1:28 It wasn't the first microprocessor designed for non-calculator computers: the Intel 8008 was designed for a "real" microcomputer, the Datapoint 2200. The Datapoint machine actually ended up using the 8008's instruction set implemented in TTL(!), but the 8008 was used in a number of early machines including the Micral N from 1973. And the 6502 had its roots in the Motorola 6800: its sister the 6501 was designed to be pin-compatible with the 6800.
I was a technician at CalComp in '73-'74. They made flatbed plotters used in the semiconductor industry. One of the test plots we would run was called IC Mask on Strippable film.
Nice talk but some definite confusion here. First and most important, the 6502 is absolutely not RISC. You could perhaps call it minimal instruction set (though nowhere near as minimal as say PDP8 or Nova), but heavily microcoded is not RISC. The C in 65C02 stands for CMOS. They currently (Aug 2020) cost $6.95 for 1, $5.09 each for 250. The 6502 did not use several times fewer clock cycles than the 6800, maybe you were thinking of 8080/Z80? Normal 6800 instructions ranged from 2 to 9 cycles with most instructions 3 or 4, just as on the 6502. JSR was maybe the 6800's worst with 9 vs 6, but RTS was actually better at 5 vs 6. SWI and RTI were a little more than 9 cycles but they're very uncommon. The other big difference was conditional branches took 4 cycles on 6800 whether taken or not, while 6502 is 2 cycles not taken and usually 3 taken. Otherwise equivalent instructions generally took the same 1 cycle per byte transferred on both. The 6800 even has shorter and faster zero page instructions with the same execution time as 6502. What the 6800 *doesn't* have is the ability to indirect through pointers in zero page, let alone index off them. So the 6502's "LDY #13; ADC (42),Y" (4 bytes, 7 cycles) becomes "LDX 42; ADC A, 13,X" (4 bytes, 9 cycles). No biggie. The problem comes when you'd like a loop such as "LDA ($42),Y; STA ($69),Y; DEY; BPL LOOP" (7 bytes, 16 cycles) when the 6800 needs something like "LDX 42; LDA A 0,X; DEX; STX 42; LDX 69; STA A 0,X; DEX; STX 69; DEC B; BPL LOOP" (17 bytes, 42 cycles) which is 2.6 times slower. (if you're really really sure you won't cross a page boundary you could do "DEC 0042" instead of "DEX; STX 42" but it's the same length and only saves 3 (6) cycles). Motorola's 6809 simplifies the programming compared to 6800 with "LDA B,X; STA B,Y; DECB; BPL LOOP" (7 bytes, 15 cycles) but still isn't significantly faster than the 6502! (and you also need to load X and Y pointers before the loop)
It wouldn't be too long after the 6502 was developed when the first computerized systems for designing chips started going into production, and Linotronic machines would be used to made the photomasks.
One of the best things about the 6502 vs the Intel processors that came before it was its clock architecture. Instead of retiring multiple clock phases, it performed a memory access on each clock, so that am Intel processor required a much higher clock frequency to execute code as fast as the 6502 running at 1MHz, and the CPU architecture was also static rather than dynamic so you could save power by allowing the clock or even "sleep" by stopping the clock and resuming it later.
The Zilog Z80 was around about 4 or 5 years before the 6502. And it's still with us - most home and business alarm systems use the Z80 or it's later versions.
Whilst I enjoyed the talk there is a lot of info here that is not accurate. However the 6502 architecture to this day is still very elegant a joy to program.
The WDC 65C816 can be bought from Mouser Electronics for $6.95 in single quantity. That would be equivalent to $1.48 in late 1975. So, in 44 years, the cost of a "6502" has dropped by a factor of 16.9, and by the way, you now get 16-bit mode for free.
