KiCad Spice, Anyone Tried It?

Strawman argument. Nobody made claims about taking it away "improperly".

Where do the billions they make come from? Thin air or what?

Reply to
Tom Gardner
Loading thread data ...

I can remember a couple of cases when I started looking at the code for a server level Power system.

Several variables were zeroed - the compiler generated an xor of a register with itself, a load immediate, and subtracted a register from itself. My first thought was why can't they make up their mind? Then I realized that they were different registers so the instructions were executed in parallel in different execution units.

The Power processors have a set of powerful (if somewhat incomprehensible) bit manipulation instructions. The compiler would routinely combine multiple bit operations and tests.

My favorite was when one of the new guys came to me when he was getting wrong results on a floating point compare, even tho he was doing the same operations on the same values. (Those that know the answer don't get ahead of me.) I found out numerical analyses apparently is not a requirement (or even strongly recommended elective) for a comp sci degree anymore. Of course one path was optimized with a multiply-add instruction, but the other path was convoluted enough that the multiply and add were separate. (The multiply-add instruction has one rounding error, the separate operations have 2, which in this case lead to a difference in the low order bit.)

Reply to
Dennis

econd levels so they don't impact me. They don't trade on information so I don't impact them.

be slightly too long for the margin of this page.

quities. Do you somehow have the understanding they are improperly taking away from others? That would be the same as saying your competitors are im properly taking away from you. Is that what this is about? Are you feelin g like a victim of unfair competition?

No strawman argument because I didn't make an argument just like you didn't make an argument. I responded to your question with my question because I don't see the point of your question.

As long as it is not improper, why do I care where it comes from? Why do y ou care where it comes from if it is not improper?

What is your point? Do you have something to say about this or not?

--
  Rick C. 

  +--+ Get 1,000 miles of free Supercharging 
 Click to see the full signature
Reply to
Ricketty C

Lots of US DoD systems, largely but not exclusively radar, some quite large. Some civilian weather and ATC radars. And the like. There was always a radar somewhere.

In the old days, a common pattern was to have a Motorola SBC running a RTOS (often VxWorks, but I've used many others) connected by a communications link (bespoke in the old days, now usually ethernet) to a main computer of some kind.

The SBC had the hard(ish) realtime code that directly controlled the radar hardware, implementing commands from the main computer and reporting back.

The main computer was everything in the old days - I've used VAX/VMS, Harris Nighthawk, Concurrent, SEL 32/55, and so on. Over the years, UNIX displaced all these, but the platform was proprietary - SGI, IBM, HP, Solaris, and so on. These have all been swept away by Linux on X86 hardware.

In parallel, there would be a Signal Processor that originally was a bespoke hardwired computer that implemented a single algorithm extremely fast, turning raw I&Q data from the REceiver-eXciter into detections, which were then passed on to the main computer, where such things as tracking was implemented.

Over time, the SP function migrated into purchased large computers, and eventually became pools of enterprise-server class x86 boxes running RHEL.

One can see where this has been and is going.

I remember them too. But see above.

Neither does embedded mean small. Some of the radars I've worked on are ten-story high buildings with 27-meter diameter antenna patches on three sides. There were offices behind those antenna faces.

It's not a design error, it's a well known property of such calculations, especially if one is forced to approximate to yield an answer fast enough to matter. Your final exam is approaching at MACH

  1. Another classic example is weather radars - how long the various algorithms take depends on just how complex the radar picture is, and there is no limit to this. So, one must design load shedding et al in from the beginning. What you cannot have is a weather radar that falls over in heavy weather, just when it gets interesting.

A predecessor to Sea Sparrow was Tartar, a shipboard anti-aircraft missile system . This was a hard-core realtime system that cycled at someting like 64 Hz, with the application code running directly on the metal - no operating system whatsoever. And this is where I first encountered the concept of a "rubber clock". I bet that engagability factored into the need.

Yep. That's the hope. If the ship survived.

Joe Gwinn

Reply to
Joe Gwinn

Well, you've chosen to snip the context, so you have chosen to obfuscate points!

