Groups | Search | Server Info | Login | Register


Groups > comp.arch.fpga > #38730

Re: fast divider?

From john larkin <jl@glen--canyon.com>
Newsgroups sci.electronics.design, comp.arch.fpga
Subject Re: fast divider?
Date 2026-04-01 09:12 -0700
Organization A noiseless patient Spider
Message-ID <26gqskthcr1qfvkm62qh6qjg7cb0ipg4bu@4ax.com> (permalink)
References (7 earlier) <10qgb98$39jr5$1@dont-email.me> <6honskhhs9lpc60c05dcn16v9pooqe8udp@4ax.com> <10qi8er$3tkci$3@dont-email.me> <cnjpsk9ev6uphcdeoqhnoeu37vb5epovd9@4ax.com> <10qjcna$ad1m$1@dont-email.me>

Cross-posted to 2 groups.

Show all headers | View raw


On Thu, 2 Apr 2026 02:13:38 +1100, Bill Sloman <bill.sloman@ieee.org>
wrote:

>On 1/04/2026 7:06 pm, john larkin wrote:
>> On Wed, 1 Apr 2026 15:54:44 +1100, Bill Sloman <bill.sloman@ieee.org>
>> wrote:
>> 
>>> On 1/04/2026 2:14 am, john larkin wrote:
>>>> On Tue, 31 Mar 2026 22:30:44 +1100, Bill Sloman <bill.sloman@ieee.org>
>>>> wrote:
>>>>
>>>>> On 31/03/2026 8:40 pm, john larkin wrote:
>>>>>> On Tue, 31 Mar 2026 16:35:49 +1100, Bill Sloman <bill.sloman@ieee.org>
>>>>>> wrote:
>>>>>>
>>>>>>> On 31/03/2026 2:00 am, john larkin wrote:
>>>>>>>> On Mon, 30 Mar 2026 16:42:12 +1100, Bill Sloman <bill.sloman@ieee.org>
>>>>>>>> wrote:
>>>>>>>>
>>>>>>>>> On 30/03/2026 2:18 am, john larkin wrote:
>>>>>>>>>> On Sun, 29 Mar 2026 15:52:53 +1100, Bill Sloman <bill.sloman@ieee.org>
>>>>>>>>>> wrote:
>>>>>>>>>>
>>>>>>>>>>> On 29/03/2026 8:38 am, john larkin wrote:
>>>>>>>>>>>> On Sat, 28 Mar 2026 16:44:40 +1100, Bill Sloman <bill.sloman@ieee.org>
>>>>>>>>>>>> wrote:
>>>>>>>>>>>>
>>>>>>>>>>>>> On 28/03/2026 5:39 am, john larkin wrote:
>>>>>>>>>>>>>> On Sun, 22 Mar 2026 03:00:16 +1100, Bill Sloman <bill.sloman@ieee.org>
>>>>>>>>>>>>>> wrote:
>>>>>>>>>>>>>>
>>>>>>>>>>>>>>> On 22/03/2026 1:52 am, john larkin wrote:
>>>>>>>>>>>>>>>> On Sat, 21 Mar 2026 16:36:43 +1100, Bill Sloman <bill.sloman@ieee.org>
>>>>>>>>>>>>>>>> wrote:
>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>> On 20/03/2026 4:05 am, john larkin wrote:
>>>>>>>>>>>>>>>>>> On Tue, 17 Mar 2026 22:30:01 +0000, someone
>>>>>>>>>>>>>>>>>> <cffbf4deb9142bce48974efc0e64dede@example.com> wrote:
>>>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>>>> I assume these are up-counters, so the thing overflows at all 1's. Then you only have the one fast carry TPD for the MS18b overflowing to all 1s when a 1 is clocked into its LSB.  One whole clock period to clock the 1 out of the DFF and meet the setup times for what I assume is a synchronous LD and its setup for the counters. So that particular timing criticality is a DFF TPD and a LD setup TSU to reliably capture the register data. The LD TPD to CLK TSU for the LS18b counter shouldn't be a problem. This must be very speedy logic for 150MHz. Do you have a simulator that displays how much margin you have on this timing, or is it just a bunch waveforms?
>>>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>>> Yes, loadable up-counter with carry chain.
>>>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>>> This would be in an FPGA, so the diagram is just a concept. The
>>>>>>>>>>>>>>>>>> reality will be VHDL code. And the FPGA boys use the Wishbone
>>>>>>>>>>>>>>>>>> architecture and want the counter to be 32 bits, which is OK with me.
>>>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>>> We are already doing a DDS at 250 MHz on this chip, an Efinix T20, so
>>>>>>>>>>>>>>>>>> I expect we could do a divider in that ballpark.  The T20 is in the
>>>>>>>>>>>>>>>>>> *slow* Efinix family.
>>>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>>> I think the T20 has 18-bit fast carry chains.
>>>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>>> After the boys code this, the tools can verify timing.
>>>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>>> FPGAs are great, but there's a cultural gap between people who draw
>>>>>>>>>>>>>>>>>> and people who type.
>>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>> Can't say I've noticed that, but since I can do both, and most of the
>>>>>>>>>>>>>>>>> engineers I've hung out with could too, John Larkin may be projecting here.
>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>> Gosh Bill, you are wonderful. You are great at everything.
>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>> I'm certainly not great at writing VHDL. There's a VHDL text-book on my
>>>>>>>>>>>>>>> bookshelf - bought for a project which didn't come off - but the stuff I
>>>>>>>>>>>>>>> did type was in a much less powerful language, but powerful enough to
>>>>>>>>>>>>>>> get the chip to do what I wanted it to.
>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>> What are you designing now?
>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>> Absolutely nothing. I do fish for work from time to time, but at 83 I'm
>>>>>>>>>>>>>>> not an attractive employee.
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> Join one of those maker space things, meet some people, offer to help
>>>>>>>>>>>>>> for free, see what happens.
>>>>>>>>>>>>>
>>>>>>>>>>>>> I'm active on the committee of NSW branch of the IEEE but I don't know
