was more like 20 times faster. Although it required more cycles per operation, the operations were vastly superior in flexibility and capability, so most tasks could be performed with a third the operations, at 4 times the width. The 68000 also had divide and multiply of 32 bit numbers, where the 6502 only had addition and subtraction of 8 bit numbers. Needless top say the 68000 was hundreds of times faster in those tasks. I programmed Assembly for both, and although the 6502 is very nimble for the time size and price, there is no competition between it and a 68000 in how powerful they really are. OK you actually got to that at 14:47, and yes you had to make your own binary math for multiplication and division and numbers above 255. But even if you have the math, you sometimes use logic operations instead for fixed numbers. At least that's how I did it.
As someone who's been a game programmer for 30 years.. I miss this era.. Coding was so much more "fun" back then.. Modern day developers struggle to understand the basics of optimization... I look at people complaining about how impossible it is to run their game on the Switch (for example) and shake my head.
I agree, but that doesn’t mean it can’t still be fun. I personally don’t do a lot of actual optimization, I mostly try to make code that doesn’t run poorly and it’s pretty fun. This makes me want to do a homebrew project for an older cpu at some point through.
@@colleagueriley860: I work with 100s of people now. And most of them think optimization is a thing you get from adding more power to the machine. You hear them bitch and moan about power of hardware or engines in interviews with press.. it's depressing... And that's before we get to the simple fact the "higher ups" dont care and dont want you spending time/money on making it go fast.. just ship it out...
well, optimization back then was absolutely neccessary, because of the slow cpus, small mem etc. so there was no other way. you HAD to do it and that's why a lot of ppl became very good at it, it wasn't because it's "fun".
You're right. Rubik's cube was more a 1978/1979 big thing. I remember when I was in 8th class in middle school when I was summond to the school director, wondering what "crime" I had committed, to just having the director asking me to solve his cube.
@@galier2 I remember 1980 because it was my first year at highschool and couldn't wait to get my hands on one. There may be a Hungarian or two here in their late 50s who got one before the rest of us.
@@andrewdunbar828same here - 1st year at the comprehensive school and …lets just say, someone was selling legit cubes cheap somehow. I was too naive to realise why they were cheap at the time and only twigged months later, that not everyone pays for things when they walk out of a store!
I joined a gaming company when the special graphics chip for the SNES came out and to make use of it was my first task. It's basically an 8 bit DSP with very limited capabilities, but it was reasonably good (at least compared to the 65xx) at spitting out pixels, line by line, so it could do some basic 3D rendering, a bit like the blitter on the Amiga. I recall it having instructions that are 1, 2 or 3 instructions long with delayed branching, which made some interesting optimisation tricks possible. Since it's a 8 bit processor, it would read instructions one byte per cycle, so if a branch wasn't followed by nop's, it could branch in the middle of reading long instructions, and that could be used to modify the instruction on the fly depending on the branch outcome. When I got this chip to play with I had very limited documentation and tools, and the first thing I had to do was to write my own assembler for it. My work resulted in the game "Winter Gold".
I saved from e-waste a ceramic MOS 6502 with a date code of 3675 which makes it one of the oldest survivors. There are very few older documented date codes, at least that I have found in my online research. It is a prized part of my vintage rescue collection.
Errata: @16:30 - The Atari 2600 only has 12 bits of Address (as shown in your diagram) and can therefore only address 4 KB of memory space. The asteroids cartridge shown is an 8 KB cartridge, but it used a very crude banking scheme.
The 6507 CPU in the Atari 2600 can address 8KiB of memory space, 4KiB is devoted to the cartridge ROM and 4KiB is devoted to RAM, I/O, timer video and audio registers inside the RIOT and TIA chips.
I came to say something similar. BrainSlugs83 is close, but NerdlyPleasures has it right. I'm currently working on my first Atari 2600 game. The 6507 also still has internal vectors for interrupts that have been moved to fit into the modified address space, even though it no longer has the external pins that trigger those interrupts.
One thing that occurs to me while I'm watching these videos, is how ubiquitous computers are in our lives. I'm a truck driver, but I still rely massively on computers to do my job... I have a cell phone, I have a SAT nav, I have a tablet, the timing, horsepower, exhaust, and RPMs of my trucks engine are monitored by computer, my dash cam is run on a computer... And, I have a Steam Deck in the bench of the truck... All of these devices Trace lineage back to the fascinating technology that my mans here is explaining, and it's all so fascinating. And frankly, a little incredible.
I don’t think it’s reasonable to say that the 6502 powered a generation. At least in the U.K. the ZX80/81 and the hugely popular Spectrum were Z80. My first micro, a TRS-80, as also Z80. The Z80 ran CP/M and was also a far easier chip to program.
@@Okurka. Poor? Not at all. It was used in many business machines. It was the basis of the MSX computer standard. It was also widely used in games consoles. It had a rich instruction set, including 16-bit instructions.
But the BBC Micro had a 6502 and that fact eventually led to the creation of ARM in the further history of the Acorn company, so I think the impact of the 6502 even in the UK is probably higher at least from today's standpoint. By "powering a generation" in the 80 of course you are correct, that probably goes to the Z80 for Sinclair's computers being cheap.
@@mudi2000a The BBC Micro was expensive. That’s why the cheaper ZX Spectrum was so popular. I’m not saying that the 6502 wasn’t hugely popular, just that the Z80 was too, and it was also very successful so that the suggestion that the 6502 powered a generation of computers was wrong. It was a mixed market, and remained like that for a long time.
@@jbponzi1 Yes, it's sad for me, too, especially as I am an old and sentimental guy. But, on the bright side, the Z180 still lives; I wrote a lot of code for the Z180. And also the eZ80 - already over 20 years since introduced; seems like yesterday. Life is so short and CPUs come and go! 🙂
Back in the early 80's I wrote many thousands of lines of 6502 assembly. Aside from being fun, it taught me a lot about programming, conserving memory, writing efficient code, coding with minimal instructions, and a lot besides. Oh and of course if you are coding exclusively in assembly and burning your code into Eprom, you certainly learn about writing error-free code
65c816 cores were also used in Micronas cdc16fxx series chips used in cars - interestingly, a large part of WV Passats and Audis from 2000 used them in dashboards and other brands in fuse boxes ;) I used these chips to reverse engineer the Audi dashboard, and I still write Atari XL/XE demos ;)
My first experience with the 6502 was the Acorn Atom computer from 1979 (I kept working with the Acorn line of computers until the RISC-PC). As the Acorn Atom has a BASIC that had a build-in compiler (you could use mnemonics like STA, JMP, ROR and so on, and even labels to jump to), I learned coding in machine code for the 6502 pretty fast. It helped, of course, that I already did machine code on my first single board computer - The Cosmac Super Elf, with an RCA 1802 processor and 265 bytes of RAM. I restored that Super Elf into working condition by the way, and am now busy doin the same for the Acorn Atom. There are still some hurdles to go, but that's the fun of it... Anyway - I really liked that 6502 processor. It was easy to write programs for (in machine code of course), and I wrote several toolkits (burned in EPROM) for the BBC-B computer. Great times...