Your response (which you have partly left in above) "There's nothing wrong with computer trading. I don't trade on millisecond levels so they don't impact me." was in response to Jeroen's statement which you snipped: "There was a time that speculation was frowned upon. These days it's the rule. Don't pretend that manipulating stocks at ms intervals can in any way be advantageous for a productive economy. They are parasites, leeching the productive work of others."

Hence the question of where the HFT mob's money comes from. If you can persuade people of the source, perhaps people would change their view as stated by Jeroen.

Reply to
Tom Gardner

isecond levels so they don't impact me. They don't trade on information so I don't impact them.

d be slightly too long for the margin of this page.

equities. Do you somehow have the understanding they are improperly takin g away from others? That would be the same as saying your competitors are improperly taking away from you. Is that what this is about? Are you feel ing like a victim of unfair competition?

".

dn't make an argument. I responded to your question with my question becau se I don't see the point of your question.

do you care where it comes from if it is not improper?

points!

Someone has to trim posts at some point. Calling people names is not very useful. Accusing people of stock manipulation without evidence is not usef ul.

So even with the inclusion of the text you quoted, nothing more has been sa id.

Stock trading happens. People who are more connected with the market have a better opportunity of making money. Using computers tied directly to the exchange provides a way to make money out of small movements in stock pric es. There is nothing inherently evil about this.

You continue to not add anything useful to the conversation. If you don't say anything relevant, I won't continue to reply.

--
  Rick C. 

  +-+- Get 1,000 miles of free Supercharging 
 Click to see the full signature
Reply to
Ricketty C

Trim: yes.

Make accusations and snip the context that disproves the accusation: no.

Avoid the question and make irrelevant points: no.

So, where /do/ you think the HFT mob's billions come from?

Reply to
Tom Gardner

s faster

w and prosper. If there were no market there would be many fewer companies and a much smaller economy.

ly crime is success?

Perhaps because all the value created by being able to trade a few millisec onds faster ends up in the pockets of those who have created that advantage .

The people who invest in companies which grow and prosper are investing for the long term. The people who make money out of rapid trading aren't.

There's an argument for applying a transaction charge to all stock market t rades.

It wouldn't need to be big - 0.1% of the value of transaction would probabl y do the trick - but it would make churning unprofitable.

--
Bill Sloman, Sydney
Reply to
Bill Sloman

e

John Larkin takes the attitude that anything he can't comprehend isn't wort h comprehending. It isn't a productive attitude.

Back when I did Theory of Computation Part 1 (as a graduate student in 1966 ) we did get taught an assembly language (for an IBM 7040/44) but only so t hat we could understand what higher-level language statements compiled to, so we could organise our calculations in a way that let the processor deal with them as quickly as possible.

I think the Cooley-Tukey Fast Fourier Transform

formatting link

came up as an example of such an organisation. Carl Friedrich Gauss first invented it in 1805. In 1966 we got told about the Lanczos version (invente d in 1942) which stuck in my mind because I'd been reading his book on nume rical methods (Applied Analysis, 1957)in the physics library and ended up c iting it in my Ph.D. thesis.

Nowhere near sorry enough.

--
Bill Sloman, Sydney
Reply to
Bill Sloman

Oh, yeah, that's what they say. But, they're also front-running, i.e. buying the lowest-bid offering just ahead of a legitimate buyer, and offering that buyer a slightly higher price on the goods they just snatched off the shelf. If you made the market equal for everybody, the front-runners wouldn't get lower prices than anybody else.

Holding a stock for microseconds sure makes them part of 'the market', all right; the parasitic part.

Reply to
whit3rd

I guess I have to interpret that a warning about what follows.

You make a good argument for the existiance of the stock exchange, no-one was disputing that. Who are the high frequency traders helping?

You tell me, you're the expert in straw-man arguments.

--
  Jasen.
Reply to
Jasen Betts

Okay. This would seem to be a very specific field - one where you are using embedded systems that need a significant amount of processing power. It's the kind of thing where you want high processing speed using whatever technology is available and best (be it multiple cpus, FPGAs, DSPs, or whatever) - mostly regardless of size, power, cost and complexity. That makes it an outlier in the embedded world.