>>>>>>>>>>>>> of any maker space things in Sydney.
>>>>>>>>>>>>
>>>>>>>>>>>> Hey, you could google
>>>>>>>>>>>>
>>>>>>>>>>>> maker spaces sydney australia
>>>>>>>>>>>
>>>>>>>>>>> https://makerspaces.com.au/nsw/sydney
>>>>>>>>>>>
>>>>>>>>>>> So they exist. Leather work and needle work (sewing) are supported. I am
>>>>>>>>>>> a tolerably competent carpenter so I might fit in. As a route into
>>>>>>>>>>> advanced electronic design it doesn't look promising.
>>>>>>>>>>
>>>>>>>>>> We have some of that handicraft stuff here, but we have a lot of
>>>>>>>>>> people who want to build things that use electronics, and those people
>>>>>>>>>> aren't usually very good circuit designers. [1]
>>>>>>>>>>
>>>>>>>>>> Some of them are actual startups with an idea and some funding. They
>>>>>>>>>> go to meetups to pitch their ideas and maybe meet people who could
>>>>>>>>>> help.
>>>>>>>>>>
>>>>>>>>>> Studio 45 near here has an occasional meetup with 500 people, free
>>>>>>>>>> food and beer. We might sponsor one.
>>>>>>>>>>
>>>>>>>>>> We had a cool one at a Rivian facility. I met a photonics consultant
>>>>>>>>>> and recommended Phil's book. And listened to yet another pitch for AI
>>>>>>>>>> circuit/pcb design.
>>>>>>>>>>
>>>>>>>>>> There's one coming up in a pier on the SF Bay, an ocean
>>>>>>>>>> instrumentetion outfit. That should be fun.
>>>>>>>>>>
>>>>>>>>>> I do one or two meetups per month and meet lots of Young Things.
>>>>>>>>>>
>>>>>>>>>> The point is that you could show up, and meet people who need
>>>>>>>>>> electronics, and see what happens.
>>>>>>>>>
>>>>>>>>> If I showed up at the right meetings, I might meet people who needed
>>>>>>>>> electronics. The odds don't look great.
>>>>>>>>>
>>>>>>>>>> Or don't.
>>>>>>>>>
>>>>>>>>> That would seem to be the rational choice.
>>>>>>>>
>>>>>>>> Exactly. Do nothing. Just post insults on forums all day.
>>>>>>>
>>>>>>> I post information. When it doesn't inform people that you are a
>>>>>>> brilliant circuit designer, you feel insulted, though you should be used
>>>>>>> to that by now.
>>>>>>>
>>>>>>>>>> [1] It's impressive how few people are good at electronic design.
>>>>>>>>>
>>>>>>>>> Even more impressive that you seem to think you can make that statement.
>>>>>>>>> If you don't think that a classical emitter-coupled monostable can work,
>>>>>>>>> your status as a judge of electronic design quality can't be all that high.
>>>>>>>>
>>>>>>>> That circuit can certainly work; it's classic=ancient. But not often
>>>>>>>> useful in this modern world. I didn't like your version because the
>>>>>>>> input trigger had to be delicately tuned to fire it, and it really
>>>>>>>> amplified the input pulse more than it one-shotted. It was Spice tuned
>>>>>>>> until it appeared to work.
>>>>>>>
>>>>>>> Many applications have well-defined trigger pulses. I certainly didn't
>>>>>>> spend any time "deliberately tuning" the circuit.
>>>>>>>
>>>>>>> The trigger pulse has to be big enough and fast enough to turn off the
>>>>>>> normally-on transistor and push enough current into the timing capacitor
>>>>>>> to get the one-shot action, and the amplification is incidental to that.
>>>>>>>
>>>>>>> You didn't recognise the circuit, couldn't see how it worked, and have
>>>>>>> been trying to deny this obvious point ever since.
>>>>>>
>>>>>> That circuit is in the 1964 GE Transistor Manual (7th edition, $2).
>>>>>> Transistors used to be expensive so their use was minimized. Nowadays
>>>>>> they cost about nothing.
>>>>>
>>>>> Using a broad-band transistor in a classic circuit can give you a much
>>>>> shorter pulse. Broad-band transistors aren't all that cheap, and there
>>>>> aren't as many of them around as there used to be, but it can be a
>>>>> useful option.
>>>>>
>>>>>>> This isn't the kind
>>>>>>> of comment you like reading, and will claim that you are being insulted.
>>>>>>> Tough. This isn't some kind of mutual approbation society
>>>>>>>
>>>>>>>> There are far better, easier, cheaper, more deliberate ways to make a
>>>>>>>> fast one-shot these days.
>>>>>>>
>>>>>>> Which is to say, to make a one-shot whose action you can understand
>>>>>>
>>>>>> I usually design circuits that I understand, but what matters is that
>>>>>> they work.
>>>>>>
>>>>>> With Spice or experiment, one can profitably design a circuit that you
>>>>>> don't understand. I have a cool new sensor simulator circuit that I
>>>>>> don't understand.
>>>>>
>>>>> You really don't want to. They have a nasty habit of doing something
>>>>> unexpected at inconvenient moments.
>>>>
>>>> That's not a nasty habit, it's a talent that I practice and teach.
>>>
>>> That's a really nasty habit. Encouraging people to build circuits that
>>> can go wrong when they run into a situation that the designer hadn't
>>> expected is - to put it kindly - unwise.
>> 
>> Invention is precisely running into - running toward - the unexpected.
>
>That's a bizarre way of looking at it. It's doing something in a way 
>that hasn't been done before, but it is goal directed, and you wouldn't 
>start the process if you didn't have a pretty clear idea of what you 
>wanted to do, if not exactly how you were going to do it.