@@ArneChristianRosenfeldt I am not sure what you mean. If you use the in-BASIC compiler, you could specify where your compiled code runs. By default it's above zero page of course. In the BBC-B computer I often used a sideways RAM board to store my compiled code, and run it like a sideways ROM. I must admit I am VERY rusty about that, because it's more than 40 years ago I did all that stuff. I hope by rebuilding the Acorn Atom into working order I get a platform to restore that rusty knowledge once again. 🙂
@@jclosed2516 Yesterday I wanted to try to JMP() , but ultimately could not use this instruction. But again it was right before my eyes: Code does not come from the zero page. Jumps are either relative or 16 bit absolute even in the indirect case or with JSR or RTS or RTI. I also found out that in the Commodore C16 the zero page is free for me to use here D0--E8 ( sic in the book I have, not --E7). Not much. I also have to PHA;TXA;PHA
My first computer was a clone of the ZX Spectrum (TK-95), so I always really liked the Z-80, but this video got me really interested in the 6502 and really gave me a lot of information about this CPU that I didn't know.
Great job on the video! I learned about the 65C816 from SNES specification and development documents many years ago then eventually learned about the original NES and C64. Your explanations about pins and such is a great for further learning how the chips were related and even cross compatible.
Thanks for the walk down memory lane! It was fun to hear about all the 6502 variants, and all the machines that used them. I bought an Apple //e in 1983 and spent several happy years programming it in assembly until moving to the MS-DOS world and C a few years later. I had no idea its 65C02 supported new instructions. I'm sure I could have taken advantage of that at the time, as speed and code size were always at a premium. Now I'm thinking I should build a 14MHz W65C02 machine and port some of that old code to run on it. Young me would have thought he died and went to heaven with that clock rate!
I really appreciate your videos. There is something about them that has awakened my passion for mid-to-late 90s, early 2000s technology. I got into computer repair in 1996 and have been in love with them since. Just started teaching myself assembly and C and wow, much different than hardware :) Btw, I wish I had a picture of my grandfather. You look exactly like he did at our age lol.
Thank you for this video! I have bought an Acorn Atom in 1981, expanded it at first with piggyback 2114's, but at the end I had rebuilt it completely in a 19 inch rack running on 2MHz. I also had extensions like a 80 column video card, an eprom programmer, multiple floppy disk drives, a ram disk, a telex signal card and even a IEEE 488 controller card.
My first computer was a KIM-1. Took a class at the Byte Shop of Palo Alto on the Z80, but never resorted to assembler/machine code on that family. Just felt bad they had more registers till I realized the 6502 had a whole PAGE of 'em... Had as OSI Superboard, eventually Apple II. My favorite processor, so clear and straightforward. 6502 FOREVAH! My KIM-1 eventually ended up in a briefcase with a power supply and a a bunch of protoboard as I tried (with no training) to build a digital synth during the day, then play amazing fusion jazz at night. FUN times!
Brilliant vid. I've picked up bits and pieces from these vids before, but it was great to have it all in one. Hopefully we'll see some similar vids on other chips in the future.
10:15 and 1.79 MHz is half the frequency of the 3.58 MHz color subcarrier used for NTSC video, which is why you see those particular numbers show up so frequently in clock speed. Some SNES games used the SA1 chip, which contains a second 65816 running at a blazing fast 10.74 MHz.
Coming from the Tandy world, I was always a Z80 person; I wish now that I had diversified into the 6502 as well. Back then it was almost a Hatfield vs McCoy's. Thank you for presenting more into my little world than I expected!
I'm not surprised that he said that. ARM was developed by people who were very much connected with the 6502 world and had been involved with the various 8-bit Acorn machines. In fact the ARM 1 was developed using BBC Micros and their Tube adapters. So you could say that the 6502's influence still impacts the computing world to this day, even if you ignore the retro world!
imagine making your own pipboy from fallout using a 6502 it has all the specs for making one of those
Yes, and I‘ve noticed that at least one of the dev teams I’ve got to know for medical grade hardware use a 6502 core instead of more modern ones in FPGA as they‘re basically feeding it the same software they did decades ago.
Can make validation simpler.
The 6502 is sure historically important and influential, but you can't see anything of it in the ARM design really.
(Yes, I have studied both instruction sets, along with many others.)
Dang, that interview was strange. The video feels like it is missing the part of why the chip was designed in the first place and why it had such wide proliferation.
My grandfather was the CTO of Zilog and raised me. He helped develop the Z80.
I am very proud of him. He made life unrealistic for me.. he raised me and spent all his free time with me building PCs and listening to whatever music I found. I’ve never found anyone who holds a candle to him. Not having him around has been tough.
Knowing his work won’t ever die in some capacity makes me happy. Not this IC.. just in general with his work. He also helped GE in the 60s build their first main frame. Dude was the most modest man you’d ever meet.
Edit: I’m not going to put my families personal info out there, had no idea this comment would get any traction. Believe me if you want or don’t.
The Z80 is very cool and fun to program. Thanks for sharing! And big props to your grandfather and Zilog.
Hell yeah, would that make you the first Game Boy?
The TRS-80 Model I was my first home computer, and the first platform I learned assembly language on. The Z-80 will always have a special place of honor in my vintage-computing collection. 🙂
The Z80 was in everything, there was nothing like it. Still not fully understood today completely. Even the military used it everywhere.
interesting my grandpa on my dad's side couldn't even find the on button on a pc and refused to use them .my grandpa on my moms side knew a lot about analogue electronics like crt tv's and vcr's but when it came to computers i didn't have one until 2008 and kept getting malware .
At 4:38 that's me closing the VCR lid in the video "VHS VCRs Revisited" ... Big fan and honored to be mentioned/shown in an 8-bit-guy video! (First programming experience was BASIC on the Apple //e on a 6502 ...)
Note: As I recall, Apple //e used a 65c02, a CMOS version of the processor.
I like how the thumbnail is him offering the cpu like it's a quest item
"6502's, cases, RAM, you want it? It's yours my friend, as long as you have enough money. Sorry fuzzy7644, can't give tickets to the retro PC store. Come back when you're a little, MMMMMMMM richer." -What you probably thought he would've said
@@arbrilliant191 This is disgustingly accurate to what was on my mind, thank you.
A few thoughts...
1. I feel this only scratches the surface. I get that this must be a monster project, but I'd watch a 2 hour version of this going into endless detail. Or a 4 part series even. The 6502 strikes me as one of the most important CPUs in history, it deserves ALL the effort and detail.
2. The segment with Bill Mensch was very short, perhaps there wasn't much to add, or he couldn't share anything new, but I wish there had been more.
3. The Z80 needs a similar video, as does the 6800/68000, as does the 8086
3:40 Nice CGI
You got some solid points, but The 8-Bit Guy already made several videos of consoles and computers that were powered by the 6502 and he did go into fair details of the chip scattered across them. I hope David covers the Z80. I own a 20mhz variant and I would love to learn more about it.
6809E here lol
@@elijahvincent985My z180 powered SC126 also would say "yes" to a video about the z80 and its descendants.
@@Mmacrossfirekenai And the 6809 family deserves a lot of love, too. OS-9 for the win!
Thanks so much for the informative videos. Your channel inspired me to transition into IT. Two years later, I’m working full time in the industry and couldn’t be happier! Your channel is the best.
6502 is a canonical piece of computer hardware, similar to classic game consoles. They will keep getting made, emulated and implemented in FPGA forever.
the 8-bit guys is correct about how cpu speed does not compare well against different cpu's because amd's cpu's where slower then the pentium 4's but could seriously match or even outperform faster pentium4's so the 6502 is not the only cpu that proves this issue it works even with modern computers cpu's
I wouldn't be so sure, machines are getting ridiculously complex, and I doubt you're going to be implementing RDNA GPUs with unified memory in an FPGA anytime soon. Let alone accurately. 😅
@@SaraMorgan-ym6ueRight, however on "modern" hardware (so even 20 years ago) the clockspeed was just a tiny bit of the puzzle. The size and usage of the cache was in most cases much more important. CPUs have reached a complexity level that makes it almost impossible to predict the performance ahead of time. They have dynamic pipeline optimisation stuff going on that's just wild. In the good old day you just pick up an instruction set sheet, look up the clock cycles of each instruction, add them up and you get a fairly good estimate how long your code would take to run. This isn't the case anymore.