The 'C' in the middle of the part number was the standard way of indicating a CMOS version of a part that originally was NMOS (or PMOS). Example: www.intersil.com/en/products/space-and-harsh-environment/-883/microprocessors-and-peripherals/80C86-883.html And why would one care? Because the CMOS part uses less power (and thus produces less heat) and because it is (much) more sensitive to static electricity. There are usually also differences in how much current the parts can sink and source and at what voltages above or below which they say "that is certainly a 1/0". CMOS parts can also (easier) output full-swing voltages (0/5V or whatever the Vcc is).
This is an interesting take on the "History of the 6502" lots right in this but could use a chat with Mr. Mensch to refine the information. This was fun to watch and hear the questions here and there. WDC has some new board level products which could be helpful to play with to become more familiar with the technology. The Xxcelr8r system is pretty cool. All WDC parts are available at Mouser Electronics.
Yep, sure is. The WDC 65C816 is $6.95 at Mouser in single quantity. That would be equivalent to $1.48 in late 1975. So, in 44 years, the cost of a "6502" has dropped by a factor of 16.9, and by the way, you now get 16-bit mode for free.
The fun part about the 6502 was that it could use page 0 memory (the first 256K) for indirect addressing which sort of gave it 128 more "registers". I wrote this BEFORE he started to talk about zero page addressing....................
The Atari 800 used BCD floating point in their operating system. When it came time to put out a Basic language, the originally bought the source code for MS Basic BUT they needed it to do two extra things : syntax checking when a line is entered and fit into an 8K cartridge. They never were able to get it to work so they hired another company (Shepardson Systems {or something like that}) to write a new basic. In order to make it fit into 8K they used the built in BCD system in the OS for all mathematics (including calculating line numbers). The end result was the slowest basic language at the time. Yes, it had other features that made it decent but it was sad when I claimed Atari was superior to Apple only to have them use a MS Basic program to "prove" their point. After a while MS came out with a MS basic for the Atari and I was able to blow them out of the water with their own demos. (their was a big debate Apple vs Atari at the time. Logic said Atari was vastly superior but Apple lovers fell back to using Basic program for their "proof".
Ok, so it wasn't the quickest BASIC available. But it was a delight to use, its refusal to accept bad syntax and highlight where on a line the problem was. Strings were blindingly fast and you could also store relocatable machine code in them. I never enjoyed using Microsoft Basic on the Atari, probably because I had become accustomed to squashing most errors/bugs before running my code.
@@monetize_this8330 actually it was the OS's floating point library that made Atari Basic so slow. The language itself was pre-compiled into P code hence it didn't have to interpret anything at run time. Turbo Basic fixes that problem and runs about 4 times faster than Atari Basic. LOL - When Microsoft Basic came out for the Atari 800 they actually used the fact that MB would find the syntax errors during run time (when they should be found according to what they wrote). That had to be one of the biggest pieces of B.S. I've ever heard.
@@bjbell52 I used to have a program called ABC "Atari BASIC compiler" which converted it directly into machine code. It wasn't ideal, because it forced you to re-write your BASIC in a certain style. But it was ok to use if you were creating stuff from scratch.
This guy really doesn't know much of anything, does he? All of his information comes from Google searches and Wikipedia. And half of what he says he even admits that he's not sure about. He doesn't even know that "0183" means the first MONTH, not the first QUARTER of the year.
Howard Scott Warshaw wasn't anybody. He created Yars Revenge and worked for Atari when he created E.T. Which is not even close to the worst game ever made. Good talk on the 6502, I always enjoy these retro cpu talks.