I strongly disagree; that is backwards. Sometimes we imagine products
or circuits that nobody ever wanted or expected. It just happens
sometimes at 2AM.

I have a folder full of ideas, most speculative and unexpected and
probably dumb. We hire smart kids, college students, to explore them
and write up a report on the possible uses, competitors specs and
pricing, any interesting offshoots that occurr to them. They get a
fixed fee when they turn in the report.

>
>> Sometimes that's accidental, but can be deliberately provoked.
>> Inventing needs the right skills and personality but improves with
>> practice in the right environment. Books have been written about that.
>
>None of them useful enough to have been touted at places that encouraged 
>inventions and applying for patents. EMI Central Research was just such 
>a place, and I worked there for three years without ever running into 
>such a book. The histories of Bell Labs
>
>https://en.wikipedia.org/wiki/The_Idea_Factory
>

I have that one; good book.

Someone said that all the great inventions at Bell in those days were
done by people who ate lunch with Harry Nyquist.

>didn't mention any such book either. People will write books with the 
>flimsiest of justifications if they think the product will sell. 
>Teaching people how to make genuine inventions would be a very good 
>thing if you could do it, and a lot of confidence tricksters claim that 
>they can. The evidence supporting such claims doesn't seem to exist.

The real evidence is purchase orders.