@@SK83RJOSH I read your comment but Im not sure what it has to do with the 6502?
@@Bunny99s well give it credit for surviving this long since no other cpu's have lasted as long as the 65c02 and 65c816 they are in it for the long haul plain and simple (for being manufactured this long) I mean lasting this long for manufacturing life all other cpu's are discontinued
My first computer was in college in 1978: the MEK6800D2 development kit. I programmed a prototype for an automotive MPG display and a Times Square marquee, each in 512 bytes. Programming was all in assembly and hexadecimal. I would measure the frequency of the odometer and fuel sensor, divide distance pulses per second by fuel pulses per second and converting that to MPG units all using bit shifting. Storage was Kansas City standard audio tapes. My productivity improved immensely when I learned how to use the 6800 assembler on the university mainframe. I was unaware of the history of the 6502 and its connection to the 6800 at the time. It is great to hear from the people like Bill Mensch, who were hands on with some these things that are now household words. Thousands of pacemaker users are happy to know they are using some of the most tried and true processors there is.
Wow! Fascinating!
My Brother owns an Original ATARI 2600, and what is left of the ACER 486 S/X PC we had for Years back in the 1980's and some of the Early 90's.
I own The NES - and Many NES games! one for sure is quite rare and I can not even find any info on the game. I think it was a prototype/Unmissed game!
I did my capstone project for my BSEE in 1988 writing assembly code for the 6502 on an Apple II+. Really fond memories!
I first programmed the 6502 on a synthesizer in 1979. A small music company in Oklahoma City called PAiA Electronics (still in business!) used a 6502 for computer control of a modular synthesizer in 1970's. John Simonton (en.wikipedia.org/wiki/John_Simonton ) was the designer. A fun fact is Larry Fast of Synergy used the Apple II to play some of Simonton's software for the PAiA platform. Fast played with Peter Gabriel and relied on Simonton for tour support on the first tour through Texas in 1977.
So many electronic music performers are very familiar with the Rockwell 6502.
@davidryle Do you remember a guy named Hal Chamberlin? He wrote a killer book called Musical Applications of Microprocessors. It was my “Bible” in 1980.
@MultiPetercool I still have my copy of that book. An amazing resource the time, and still full of good info.
@@GodmanchesterGoblin I met Synth Legend Roger Powell about that time. He was working with an Apple 2 with a Mountain Hardware 16 channel A/D-D/A converter board in it. He used to to handle control voltages and triggers for his Moog and Serge modules. Cool stuff! I think he used it on his second solo album Air Pocket.
That is super cool! I've always wanted to get one of PAiA's theremin kits!
Larry Fast rocked.
David please post that whole interview/chat. Please
Yes, please. It felt very edited down! I'd love to see more.
+1
Agreed.
Yes, PLEASE!
Seconded
Fascinating David I learnt a lot and I had no idea that the 6502 was still being made and it makes me happy that it is. Back in the day I had a Commodore 64 and Commodore Plus 4 at home and used BBC Computers at school, all 6502 based and used to write programs in basic for them all so I understand about memory access and some of the CPU instructions. But what I never really understood is how it all fitted together. You turned on and there was a BASIC prompt and a flashing cursor but it all seemed a bit magic as to how that all happened. I never knew about how memory was addressed and stuff like chip select lines. Between yours and Adrian's channel I have learnt so much over the last few years and now understand much better how the computers of my youth actually worked.
Really enjoyed the video, thanks!
Ben Eater has a very good video on the 6502, explaining all the decoding logic and stuff.
Thank you! This was one of the most interesting 8-Bit Guy videos I watched in a long time. The only thing is that it kind of left me longing for more. Also, I expected a bit more of the Bill Mensch interview. I appreciate the effort that goes into making a video like this, but if it was up to me it could be at least twice as long without being boring. Would love to see more of this!
Yes David you don't need to apologize at the end, it wasn't too long and many of us would be happy too watch a "Part Two" video if you make one.
Great video!
Fun fact: The 6502 is essentially a simplified Motorola 6800 implemented using NMOS with a different pinout and a slightly modified instruction set.
That's because most of its engineers came from the Motorola 6800 team!
Those simplifications were super important though: 6800 was eye-meltingly expensive at the time and the 6502 was cheap enough to practically sell by the barrel!
@@talideoni think it was $25 vs $125. I have the feeling that production cost was not the only reason why the motorola was so expensive. Misunderstanding the market costed motorola to lose the industry
Fun fact:(s) Similar to the 6800, the Data General Nova (1969) had two accumulators and two index registers and a zero page! But historians and Chuck Peddle himself (who worked on both the 6800 and 6502) said the 6800 was inspired by the PDP-11 although from a modern perspective the PDP-11 is a lot closer to the 68000. However when you are coding the Motorola processors (6800, 6809 or 68000) vs the MOS Technology 6502 it is apparent that the C style structs are prefered by the addressing mode design and that does have PDP-11 roots. On the 6502 you end up wanting structure of arrays.
Yep. In fact Motorola sued them because of the extreme similarities.
Yeah, as an electrical engineering student, I cut my teeth on assembly on the 6800. I guess the college must have gotten a deal going with Motorola so they could “hook” us early. We soon moved to the TI TMS9900. And holy cow that thing was 16 bit! Wow. …lol
You got tantalising close to giving away a key reason why the 6502 was so popular, and that's that it leaves the bus available for video half the time in a predictable way. By luck (or judgement?) one clock cycle equalled one character time on the CRT, so one half of the cycle the CPU guzzled data, and the other half was left to video. Essentially it gave way to an early form of DMA.
This wouldn't work on a Z80 as it doesn't have a predictable bus, so instead you have to implement other segregated ways for video.
6502 designs were often UMA like current consoles.
I wish someone could create a 6502 religion...
Of course it kinda already exists, but someone needs to make it official.
The Gameboy ran a Z80 variant so I wonder why Nintendo made the switch there.
@@Alfred-Neuman Bender did! (Bender was powered by the 6502): Bender founded Robotology
It was just as easy to implement memory mapped graphics in Z80 machines. All clones of Sinclair Spectrum implemented it with a handful of ttl logic chips. Overall the complexity and performance were very similar.
In 1979 a I bought 6502 based kit computer called the Tangerine Microtan 65 . Every single component had to be soldered by hand onto the circuit-board etc.. I had never soldered before and was learning as I went. It worked the first time I switched it on; to my utter astonishment.
Fondly remembered.
@The8BitGuy You do an excellent job on these computer tech explainers. You provide the best quality videos of this type on RUclips. Great job buddy!
1:22 I wouldn't be surprised if there's cars running around with 6502s under their dashboard. My 85 F150 is one of the last vehicles Ford ever sold with a carb, but it had an Intel 8061 under the dash in the primitive EEC-IV system. It would be shocking if there ISNT a ton of cars rolling around with 6502s running their primitive EFI setups/last gen carb setups.
I heard rumors that some Chryslers had RCA 1802s under the hood for fuel management.
