Talk:74181

Comment 1

If you are aware of other computers or devices that used the 74181 or 74S181 please add them to the list. 74s181 13:51, 27 October 2007 (UTC)

If you have a citable reference on the history of the 74181, like, who first introduced it, when, etc., please add a history section. Right now I think that Fairchild introduced it in 1967 but I can't prove it. 74s181 13:51, 27 October 2007 (UTC)

I'm also thinking of adding some info about Wayne D. Pickette, who supposedly was inspired by the 74181 to invent the microprocessor... 74s181 13:51, 27 October 2007 (UTC)

[1] hints at some of these things. -- RoySmith (talk) 02:46, 28 October 2007 (UTC)

Also look at [2] -- RoySmith (talk) 02:51, 28 October 2007 (UTC)

I am Wayne Pickette, the true story is that I was inspired by Asimov's I Robot, I then received validation for my thoughts with the 74181 that a computer could be formed on a chip or at least on a multi-chip hybrid. I uploaded image:100_737 it is a picture of the original drawing I presented to Dr. Noyce in 1970 then Dr. Hoff to whom Dr. Noyce took me to. Any further questions you may contact me at waynedougpick@yahoo.com/ I'm currently patenting items to address climate change from 2008 onward. -- (talk)Wankap (talk) 08:46, 22 November 2007 (UTC)

Thanks for your support. I have seen both of these articles, do you think they meet WP:RS? 74s181 02:58, 28 October 2007 (UTC)

Here's another one [3]. They require a little interpretation, but as first-hand recollections of the man himself, I don't see how you're going to get much more reliable than than. -- RoySmith (talk) 02:59, 28 October 2007 (UTC)

Uh, RoySmith, the SpaceWars reference you added is my personal website. Not exactly a WP:RS. 74s181 05:44, 28 October 2007 (UTC)

Other references are great! Thanks, RoySmith! 74s181 06:17, 28 October 2007 (UTC)

When

The {{when}} template does not require an entry on the talk page; it's just a suggestion. I've replaced it again, since "today's microprocessors" is a temporal phrase. The microprocessors of "today" evolve pretty continuously with new releases, designs, and fabrication processes. As worded, the article is not clear in what it means, and is using this vague wording to avoid being specific and accurate. The uncited text in the "today" section should explain the specific transition involved, and what evolution caused it. Otherwise, it should be removed from the article. -- Mikeblas 01:44, 2 November 2007 (UTC)

I think you're being a bit silly with this {{when}} business. It is important to provide a time frame when a statement may only be transiently true. If I say, Super Gonzo Space Invaders is the world's best selling video game, that's the kind of statement which needs to be qualified with, as of November 2007, because it's the sort of thing which changes often. On the other hand, Antarctica contains the statement, Antarctica is Earth's southernmost continent. Technically, that's only true for now. In another 200 million years, the plates will move around and it may no longer be true. Would you put a {{when}} tag on Antarctica? No, of course not. The same logic applies here. The 74181, however notable it is for its historic uses, is effectively obsolete. It will never again become cost-competitive with contemporary modern chips. Yes, it would be useful and interesting for the article to explore the exact history of when the 74181 faded from being the top of the heap, but the lack of such an exploration does not detract from the current statement. I'm removing the tag. -- RoySmith (talk) 14:05, 2 November 2007 (UTC)

Your analogy doesn't apply very well because of the obvious difference in time scope. The "today's" problem does detract from the article because it's not concrete when a concrete statement can easily be written. As such, I've taken a crack at resolving the temporal problems in this section myself. -- Mikeblas 14:30, 2 November 2007 (UTC)

I don't have time to fix this right now, but this statement is still incorrect. The '181 continued to be competitive with early microprocessors due to performance reasons, but when the microprocessors got to be faster than the fastest CPU you could build with the '181 (probably around the time of the 80486), the '181 was no longer commercially competitive. It definitely wasn't because microprocessors got their own ALUs, they've had ALUs since the beginning, but they were slow in comparison. Or, to put it another way, microprocessors have always had the advantage of simplicity, one chip replaced an entire circuit board, but as long as uPs were slower, the '181 continued to be used in 'real' computers. When the uP-based computers started outperforming 'real' computers, the '181 was no longer commercially viable. —Preceding unsigned comment added by 74s181 (talk • contribs) 19:07, 2 November 2007 (UTC)

See Killer micro -- RoySmith (talk) 19:10, 2 November 2007 (UTC)

4 bit slice of an arithmetic logic unit

Although "4 bit slice of an arithmetic logic unit" may sound more grammatically correct, it isn't quite factually correct. It isn't a 'slice' of an 'ALU'. See Bit slicing. It is a full blown 4 bit wide ALU that is designed to be connected to with other identical parts for larger word lengths, that is what 'bit slice' means.