>
>> Some people invent things. Some intelligent and (over)educated people
>> actively resent invention, because they can't do it.
>
>I can't say that I've met any of them. My father and two of my friends 
>have each got their names onto about 25 patents and none of them ever 
>talked about people resenting that work.
>
>>>> Given an enormous space of undiscovered ideas, one profits from a
>>>> method of exploring them in parallel with minimal filtering.
>
>At EMI Central Research we were encourage to submit patent queries. One 
>of my colleagues put in a record number of patent queries - about fifty 
>in one year - and was seen as having rather poor judgement. None of them 
>turned into a patent. He would have benefited from better filtering.

The real evidence is purchase orders.


John Larkin
Highland Tech Glen Canyon Design Center
Lunatic Fringe Electronics

Back to comp.arch.fpga | Previous | NextPrevious in thread | Next in thread | Find similar


Thread

fast divider? john larkin <jl@glen--canyon.com> - 2026-03-15 11:47 -0700
  Re: fast divider? someone <cffbf4deb9142bce48974efc0e64dede@example.com> - 2026-03-17 22:30 +0000
    Re: fast divider? john larkin <jl@glen--canyon.com> - 2026-03-19 10:05 -0700
      Re: fast divider? Bill Sloman <bill.sloman@ieee.org> - 2026-03-21 16:36 +1100
        Re: fast divider? john larkin <jl@glen--canyon.com> - 2026-03-21 07:52 -0700
          Re: fast divider? Bill Sloman <bill.sloman@ieee.org> - 2026-03-22 03:00 +1100
            Re: fast divider? john larkin <jl@glen--canyon.com> - 2026-03-27 11:39 -0700
              Re: fast divider? Bill Sloman <bill.sloman@ieee.org> - 2026-03-28 16:44 +1100
                Re: fast divider? john larkin <jl@glen--canyon.com> - 2026-03-28 14:38 -0700
                Re: fast divider? Bill Sloman <bill.sloman@ieee.org> - 2026-03-29 15:52 +1100
                Re: fast divider? john larkin <jl@glen--canyon.com> - 2026-03-29 08:18 -0700
                Re: fast divider? Bill Sloman <bill.sloman@ieee.org> - 2026-03-30 16:42 +1100
                Re: fast divider? john larkin <jl@glen--canyon.com> - 2026-03-30 08:00 -0700
                Re: fast divider? Bill Sloman <bill.sloman@ieee.org> - 2026-03-31 16:35 +1100
                Re: fast divider? john larkin <jl@glen--canyon.com> - 2026-03-31 02:40 -0700
                Re: fast divider? Bill Sloman <bill.sloman@ieee.org> - 2026-03-31 22:30 +1100
                Re: fast divider? john larkin <jl@glen--canyon.com> - 2026-03-31 08:14 -0700
                Re: fast divider? Bill Sloman <bill.sloman@ieee.org> - 2026-04-01 15:54 +1100
                Re: fast divider? john larkin <jl@glen--canyon.com> - 2026-04-01 01:06 -0700
                Re: fast divider? Bill Sloman <bill.sloman@ieee.org> - 2026-04-02 02:13 +1100
                Re: fast divider? john larkin <jl@glen--canyon.com> - 2026-04-01 09:12 -0700
                Re: fast divider? Bill Sloman <bill.sloman@ieee.org> - 2026-04-02 14:41 +1100
                Re: fast divider? john larkin <jl@glen--canyon.com> - 2026-04-02 07:53 -0700
                Re: fast divider? Bill Sloman <bill.sloman@ieee.org> - 2026-04-03 02:21 +1100
                Re: fast divider? john larkin <jl@glen--canyon.com> - 2026-04-02 08:57 -0700
                Re: fast divider? Bill Sloman <bill.sloman@ieee.org> - 2026-04-03 04:05 +1100
          Re: fast divider? someone <cffbf4deb9142bce48974efc0e64dede@example.com> - 2026-03-26 23:30 +0000
        Re: fast divider? Ross Finlayson <ross.a.finlayson@gmail.com> - 2026-03-21 09:38 -0700
          Re: fast divider? john larkin <jl@glen--canyon.com> - 2026-03-21 10:29 -0700
            Re: fast divider? Bill Sloman <bill.sloman@ieee.org> - 2026-03-23 23:49 +1100
              Re: fast divider? john larkin <jl@glen--canyon.com> - 2026-03-23 08:20 -0700
                Re: fast divider? Bill Sloman <bill.sloman@ieee.org> - 2026-03-24 23:04 +1100
                Re: fast divider? john larkin <jl@glen--canyon.com> - 2026-03-24 05:56 -0700
                Re: fast divider? Bill Sloman <bill.sloman@ieee.org> - 2026-03-25 02:47 +1100

csiph-web