Ditto, I drive an '84 F150, also with the same 8061 EEC-IV feedback carb (300 I6). They only offered feedback carbs for 3 years (300 and 351 84-86, 302 84-85). A bit of a weird technology, but for an eccentric like me, it fits!
You wouldn't be wrong... GM actually commissioned a 6800 variant for their earliest engine control modules in the early '80s! (the 6502 began as a low-cost version of the 6800 after all)
6502 were used for ICBMs (Rocketdyne) as well as the spaceshuttles.
There are a lot of cars using 8051-compatible processors for various tasks. In recent times, most of those have moved to ARM, but, possibly some simple tasks like seat adjustment might be done with a specialized motor controller running 8051 code.
10:50 The TurboGrafx/PC Engine's HuC6280 (A variant of the 65SC02) runs at 7.16 MHz & in 1987 may have been one of the fastest 6502 variants. You mentioned the SNES and Apple IIgs use the 65c816 a "16 bit" variation which may make them a bad benchmark. Where the TG16/PCE CPU is closer to the 8 bit 6502.
11:35 Although the TG16/PCE had a few RAM expansion HuCards used by the CD-ROM attachment, there were no cartridge based enhancement chips ( 11:11 ) such as those used in SNES games. The TG16/PCE 65SC02 does make use of a video controller (HuC6270) and a color encoder (HuC6260) which certainly do some heavy lifting, but I still think the TG16/PCE would be a very good showcase of what the 6502 can do.
Oh, and there is a mouse available if someone decided to port some games over.
This statement is a bit deceptive. The HuC6280 had an instruction set extension for copying memory to / from its video controller, which was far faster than doing it in software with the regular LDA/STA instructions.
@@glenndoiron9317 Deceptive is a bit deceptive. I’m just being concise in the internet. So sue me.
@@grafxgearexpect a summons in the mail.
That video co-processor did a lot of the heavy lifting on the PC Engine, much like most consoles of the time.
The same even went for the C64. Things such as scrolling the screen, throwing sprites around the screen and collision detection, all of which could be very CPU intensive, were handled by these chips allowing the CPU to handle other things. That's why if you look at C64 versions of games vs. a Spectrum or an Amstrad version, no matter the quality of the gameplay itself(I can tell you from experience some Spectrum versions play better than their C64 versions), the C64 would typically have a smoother framerate and better scrolling. The only games I can think of where that isn't true is games that use 3D rendering.
I don't think the OP was deceptive, but I always considered the TG16's name to be deceptive. :)
I got my game dev career start on the Apple II, back in 1983 and coded the 6502 on the Apple, Vic-20, C64, Atari 8's, etc. until moving to 68000 on the ST, Amiga and Sega Genesis. 40 years later and I kind of miss the 6502 and regardless of the struggle and limitations, there was something really "fun" about getting things working back then vs. now.
You coded the C64's 6502?
@@Okurka.oh here we go... Well, I've coded several machines over my 40 year career. I guess I could get into the minutia of every specific variant, but I was being general. But I guess I should be more pedantic, considering the video and likely audience. 🙄
I *love* these breakdowns of CPUs and other chips, genuinely. Thank you for making my day! (I had a crappy week and this actually made up for it!)
I love the way you do nitty technical deep dives. Your ability to take incredibly complicated subjects and explain them in such a way that they are understandable withought omitting important information is truly awe inspiring.
David, could you please upload the entire unedited Bill Mensch interview? I would love to hear more of his thoughts. Such a brilliant man. Thanks!
Writing a 6502 emulator is practically a rite of passage.
I have not done this yet. What have I not passed?
I skipped that and wrote an 8086 emulator for MIPS and ARM 25 years ago so you could run PC programs on those processors for Windows CE. Found an undocumented feature of the 8086 on the way that MS-DOS was dependent on too.
@@nicholaslarson3778 Once you finish the emulator a cutscene plays while you're sleeping.
I skipped both and touched a boob
Writing the correct Overflow handling in a 6502 emulator is a rite of passage.
David, this video was AWESOME I learnd soo much, as always thanks a ton 🙏 from the guns to the keys and cpus the sheer qualify of you videos is admirable! Thanks!
I have been a subscriber for a couple of years and come here every so often when I'm on RUclips . I have many other channels that I watch for different topics but I always enjoy spending time here as in this video. I remember all the hardware that used the 6502 back then so it's like a trip down memory lane. You put your videos together in such a way that makes them very enjoyable to watch and this one is no exception. Thanks a lot for this and please keep giving us the great content that you have been providing.
so sometime in your life you have interacted with a 6502 processor you know🤣🤣
Another awesome video! Thank you!
I'd like to hear more of the interview with Bill Mensch and what he had to say about the 6502 architecture and the direction it might have gone.
1975 vintage here too! Happy early birthday, young man.
Wait. The damn furby is running on a 6502?! That's the kind of random factoid that I live for.
The 6502 and Z80 are immortal in a way. These ancient chips from the 70s keep finding uses. You've listed many of the uses of the 6502. The Z80 also has some fun ones. Maybe you all remember those old MP3 players, the ones that were basically USB sticks with a battery and headphone jack. Those things use a Z80. It's fast enough to decode an MP3 as long as it doesn't need to do anything else.
Pretty sure it used a DSP or something to decode MP3.
no 8 bit of the era can decode mp3. do not tell us such kind of bullshits
There might be a Z80 core in some mp3 players for controlling them, but no way that thing does the mp3 decoding. That is done in a hardware decoder on the same chip.
@@mrnmrn1As i said. no 8 bit processor has enough power to sw-decode a mp3 stream.
@@gasparinizuzzurro6306 And that's exactly what I said, too...
This was a great video, thanks for going the extra mile and even interview one of its creators.
I'm glad we are all on this journey together
8bit guy has taken us a long way from, vintage PCs to now ancient classic computer architecture
In a way, I feel like we finally arrived.
If this was the intro to vintage computing I had at the start of your channel, I can say I would be able to follow you. But, here we are!
I feel like we are getting closer, if not already, at the heart of vintage computing.
I think, the future is bright. People can now take the first steps in the right direction, of where we need to go if we are to grasp the future of computing while still connecting to the real legacy of computer technology and architecture!!!!
1977 - the year I moved to Colombia. 1980 - the year I wrote my first computer program. It was written in 6502 assembly on an Apple II. Great memories!
This was informative and approachable. Nice work, 8BG!
Do you have any plans to release the full interview with Bill Mensch? That segment was shorter than I was hoping it would be.
The interview with Bill didn't go so well. He goes off on a lot of tangents, and there wasn't a lot of great material to share.
@@The8BitGuy Appreciate the honest reply. It's a shame the video didn't go well. Best of luck to Bill anyway. He's a piece of history.
FR
I, too, truly appreciate the honesty. Mensch is obviously an architecture legend, but that doesn’t automatically translate into eloquence.
Tangents are fun too, sometimes!
Love your videos!!
This is a great video
I never realized that modern technology still uses the 6502
It really powers the whole world!
And all because some guys wanted to make a copy of the Motorola 6800
Well I'm a big 6502 fan but the world is probably powered by ARM. Which is not bad because at the end of the day ARM was designed by Acorn which previously used a 6502 but then needed something better which still doesn't use much power, and the result was ARM. There is an excellent video by the channel www.youtube.com/@LowSpecGamer about the history of ARM; he also has very good videos on the history of the 6502 and the Z80.
They didn't want to make a copy, they wanted to make a much cheaper CPU! :)
Wow, I knew the NES used a 6502 but I didn't know about the furby and Tama gotcha, that chip powered my '90s childhood. Perhaps the 6502 was also in the "digivice" that had a pedometer and you walked or shook the thing to get your Digimon to walk and play the game.