Wasn't the TRS-80 from Radio Shack which used a Z80 processor come out before computers with a 6502 Apple, or Atari 800? I always loved people who clamed the Apple was a superior computer to an Atari 800. I always asked why and they couldn't ever tell me (even I knew on area the Apple was better than the Atari - they used a parallel I/O instead of a serial I/O Atari used). Otherwise, they used the same processor. Early Atari games were ports from Apple which didn't need very much changing to translate from Apple to Atari. The difference (when they didn't use the Atari's built in graphics co-processor) was that the Apple ran at 1mhz where the Atari ran almost 80% faster at 1.79mhz.
bjbell52 Wikipedia says: The TRS-80 was introduced in august 1977, and the Atari-800 in 1979. The Apple ][ was on the market in april of 1977. (The Apple one was april 1976)
The TRS-80 was announced first but ended up shipping at the very end of 1977 after the Apple ][ and Commodore Pet. The thing which I loved about the Apple ][ at the time was the cheap, very fast (15 KB/s -- fill or dump the whole RAM in 4 seconds), high capacity floppy disk drives and 80 column cards that were available very early on (mid 1978). It was a serious computer with a good keyboard that could be used for work all day for word processing, spreadsheets, databases, accounting, and custom programming in proper languages such as Pascal.
First real microprocessor ... FirstMicroprocessor.com MOS 6502 because popular around 72-74. I used it in the Jolt, Super Jolt and SYM Computer Board. Ray Holt
This talk is an outstanding example of why everything does not need to end up on TH-cam.
"0183" indicates the 6502A pictured was made in the first week of 1983.
1:28 It wasn't the first microprocessor designed for non-calculator computers: the Intel 8008 was designed for a "real" microcomputer, the Datapoint 2200. The Datapoint machine actually ended up using the 8008's instruction set implemented in TTL(!), but the 8008 was used in a number of early machines including the Micral N from 1973. And the 6502 had its roots in the Motorola 6800: its sister the 6501 was designed to be pin-compatible with the 6800.
did Bill gates write programs for the 8008?
I was a technician at CalComp in '73-'74. They made flatbed plotters used in the semiconductor industry. One of the test plots we would run was called IC Mask on Strippable film.
great seminar video of the MOS
Nice talk but some definite confusion here. First and most important, the 6502 is absolutely not RISC. You could perhaps call it minimal instruction set (though nowhere near as minimal as say PDP8 or Nova), but heavily microcoded is not RISC. The C in 65C02 stands for CMOS. They currently (Aug 2020) cost $6.95 for 1, $5.09 each for 250. The 6502 did not use several times fewer clock cycles than the 6800, maybe you were thinking of 8080/Z80? Normal 6800 instructions ranged from 2 to 9 cycles with most instructions 3 or 4, just as on the 6502. JSR was maybe the 6800's worst with 9 vs 6, but RTS was actually better at 5 vs 6. SWI and RTI were a little more than 9 cycles but they're very uncommon. The other big difference was conditional branches took 4 cycles on 6800 whether taken or not, while 6502 is 2 cycles not taken and usually 3 taken. Otherwise equivalent instructions generally took the same 1 cycle per byte transferred on both. The 6800 even has shorter and faster zero page instructions with the same execution time as 6502. What the 6800 *doesn't* have is the ability to indirect through pointers in zero page, let alone index off them. So the 6502's "LDY #13; ADC (42),Y" (4 bytes, 7 cycles) becomes "LDX 42; ADC A, 13,X" (4 bytes, 9 cycles). No biggie. The problem comes when you'd like a loop such as "LDA ($42),Y; STA ($69),Y; DEY; BPL LOOP" (7 bytes, 16 cycles) when the 6800 needs something like "LDX 42; LDA A 0,X; DEX; STX 42; LDX 69; STA A 0,X; DEX; STX 69; DEC B; BPL LOOP" (17 bytes, 42 cycles) which is 2.6 times slower. (if you're really really sure you won't cross a page boundary you could do "DEC 0042" instead of "DEX; STX 42" but it's the same length and only saves 3 (6) cycles). Motorola's 6809 simplifies the programming compared to 6800 with "LDA B,X; STA B,Y; DECB; BPL LOOP" (7 bytes, 15 cycles) but still isn't significantly faster than the 6502! (and you also need to load X and Y pointers before the loop)
It wouldn't be too long after the 6502 was developed when the first computerized systems for designing chips started going into production, and Linotronic machines would be used to made the photomasks.