Let's talk if you disagree with the current wording, "The 74181 is a bit slice arithmetic logic unit (ALU), implemented as a 7400 series TTL integrated circuit. " 74s181 00:16, 3 November 2007 (UTC)

microprocessors that have integrated ALUs

I have never heard of a microprocessor without an integrated ALU. I think this statement comes from a misunderstanding of what an ALU is, or what a microprocessor is, or both.

There are various reasons why the '181 is no longer 'commercially viable' compared to microprocessors today.

1. A CPU built using the 74181 requires many components, even the Pathetic Instruction Set Computer (PISC) requires 22 chips. A nicroprocessor is a single component, it is inherently smaller, more reliable and cheaper to mass produce.
2. A CPU built using the 74181 will use more power and dissipate more heat than any modern microprocessor of equal computing power.
3. A CPU built using the 74181 or even faster discrete logic families (like ECL) is limited by speed of light delays inherent in the size of the chips and the spacing between them. Believe it or not, part of the reason that today's microprocessors are able to run faster than the fastest supercomputers built from individual components (like the '181) is because all the parts of the microprocessor CPU are within a few millimeters of each other.

Even though #1 and #2 were true from the very beginning of the microprocessor age, computers based on the '181 and later TTL and ECL bit slice components still were commercially viable for several years because the first microprocessors were so slow. The 8 bit 8080 could execute up to 500,000 instructions per second and was popular from 1975 until the IBM PC came out in 1980. The 16 bit Data General Nova 3, a popular minicomputer of that time period based on the '181 executed instructions in about 800 nanoseconds, or about 1.2 million instructions per second. Not only was it three times faster in raw instruction execution, it was a true 16 bit computer, my experience was that it was 10-20 times faster when running real applications. But the IBM AT (based on the 80286) came close to the Nova 3 in performance, supported more RAM, and was 1/10th the cost ($50,000 for a Nova 3 with hard drive, terminal, and printer, vs $5,000 for a similarly equipped IBM AT). 74s181 01:26, 3 November 2007 (UTC)

Also, because the section is about the present status of the '181, I changed the section title back to 'Today'. 74s181 01:27, 3 November 2007 (UTC)

Tek 4052

I've removed this. A message at http://www.classiccmp.org/pipermail/cctalk/1999-April.txt.gz indicates that the 4052 used the 2901, not the 74181. -- RoySmith (talk) 02:35, 3 November 2007 (UTC)

Why 4 bits?

It would be interesting to explore why 4 bits became the standard for how wide to make a bit-slice part? Why did we never see 8 bit slices? -- RoySmith (talk) 14:36, 3 November 2007 (UTC)

Speculation:

1. 4 bits = binary coded decimal digit.
2. 8 bit ALU would have required a 36 pin DIP, no such thing at that time.
3. 4 bits = hexadecimal digit.
4. 8 bit ALU would have more than twice as many gates (for look-ahead carry) and wouldn't fit on a 74xx TTL chip manufactured with the processes available at that time.

I think that #1 is most likely, but this is speculation. 74s181 22:34, 3 November 2007 (UTC)

There were 8-bit slices, e.g., the Fairchild 3800 MOS 8-bit parallel accumulator from around 1969, and the TI SN74AS888 from the late 1970s. There were also 2-bit slices, e.g., Intel 3002.

Also Fairchild did made some early parts in a 36-bit DIP, e.g., the U6A903059X (later renumbered MuL9030) eight-bit RAM, which was in production in 1967 or earlier, and used in the DEC KA10 processor for the PDP-10. It is in the Fairchild 1969 Data Catalog. --Brouhaha (talk) 20:18, 29 January 2011 (UTC)

Less active.

There is only so much time in the day, and there are priorities in my life that need more time and attention right now. So, while I don't expect to be completely absent from Wikipedia, don't be surprised if days go by without any contributions from me. 74s181 14:19, 14 November 2007 (UTC)

History of 74181

The article lacks a history of 74181, such as: who or which manufacturer developed it first etc. Cogiati (talk) 15:07, 17 June 2011 (UTC)

I agree, should have a history. I believe that the chip was developed by Fairchild but when I wrote the article I couldn't find any references to support this. 74s181 (talk) 18:53, 22 July 2011 (UTC)

Logic table is wrong

The logic table is wrong and incomplete.

It is wrong because, for active-low, the signal "L" from figure 1 of the data sheet translates to logical one, not to zero.

It is incomplete gives half the data, because only half the data is given (Cn=0 for active-low, Cn=1 for active-high). --Trex4321 (talk) 16:07, 12 September 2021 (UTC)

OK, I fixed it. Trex4321 (talk) 15:38, 17 September 2021 (UTC)