A lot of toys use it. For example Tamagotchis ruclips.net/video/c4PkcZScBV8/видео.html
Variants of the 6502 are still in use, one of the most versatile CPUs ever designed.
Fun fact, the code for the Furby is available and is fun to look through if you know 6502 :)
Back in 1988 on the first year of my computer science degree course we had to program 6502 processors using assembly to create a traffic light system.
I had a 8080 at home, self build, with 2MHz, at school we have a CBM3032 with a 6502, 1MHz.
I created two almost identical programs in assembler, calculating the 92 (or were it 72?) ways to put 8 draughts on a checker-board. One on my 8080, the other on the 6502.
And the calculation-time for this was the same on both machines. 4 Seconds for all solutions, a Pascal-Program on the CBM needed 30 minutes.
The 6502 was almost a RISC-processor, all commands were done much faster.
Please consider the Motorola 68k chip next. It was in the Mac and the Neo Geo and several arcade games
And a little known thing called the Amiga.
But it's not 8 bit!
And the TI-89/92 series of graphing calculator, which are still in production in some form to this day.
@@-taz- There was the 68008 variant with an 8 bit data bus, but otherwise fully compatible.
And Atari ST
I grew up with the 6502. Nostalgia makes me love it. Its abysmal registers brings me back to reality.
I'd like to see a deeper dive. It was a transformative chip.
I grew up with many of the systems it was in. C64, NES, Apple II, etc. It was awesome.
First time I got to use a PC in real life was a OSI C1p and 41P at 12 years old in the late seventies. I still remember the first words OSI 6502 BASIC VERSION 3.2 (c) MS etc.
The MS part doomed me for life lol
@@axle.student Was probably pretty similar to the commodore basic too, as both were made by MS.
@@maxxdahl6062 Similar. The Commodore also had a more advanced color display as well as the advanced sound mapping, but even the base code had subtle differences. Apple II, Amstrad etc had subtle differences. There is a page somewhere in wiki that has a list of the different 6502 BASIC variants.
From the OSI BASIC start I can code in any form OF BASIC ever created. I Often use FreeBASIC beside C and Python.
Once you realise all the zero page instructions are there to provide loads of “registers” that can be used in lots of flexible ways, the 6502 is actually replete with registers ;)
The only CPU I learned to program in Assembly language!!!!!! I’ve got a soft spot in my heart for the 6502, and I’m amazed it is still alive and kicking after all this time!
and the 8_bit guy was born that's why 1975 was a great year indeed
Me too, the 6502 was the only CPU I did Assembly on
Hello David.
Since I am so broke, I can't really afford Patreon, so this is the only place I can say it, but I wanted to express my deep thankfulness for you, LGR, and Techmoan changing my life. I have really wanted to meet you all, and get to know you a little better, and maybe send you something from the goodness of my heart, just to say thanks for getting me to where I am now. I am a sixteen-year-old nerd who from you three got really interested in old tech and began collecting old software, computers, typewriters, hi-fi equipment, etc. At this time I am now thinking about making my first video, but I haven't thought about it. I wanted to come on here and tell you guys I love all of you. I love watching your videos and learning new things. Always learning new things and weirding out teachers at school because they dont expect someone like me talking about these. I am talking about things they grew up with (lol)! And it makes me happy that I am now so interested in something I have spent literal years to dive into. Sounds weird to say but I enjoy growing up with you and the other two's videos. It makes me proud. Anyway, that's all I have right now.
Thank you! ❤📼💾
I appreciate these guys’ work too. But if you want real, eternal fulfilment I highly recommend you turn to the Lord Jesus. Read one of the Gospels today, perhaps. God bless. 🙏
1:58 Thank you for including the OSI 4P. Warms my heart to see that :)
7:40 Thank your for the excellent illustrations so far. I am hoping to replicate some of this with the Atmel mega328 some time soon. The bus CS I recognized instantly, but I was curious about how they were getting the 16-bit pointers and integers.
25:16 thanks for the presentation. That was awesome :)
Great Video!
Last year I actually did a project where I took a W65C02S, and connected it to a microcontroller, and monitored its inputs and outputs.
I used the microcontroller's memory space as the addressable memory that the 65C02S used, and included peek and poke commands to monitor the memory addresses, as well as a clock step functionality and breakpoint functionality.
I even wrote a disassembler for 6502 Assembly to go along with it.
This was a fun project that i taught myself 6502 assembly with. It was a very cool experience.
I'm now thinking of designing my own side project device around operating on the W65C02S.
This is the best kind of 8-Bit Guy videos. Thanks!
I remember reading a book about the history of MOS, in one chapter which oddly you can hardly find any reference to online, it talks about the toxic chemicals used in the creation of the masks, the fumes from which would turn paper yellow. In 1974 the MOS plant in Audubon was listed as a hazardous waste site due to a leak of trichloroethylene (TCE) which caused local groundwater contamination. It stood out to me as I knew people who worked in the area who ended up with a profound life changing chemical sensitivity. While the 6502 may have powered a generation, it's creation may also have damaged one. I just think it's important that those things are not forgotten too.
Wikipedia does mention the leak, and the factory becoming an EPA Superfund site just as Commodore was in the process of taking over. Sounds like MOS hadn't informed the new owners and left them holding the bag. It doesn't say what happened after that; if I were in Jack Tramiel's shoes, I'd have sued the former owners into the middle ages!
That's life in the 70's! I grew up where a chrome plant dumped their plating chemicals into the town aquifer to help boost testicular cancer rates. Probably why the EPA came about in the 70's, too.
Did they ever want to dispose of TCE properly, or just pay the shareholders and then go out of business? I don’t understand why everyone wants humans in a manufacturing or chemical plant. I also kinda hate the ISS.
@@coldlyanalytical1351 red LEDs are made of Gallium Arsenide . DVD lasers.
You know it's a good day when the 8-Bit Guy uploads!
I did electrical and electronic engineering at college in the late 80's/ early 90's, and we were taught about computer architecture and assembly language using a 6502 dev board. Really interesting and loved it. Although I cant remember much of the assembly language now.
My first experience with programming 8 bit CPUs was with the Signetics 2650. I built a system based on an Electronics Australia project based on the S100 bus. It had 4k of RAM and was clocked at 2Mhz. The system could be expanded as the S100 was already a standard for expansion cards. I never hear anyone discussing this CPU.
I later switch to the 6502 (6510) when the C64 was released. A much easier instruction set to use.
The PCEngine/TG16 ran a HuC6280 CPU, which is an upgraded CMOS version of the 6502. It ran at 7.18mhz and was released in 1987. I believe that was the fastest clocked 6502 based CPU of the time period you listed in the chart that has the SNES and Apple IIc Plus.
I think that it is weird that the cartridge also ran at that clock. Hudson added an instruction to burst copy data to VRAM ( and RAM). So level loading was fast and you can spend the splash screen on decompression. Or stream a level. Only problem: all operands are immediate. So jump indirect or self modify code.
Did it really run at 7.18 MHz, or did the system insert waitstates? The NES and SNES have a system clock at 21.47{72} MHz (6 times the NTSC color subcarrier frequency), but that doesn't mean that anything else in the system runs that fast. The graphics chips were the fastest at 5.3693{18} MHz. Also, the SEGA Genesis had a system clock of 53.693{18} MHz (15 x color subcarrier), with various dividers for the components.