At the current state or the art you could conceivably fit a 6502 in a white blood cell and still have room left over.
One of the best things about the 6502 vs the Intel processors that came before it was its clock architecture. Instead of retiring multiple clock phases, it performed a memory access on each clock, so that am Intel processor required a much higher clock frequency to execute code as fast as the 6502 running at 1MHz, and the CPU architecture was also static rather than dynamic so you could save power by allowing the clock or even "sleep" by stopping the clock and resuming it later.
The Zilog Z80 was around about 4 or 5 years before the 6502. And it's still with us - most home and business alarm systems use the Z80 or it's later versions.
No, 6502 came in 1975, Z80 in 1976. Though 8080 was in 1974.
6502 is still with us also and used in a lot of embedded systems too but your dates are wrong there
Whilst I enjoyed the talk there is a lot of info here that is not accurate. However the 6502 architecture to this day is still very elegant a joy to program.
The WDC 65C816 can be bought from Mouser Electronics for $6.95 in single quantity. That would be equivalent to $1.48 in late 1975. So, in 44 years, the cost of a "6502" has dropped by a factor of 16.9, and by the way, you now get 16-bit mode for free.
super guter vortrag! gut, dass auch die kosten und entwicklung besprochen wurden.
Pretty sure the 'C' in 65C02 is for CMOS.
That makes sense. I think I got the idea that it could mean "compatible" from descriptions like this one: en.wiki2.org/wiki/WDC_65816/65802
The 'C' in the middle of the part number was the standard way of indicating a CMOS version of a part that originally was NMOS (or PMOS).
Example:
www.intersil.com/en/products/space-and-harsh-environment/-883/microprocessors-and-peripherals/80C86-883.html
And why would one care? Because the CMOS part uses less power (and thus produces less heat) and because it is (much) more sensitive to static electricity. There are usually also differences in how much current the parts can sink and source and at what voltages above or below which they say "that is certainly a 1/0". CMOS parts can also (easier) output full-swing voltages (0/5V or whatever the Vcc is).
This is an interesting take on the "History of the 6502" lots right in this but could use a chat with Mr. Mensch to refine the information. This was fun to watch and hear the questions here and there. WDC has some new board level products which could be helpful to play with to become more familiar with the technology. The Xxcelr8r system is pretty cool. All WDC parts are available at Mouser Electronics.
Yep, sure is. The WDC 65C816 is $6.95 at Mouser in single quantity. That would be equivalent to $1.48 in late 1975. So, in 44 years, the cost of a "6502" has dropped by a factor of 16.9, and by the way, you now get 16-bit mode for free.
The fun part about the 6502 was that it could use page 0 memory (the first 256K) for indirect addressing which sort of gave it 128 more "registers".
I wrote this BEFORE he started to talk about zero page addressing....................
@BOOZE & METAL I meant bytes and typed K. Opps.
I think everybody "felt" they needed BCD because IBM used it in their big iron.
I worked with Chuck Peddle in the 1990s.
What did you do?
The Atari 800 used BCD floating point in their operating system. When it came time to put out a Basic language, the originally bought the source code for MS Basic BUT they needed it to do two extra things : syntax checking when a line is entered and fit into an 8K cartridge. They never were able to get it to work so they hired another company (Shepardson Systems {or something like that}) to write a new basic. In order to make it fit into 8K they used the built in BCD system in the OS for all mathematics (including calculating line numbers). The end result was the slowest basic language at the time. Yes, it had other features that made it decent but it was sad when I claimed Atari was superior to Apple only to have them use a MS Basic program to "prove" their point. After a while MS came out with a MS basic for the Atari and I was able to blow them out of the water with their own demos. (their was a big debate Apple vs Atari at the time. Logic said Atari was vastly superior but Apple lovers fell back to using Basic program for their "proof".