If you say RHEL is used on these systems, I'll accept your word for that

- you are the expert there. But if you were to look at these as a proportion of embedded systems that are made, they would not register on a parts-per-million scale.

(Smaller radar systems are becoming a lot more common, especially in the automotive industry. These systems don't run Linux on the radar detection part - though the "brains" of the driver assistant system probably does. But it won't be RHEL.)

Sounds reasonable.

What you are talking about here, assuming I understand you correctly, is the machine that is tasked with handling the data, doing calculations, and presenting it - not the part doing the realtime work (that's the SBC controlling the radar hardware).

By this stage, you are not really talking about embedded systems. (The SBC running the radar is an embedded system.) You are talking about a server system - a compute farm. I don't think there is any clear definition of what an "embedded system" is, but I'm reasonably sure that a definition would not include banks of off-the-shelf x86 systems running an off-the-shelf server OS.

I think we simply have a different meaning of the term "embedded". As I said earlier, I don't think there is a clear definition. But IMHO, these computers are not embedded systems.

Then to me, it sounds like you are doing the wrong calculations or doing them in the wrong way. Maybe you should be using a different coordinate set. Maybe you should be using quaternions instead of Euler angles. I don't know the task in hand, and maybe the way you describe is the most practical choice. But here you have a real world, physical system - and if your model of that physical system involves invalid calculations and undefined results, the model is wrong.

Sure.

Reply to
David Brown

Stored charges are energy, G.R. tells us that energy has mass, therfore it's hardware?

--
  Jasen.
Reply to
Jasen Betts

If people really work at it, they can paralyze themselves and get nothing done.

--
John Larkin         Highland Technology, Inc 

Science teaches us to doubt. 
 Click to see the full signature
Reply to
jlarkin

If people really work at it, they can refuse to accept solid information that doesn't fit in with their preconceptions.

That can lead to loss of productivity.

Reply to
Tom Gardner

How many hardware/software products have you designed in the last year? I've averaged about one a month.

Confusion about what is hardware and what is software hasn't slowed me down much. Well, two of the gadgets, the tiny delay generator and the GHz o/e converter, don't have any software. Sometimes a trimpot is just the nonvolatile storage that you need.

One lockdown project:

formatting link

formatting link

It is fun to design something now and then that has no software, no FPGA.

Software seems almost entirely artificial and arbitrary to me. It has no basis in physical reality, or science, or math. It's usually very far, out of sight, from the instruction sets and computer hardware that support it. Which is why it's faddish, convoluted, and buggy.

Necessary evil these days.

--
John Larkin         Highland Technology, Inc 

Science teaches us to doubt. 
 Click to see the full signature
Reply to
jlarkin

Well, bully for you.

But such questions are, as everybody is well aware, merely flack intended to divert attention from the point in hand.

We've noticed - but it is good to see you acknowledge your limitation.

Yes. Noting the similarities is /a/ good way to realise where alternate partitioning strategies could be employed.

Reply to
Tom Gardner

Once again I had to do your trimming for you.

Just like I can't rewrite history, I can't remove the context of usenet. You can always read the prior posts. Stop being silly.

You continue to ignore my question of how this is a relevant question. I will continue to say I don't care about the answer.

You are the one trying to make irrelevant points, or more accurately ask irrelevant questions. So clearly you have nothing relevant to say and I won't respond to this issue any further.

--
  Rick C. 

  +-++ Get 1,000 miles of free Supercharging 
 Click to see the full signature
Reply to
Ricketty C

When in doubt, put it in the FPGA. I'm often amazed by the things my FPGA kids are willing to do. Essentially map procedural code into state machines that execute at 80 MHz. Our smaller ARMs execute one instruction in that time.

--
John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  
 Click to see the full signature
Reply to
John Larkin

The High Frequency Trading mob have more-or-less unlimited money at their disposal. They have been known to put trading

*business rules* in FPGA logic, things like "if price gone down 1% then sell at 1% below the last traded price".

As I said, there's a very grey area between hardware and software. Frequently a function can be realised in either, as convenient.

Reply to
Tom Gardner

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.