This is where MIPS benchmarks become more useful...
@@shinyhappyrem8728 Fanboys assured me that it does run at this clock. Remember: that was in 1987. Even the first draft/prototype of any circuit made in a NEC fab would run at that clock.
@@shinyhappyrem8728 You can easily tell the way games look. Most SNES games depend heavily on "eye candy": rotating, scaling, particle effects. But underneath, not much happens, as the CPU is weak. A shmup on the PC Engine easily has twice the bullets and enemies, and all the calculations for the positions are CPU work. Weird enough, even the NES is able to handle more complex patterns of objects than the SNES can. Many games that came out on the NES and SNES, are "emptier" on the SNES. 2 or 3 less enemies, fewer bullits on screen, stuff like that.
@@lovemadeinjapan: I assume it also depends on the programmer(s) - check out the SNES game Rendering Ranger R2.
The 6502 was a legendary CPU is so many ways. In a way I think it's use just about everywhere well into the 80s hurt the 65C816's chances of success since it seemed like there was no urgency in designing a successor IMO. The 65C816 it wasn't available until 1985 and it's chief competitor was the 68000 was introduced 6 years earlier. The 68000 was comparatively much more expensive than a 6502 but once it's price started to decline in the early 80s interest in using it went up dramatically!
If MOS or WDC had started working on what became the 65C816 years earlier maybe it could have scored wins in devices that ended up using the 68000.
Also, I can only imagine what the SNES could have been capable of if it would have come standard with the version of the 65C816 in the SA1 chips that was used in a few dozen games. The SA1 ran at 10.74 MHz which is 3x the 3.58MHz the CPU in the stock SNES ran at.
Still I wonder what would have happened if Commodore had released a successor to the C64 based on the 65c816 in 1985...
CPU's back then: "I was built for an arcarde maschine, but can also run your car, home control system and lunar rovers as well!"
CPU's today: "I am sorry, but you cant swap me from an identical Maschine, since our serial numbers dont match. Please go to an authorized store that forces you to buy an entirely new motherboard assembly thats like twice as expensive than a new computer"
Thank you for the sidebar into multiplication with rotate and/or lookup tables, it was really elegantly explained.
I gained a new appreciation for how the 6502 instruction set allows for a clever programmer (or compiler) to do multiplication in just a handful of clock cycles. I had previously assumed you were just stuck with "a series of additions" (with large-number multiplications just taking an absurd number of cycles) or "paying for die space with extra-fancy chip circuitry", with no middle ground.
Wow. I am impressed to hear about Ben. His videos are great and an excellent supplement to Davids about older computer. Bens videos explain in detail how things in an 8 bit computer work, because he is building on in front of our eyes.
The most important company that used the 6502 was Acorn (BBC Micro) who then went on to create ARM!
Mine still works, the two pages of memory space for mapping external peripherals is brilliant.
Fun fact. The UFOs at Roswell ran on 6502s. Bob Lazar told me
UFO Phil heard it from George noory
Fun research: look up robert lazar's criminal history.
@@Peter_S_ propaganda?
@@Nobbie248 No, court and police record. Public records. He's not a nice person and he's a typical con.
And a 68030 and AOL took down the aliens. Just watch Independence Day
The ARM was envisioned as a 24/32 bit 6502, with a few more registers, so I understand Bill's, no need to reinvent the wheel, pov.
The big idea behind ARM was to exploit the memory bandwidth to the max. When they evaluated all the current 16 bits CPUS (80285, 16032, 68000, 65816) they were appalled by their inefficient use of the memory bandwidth.
As did the 6502, with the rising / falling edge CLK logic almost doubling memory bandwidth (as i believe a video somewhere on RUclips mentions). Acorn's MEMC 1 / 1A may have been more efficient, with a near single clock cycle memory accesses possible, when using the read / store multiple instructions (to push / pull fifteen of the twenty seven, 32 bit registers, as a block, taking 18 clk cycles = 3.33 bytes per 8MHz clock tick, rather than the standard 2 clk ticks per single ARM read / store op = 2 bytes per clock tick, let alone the 4 clk ticks needed on the like of a 68000 to read a 32bit word = one bytes per tick. Memory access was 2 to 3.33 times faster on an Acorn Archimedes, at the same clock speed, than a 68000 based Amiga, ST, Mac, Sun One, ..., the Archimedes were also marginally faster clocked than the PAL variants of the above), making for some quick stack accessing wizardry on a branch or return, along with some weird, 7-bit, reverse page logic, initially eliminating the need for much of the slow blocking logic found in the CISC toys. Though only 22 bits (4 MB) of the address bus went anywhere (without a multi MEMC and motherboard fudge), and the VLSI logic / DRAM couldn't be clocked above 8MHz, for the first six years of ARM chippery. Till the ARM 610 appeared, you couldn't source a true 32bit ARM SoC.
@@galier2I remember reading somewhere around the time ARM 1 came out, that they tested the other CPUs using the TUBE on the BBC micro - because that allowed them to separate specific aspects of performance which exposed the issues. I held on to that idea that the BBC micro has that important role in the dev of ARM which is now in billions of devices - but I heard something more recently that cast doubt on that. I’d love to have confirmation either way :)
@@sputukgmail The BBC micro was developed by Acorn the same people who created the ARM cpu. (ARM == Acorn Risc Machine). And yes they used BBC Micros to help develop the first ARM CPU and the software for it.
@@andrewgrant788 yes, I’m aware of of the origins of ARM - used one of the first Archimedes, followed the dev very closely- but it’s the conflicting stories of how important using the Tube with second processors to inform the team realising how to focus on Arm being the right path that I am hoping someone can clarify. As I say, I recall an article/interview were someone said how important it was, but I also watched a documentary about the dev of Arm that suggested they had already decided RISC was the way to go before they got the other processors to experiment with and test using the Tube.
Lol, that fired up some neurons that were dormant for a while.
The only thing i can remember from my assembly days is 1 accumulator, 2 registors?
My year end project in university was an ATM machine... i think i still have a printout of that project... got to look for it.
Come to think about it that may have been the 8086
That thumbnail you created, though, I saw you hold up the "6502" thing as if you were taking a mugshot! I thought you were jailed 😆🤣🤣
Great documentary video.
2 comments.
I’ve been a DJ for 36 years. Disco peaked in 1978, not 1975 as you mentioned as a notable moment of 1975.
I thought the interview of the inventor of the 5602 could have been longer. It feels like it was cut short but that’s just my opinion.
Overall A+ doc-video.
1 comment.
It's 6502, not 5602.
One comment. He answered the interview problem in another comment, but I won't say which comment.
One little but not insignificant mistake in the video is that the WDC 65C816 would be from 1985, but it actually is from 1983 - which means a backwards compatible 16 bit successor to the C64 could have been done early enough to become the ancestor of our current PCs instead of the IBM PC, with Commodore still ruling the world now.
@@NuntiusLegis The design of the WDC65C816 started in 1983, Atari and Apple received prototype samples in the second half of 1984 and the official release of the CPU was early 1985.
@@Okurka. Ok, thanks, German Wikipedia didn't tell me that. But perhaps 1984 or 1985 would have been early enough as well.
These historical perspective computing-tech videos, are why I enjoy your work.
Anything to show off the game and computer he sells.
Wow, that interview was short... seems like a missed opportunity...
I agree - I hope there's a longer unedited version to share at some point.
@@EddieSheffieldI read elsewhere here that the full interview didn't exactly... go so well. Bill went on a lot of tangents, apparently, as an example.