Ok, so it wasn't the quickest BASIC available. But it was a delight to use, its refusal to accept bad syntax and highlight where on a line the problem was. Strings were blindingly fast and you could also store relocatable machine code in them.
I never enjoyed using Microsoft Basic on the Atari, probably because I had become accustomed to squashing most errors/bugs before running my code.
@@monetize_this8330 actually it was the OS's floating point library that made Atari Basic so slow. The language itself was pre-compiled into P code hence it didn't have to interpret anything at run time. Turbo Basic fixes that problem and runs about 4 times faster than Atari Basic.
LOL - When Microsoft Basic came out for the Atari 800 they actually used the fact that MB would find the syntax errors during run time (when they should be found according to what they wrote). That had to be one of the biggest pieces of B.S. I've ever heard.
@@bjbell52 I used to have a program called ABC "Atari BASIC compiler" which converted it directly into machine code.
It wasn't ideal, because it forced you to re-write your BASIC in a certain style. But it was ok to use if you were creating stuff from scratch.
This guy really doesn't know much of anything, does he? All of his information comes from Google searches and Wikipedia. And half of what he says he even admits that he's not sure about. He doesn't even know that "0183" means the first MONTH, not the first QUARTER of the year.
Yeah - the 01 means 1st week
He seemed vaguely interested, but.. not interested enough to actually remember details.
The blind leading the blind.
I agree. He often admits to not being familiar with... not sure about...
First week actually.
It was a talk to a local group, and as a talk to a local group it's ok.
Howard Scott Warshaw wasn't anybody. He created Yars Revenge and worked for Atari when he created E.T. Which is not even close to the worst game ever made. Good talk on the 6502, I always enjoy these retro cpu talks.
I agree 👍💯
Wasn't the TRS-80 from Radio Shack which used a Z80 processor come out before computers with a 6502 Apple, or Atari 800?
I always loved people who clamed the Apple was a superior computer to an Atari 800. I always asked why and they couldn't ever tell me (even I knew on area the Apple was better than the Atari - they used a parallel I/O instead of a serial I/O Atari used). Otherwise, they used the same processor. Early Atari games were ports from Apple which didn't need very much changing to translate from Apple to Atari. The difference (when they didn't use the Atari's built in graphics co-processor) was that the Apple ran at 1mhz where the Atari ran almost 80% faster at 1.79mhz.
bjbell52 Wikipedia says:
The TRS-80 was introduced in august 1977, and the Atari-800 in 1979.
The Apple ][ was on the market in april of 1977. (The Apple one was april 1976)
The TRS-80 was announced first but ended up shipping at the very end of 1977 after the Apple ][ and Commodore Pet. The thing which I loved about the Apple ][ at the time was the cheap, very fast (15 KB/s -- fill or dump the whole RAM in 4 seconds), high capacity floppy disk drives and 80 column cards that were available very early on (mid 1978). It was a serious computer with a good keyboard that could be used for work all day for word processing, spreadsheets, databases, accounting, and custom programming in proper languages such as Pascal.
BCD is great.
I LOVE 6502 😍❤️👍🥂🎩
E.T. was made by Atari by Howard Scott Warshaw. It wasn’t shovel ware created by some 3rd party.
What does the "65xx" mean?
The closest to 6800 to be not considered as a copy.
The company name is pronounced M, O, S, not moss. I know several former employees there. Let's please get the history correct.
I tried the WiFi password, but.. it does not work...
...j/k nice talk :)
First real microprocessor ... FirstMicroprocessor.com MOS 6502 because popular around 72-74. I used it in the Jolt, Super Jolt and SYM Computer Board. Ray Holt
C-
Valley girl hipster says what?
interesting talk let down lazy and poor editing. Show the slides as the main part of the screen.
What? There is no editing.
Seems like this is not well researched.
Dreadful!
Giy doesn't know anything and can't speak smoothly.