If anything, it felt "akward"
Fun fact, Wolfenstine for the SNES had tech in it that made it into Doom. Carmack Used it to prototype stuff like BSP-trees (which is what made Doom at the time sooo fast)
I was writing assembler for the 6502 as a 14 year old in 1980, and was hooked !!
I loved my 6510, was it, in my C64. Nice video!
I bet you loved the VIC-II and SID more.
Yes, I liked them so much I licked their shells.@@Okurka.
I would have liked to hear more about ARM from Bill Mensch. Ironic now that ARM chips are emulating 6502 game consoles.
Today I learned that due to a medical implant, I'm a cyborg powered by a 6502.
Your explanations of these chips are so inspiring. I started with a Vic-20, but only as a user with some stabs at assembly programming. And to see how such humble chips like the 6502 were the basis for so much is amazing. I personally could never be able to mentally handle all the logic you know about these chips. But I enjoy seeing how it all comes together. Seeing how things work a the chip level is astounding to me. Then knowing the current chips are 'simply' expanding on these humble beginnings.
I am so happy I made the time to sit down and watch this video. There was a ton of great tech heavy information here that was extremely easy to understand, and I feel like such a casual "retro" fan having a murky understanding of the fundamental functionalities of these computers.
The Z-80 changed a generation. In fact, it is still used embedded today.
I mainly think of it from the TRS-80, disk drives, Sega Master System, and Sega Genesis. Some arcade games, too, I think, but I don't know off hand.
Still used on TI-84 graphing calculators which are still in current production. Actually, was used for most TI graphing calculators before the TI92 (and 89 series) moved to the 68K.
@@ratdude747 Like half of the Jeopardy questions, I'm sure I *used* to know that!
I love the whole structure of the Z80, and it was indeed used by lots of older arcade machines, also don't forget the 9 million MSX home computers.
@@ZenithMusicNet The MSX was sold worldwide (but not in England or the United States).
That's why many people today don't know it. They don't know the rest of the world.
Great video David, it was very informative.
Ben Eater is the Bob Ross of breadboards
For your demonstration of 8 bit multiplication, I personally prefer to show it in terms of long multiplication, just with a different base unit, but your explanation digs a little deeper into how the logic actually works, which I really like as well. I've been playing the video game Turing Complete (on steam) and I liked taking each of the 8 different bit shift components and adding them.
Ah memories. I was taught assembly on a BBC Micro at college, then asked to recreate a 6502 in Logicworks(?) at Uni.
There is no way a 7 MHz 68000 was only 3 times faster than a 1 MHz 6502, in practical use it was more like 20 times faster. Although it required more cycles per operation, the operations were vastly superior in flexibility and capability, so most tasks could be performed with a third the operations, at 4 times the width. The 68000 also had divide and multiply of 32 bit numbers, where the 6502 only had addition and subtraction of 8 bit numbers. Needless top say the 68000 was hundreds of times faster in those tasks.
I programmed Assembly for both, and although the 6502 is very nimble for the time size and price, there is no competition between it and a 68000 in how powerful they really are.
OK you actually got to that at 14:47, and yes you had to make your own binary math for multiplication and division and numbers above 255. But even if you have the math, you sometimes use logic operations instead for fixed numbers. At least that's how I did it.
It also looks mighty. An original 68000 is an impressive beauty.
Yeah, I checked some OCS and ST demos on YT lately.
As someone who's been a game programmer for 30 years.. I miss this era.. Coding was so much more "fun" back then.. Modern day developers struggle to understand the basics of optimization... I look at people complaining about how impossible it is to run their game on the Switch (for example) and shake my head.
YES! Optimization was king back then, to the point it was more an "artform" than a skill set. And there were some damn fine "artists" back then
I agree, but that doesn’t mean it can’t still be fun. I personally don’t do a lot of actual optimization, I mostly try to make code that doesn’t run poorly and it’s pretty fun.
This makes me want to do a homebrew project for an older cpu at some point through.
@@colleagueriley860: I work with 100s of people now. And most of them think optimization is a thing you get from adding more power to the machine.
You hear them bitch and moan about power of hardware or engines in interviews with press.. it's depressing... And that's before we get to the simple fact the "higher ups" dont care and dont want you spending time/money on making it go fast.. just ship it out...
If people truly optimised code today it would be something else. Modern CPUs are so under utilised it's crazy
well, optimization back then was absolutely neccessary, because of the slow cpus, small mem etc. so there was no other way. you HAD to do it and that's why a lot of ppl became very good at it, it wasn't because it's "fun".
I'll be quite shocked if a single person here saw a Rubik's Cube in 1975.
You're right. Rubik's cube was more a 1978/1979 big thing. I remember when I was in 8th class in middle school when I was summond to the school director, wondering what "crime" I had committed, to just having the director asking me to solve his cube.
@@galier2 I remember 1980 because it was my first year at highschool and couldn't wait to get my hands on one. There may be a Hungarian or two here in their late 50s who got one before the rest of us.
Designed in 1974, licensed for sale in the UK in 1978 and worldwide in 1980 (according Wikipedia). Bought my first cube in October 1980.
@@andrewdunbar828same here - 1st year at the comprehensive school and …lets just say, someone was selling legit cubes cheap somehow. I was too naive to realise why they were cheap at the time and only twigged months later, that not everyone pays for things when they walk out of a store!
I joined a gaming company when the special graphics chip for the SNES came out and to make use of it was my first task. It's basically an 8 bit DSP with very limited capabilities, but it was reasonably good (at least compared to the 65xx) at spitting out pixels, line by line, so it could do some basic 3D rendering, a bit like the blitter on the Amiga. I recall it having instructions that are 1, 2 or 3 instructions long with delayed branching, which made some interesting optimisation tricks possible. Since it's a 8 bit processor, it would read instructions one byte per cycle, so if a branch wasn't followed by nop's, it could branch in the middle of reading long instructions, and that could be used to modify the instruction on the fly depending on the branch outcome.
When I got this chip to play with I had very limited documentation and tools, and the first thing I had to do was to write my own assembler for it. My work resulted in the game "Winter Gold".
I saved from e-waste a ceramic MOS 6502 with a date code of 3675 which makes it one of the oldest survivors. There are very few older documented date codes, at least that I have found in my online research. It is a prized part of my vintage rescue collection.
Errata: @16:30 - The Atari 2600 only has 12 bits of Address (as shown in your diagram) and can therefore only address 4 KB of memory space. The asteroids cartridge shown is an 8 KB cartridge, but it used a very crude banking scheme.
I count 13 address bits, A0 through A12. 😕
The 6507 CPU in the Atari 2600 can address 8KiB of memory space, 4KiB is devoted to the cartridge ROM and 4KiB is devoted to RAM, I/O, timer video and audio registers inside the RIOT and TIA chips.
I came to say something similar. BrainSlugs83 is close, but NerdlyPleasures has it right. I'm currently working on my first Atari 2600 game.
The 6507 also still has internal vectors for interrupts that have been moved to fit into the modified address space, even though it no longer has the external pins that trigger those interrupts.
@@NerdlyPleasures This is correct. A12 is mapped to the cartridge ROM chip select.
The 8 bit baby, awesome! You were so cute!
Jeffrey Dahmer was also cute as a baby.
there are 6502 cores in many household junk like an electric toothbrush. I put them there, sorry
What is a toothbrush using a 6502 for? Am legitimately curious what computation is necessary for say a timer function!
@@WomblingFreely battery controller is my guess
That was a good video! Congratulations David, looks like the 8-bit guy is back!
One thing that occurs to me while I'm watching these videos, is how ubiquitous computers are in our lives.
I'm a truck driver, but I still rely massively on computers to do my job...
I have a cell phone, I have a SAT nav, I have a tablet, the timing, horsepower, exhaust, and RPMs of my trucks engine are monitored by computer, my dash cam is run on a computer...
And, I have a Steam Deck in the bench of the truck...
All of these devices Trace lineage back to the fascinating technology that my mans here is explaining, and it's all so fascinating. And frankly, a little incredible.
yeah they are more ingrained in our lives then we realize🤣🤣
I don’t think it’s reasonable to say that the 6502 powered a generation. At least in the U.K. the ZX80/81 and the hugely popular Spectrum were Z80. My first micro, a TRS-80, as also Z80. The Z80 ran CP/M and was also a far easier chip to program.
The Z80 powered the poor part of that generation.
@@Okurka. Poor? Not at all. It was used in many business machines. It was the basis of the MSX computer standard. It was also widely used in games consoles. It had a rich instruction set, including 16-bit instructions.
But the BBC Micro had a 6502 and that fact eventually led to the creation of ARM in the further history of the Acorn company, so I think the impact of the 6502 even in the UK is probably higher at least from today's standpoint. By "powering a generation" in the 80 of course you are correct, that probably goes to the Z80 for Sinclair's computers being cheap.
@@mudi2000a The BBC Micro was expensive. That’s why the cheaper ZX Spectrum was so popular.
I’m not saying that the 6502 wasn’t hugely popular, just that the Z80 was too, and it was also very successful so that the suggestion that the 6502 powered a generation of computers was wrong. It was a mixed market, and remained like that for a long time.
@@sjzara yes I didn’t deny that but at the end of the day the ARM processor was developed by the people that created the BBC Micro.
Yes, but sorry - I still prefer Z80.
It is true. Everyone has their preference. If someone insults you because of this comment, they are wrong.
Exactly! Z80 is much more convenient to use than 6502.
Have you heard that Z80 will be discontinued in a few weeks. Sad really.
@@jbponzi1 Yes, it's sad for me, too, especially as I am an old and sentimental guy. But, on the bright side, the Z180 still lives; I wrote a lot of code for the Z180. And also the eZ80 - already over 20 years since introduced; seems like yesterday. Life is so short and CPUs come and go! 🙂
@@jbponzi1No, I have not heard that! That's terrible!
Back in the early 80's I wrote many thousands of lines of 6502 assembly. Aside from being fun, it taught me a lot about programming, conserving memory, writing efficient code, coding with minimal instructions, and a lot besides.
Oh and of course if you are coding exclusively in assembly and burning your code into Eprom, you certainly learn about writing error-free code
65c816 cores were also used in Micronas cdc16fxx series chips used in cars - interestingly, a large part of WV Passats and Audis from 2000 used them in dashboards and other brands in fuse boxes ;) I used these chips to reverse engineer the Audi dashboard, and I still write Atari XL/XE demos ;)
My first experience with the 6502 was the Acorn Atom computer from 1979 (I kept working with the Acorn line of computers until the RISC-PC). As the Acorn Atom has a BASIC that had a build-in compiler (you could use mnemonics like STA, JMP, ROR and so on, and even labels to jump to), I learned coding in machine code for the 6502 pretty fast. It helped, of course, that I already did machine code on my first single board computer - The Cosmac Super Elf, with an RCA 1802 processor and 265 bytes of RAM. I restored that Super Elf into working condition by the way, and am now busy doin the same for the Acorn Atom. There are still some hurdles to go, but that's the fun of it...
Anyway - I really liked that 6502 processor. It was easy to write programs for (in machine code of course), and I wrote several toolkits (burned in EPROM) for the BBC-B computer. Great times...
How don’t you trash the BASIC registers and zero page?
@@ArneChristianRosenfeldt I am not sure what you mean. If you use the in-BASIC compiler, you could specify where your compiled code runs. By default it's above zero page of course. In the BBC-B computer I often used a sideways RAM board to store my compiled code, and run it like a sideways ROM. I must admit I am VERY rusty about that, because it's more than 40 years ago I did all that stuff. I hope by rebuilding the Acorn Atom into working order I get a platform to restore that rusty knowledge once again. 🙂
@@jclosed2516 Yesterday I wanted to try to JMP() , but ultimately could not use this instruction. But again it was right before my eyes: Code does not come from the zero page. Jumps are either relative or 16 bit absolute even in the indirect case or with JSR or RTS or RTI.
I also found out that in the Commodore C16 the zero page is free for me to use here D0--E8 ( sic in the book I have, not --E7). Not much. I also have to PHA;TXA;PHA
My first computer was a clone of the ZX Spectrum (TK-95), so I always really liked the Z-80, but this video got me really interested in the 6502 and really gave me a lot of information about this CPU that I didn't know.
Great job on the video! I learned about the 65C816 from SNES specification and development documents many years ago then eventually learned about the original NES and C64. Your explanations about pins and such is a great for further learning how the chips were related and even cross compatible.
Great Video!!! It's like old times, learning stuff in a simple well laid out video. Thanks!!!
Thanks for the walk down memory lane! It was fun to hear about all the 6502 variants, and all the machines that used them. I bought an Apple //e in 1983 and spent several happy years programming it in assembly until moving to the MS-DOS world and C a few years later. I had no idea its 65C02 supported new instructions. I'm sure I could have taken advantage of that at the time, as speed and code size were always at a premium. Now I'm thinking I should build a 14MHz W65C02 machine and port some of that old code to run on it. Young me would have thought he died and went to heaven with that clock rate!
I really appreciate your videos. There is something about them that has awakened my passion for mid-to-late 90s, early 2000s technology. I got into computer repair in 1996 and have been in love with them since. Just started teaching myself assembly and C and wow, much different than hardware :) Btw, I wish I had a picture of my grandfather. You look exactly like he did at our age lol.
Thank you for this video!
I have bought an Acorn Atom in 1981, expanded it at first with piggyback 2114's, but at the end I had rebuilt it completely in a 19 inch rack running on 2MHz. I also had extensions like a 80 column video card, an eprom programmer, multiple floppy disk drives, a ram disk, a telex signal card and even a IEEE 488 controller card.
My first computer was a KIM-1. Took a class at the Byte Shop of Palo Alto on the Z80, but never resorted to assembler/machine code on that family. Just felt bad they had more registers till I realized the 6502 had a whole PAGE of 'em... Had as OSI Superboard, eventually Apple II. My favorite processor, so clear and straightforward. 6502 FOREVAH! My KIM-1 eventually ended up in a briefcase with a power supply and a a bunch of protoboard as I tried (with no training) to build a digital synth during the day, then play amazing fusion jazz at night. FUN times!
BTW, 8-Bit Guy, don't forget the Ohio Scientific Superboard series. My first Basic on-board.
Brilliant vid. I've picked up bits and pieces from these vids before, but it was great to have it all in one. Hopefully we'll see some similar vids on other chips in the future.
10:15 and 1.79 MHz is half the frequency of the 3.58 MHz color subcarrier used for NTSC video, which is why you see those particular numbers show up so frequently in clock speed. Some SNES games used the SA1 chip, which contains a second 65816 running at a blazing fast 10.74 MHz.
Coming from the Tandy world, I was always a Z80 person; I wish now that I had diversified into the 6502 as well. Back then it was almost a Hatfield vs McCoy's. Thank you for presenting more into my little world than I expected!