the hot new programming language

[rickman]

On 7/3/2015 8:38 AM, Martin Brown wrote:

On 02/07/2015 23:07, rickman wrote:
On 7/2/2015 6:03 PM, John Larkin wrote:

We can always buy faster CPUs.

That is literally the stupidest thing I've ever seen come from you.


I don't often defend Johns comments but on this he does have a point.
For most home and office kit these days the CPU power available is so
huge that efficiency literally does not matter in end user code.

The notable exceptions are video editing and 3D gaming which really do
push the performance envelope. Word processing and general stuff you can
trade a few percent of speed for safety without any problems.

Fast CPUs are cheap and getting cheaper in line with Moore's law whereas
good software engineers are rare, expensive and getting more so.

TBH I am amazed that Moore's law has held good for as long as it has.

Moore's law has not been holding as of late when you consider the speed.
Over the last 10 years or so they have been focusing on eliminating
the massive heat sinks required to continue on the earlier path (there
have been CPUs dissipating as much heat as a 100 watt lightbulb) with
the result that CPU clock speed has not improved at all and CPU
performance only trickles ahead each year. So process improvements are
no longer a salvation.

The real issue is that it is silly to think there is no cost to using
ever increasing amounts of CPU performance. John's own project used a
CPU/FPGA which was designed in, only to find he unexpectedly required a
bloody CPU fan over the chip. Yeah, he still has room to replace that
with an x86 and a GPU or even a bank of GPUs. But at some point
Samaritan will decide the computer is a threat and have Decima kill
John. So it is a slippery slope.

--

Rick

 

[DecadentLinuxUserNumeroUn]

On Fri, 03 Jul 2015 12:23:56 -0400, Joe Gwinn <joegwinn@comcast.net>
Gave us:

So, I wrote Fifth.

Joe Gwinn

That would make *you* "The Fifth Element". (or Fifth Elemental)

 

[John Larkin]

On Fri, 03 Jul 2015 12:21:53 -0400, DecadentLinuxUserNumeroUno
<DLU1@DecadentLinuxUser.org> wrote:

On Fri, 03 Jul 2015 09:17:48 -0700, John Larkin
jlarkin@highlandtechnology.com> Gave us:

I did that kilovolt switcher design recently, and it was running 4
minutes per pass, until I removed the leakage inductance and got it
down below 1 minute. Still, I got the design done in a few hours.
Since I'll breadboard it anyhow, I didn't really need Spice.

Did you post the fixed final design?

I've got several versions, fiddled with now and then in my spare time.
Can't recall if I posted the "latest" one. It will certainly change
some when it gets to be a real-world PCB schematic. The customer(s)
already want more voltage at the Widget, so I need maybe 1400 from the
supply, no big deal, just tweak the feedback.

This is about it for now, until I get a purchase order. The sim helped
estimate parts cost and board size and such.


Version 4
SHEET 1 2184 704
WIRE 1248 -608 1184 -608
WIRE 1376 -608 1312 -608
WIRE 1440 -608 1376 -608
WIRE 1568 -608 1504 -608
WIRE 1376 -576 1376 -608
WIRE 1024 -432 944 -432
WIRE 1184 -432 1184 -608
WIRE 1184 -432 1088 -432
WIRE 1248 -432 1184 -432
WIRE 1344 -432 1312 -432
WIRE 1376 -432 1376 -512
WIRE 1376 -432 1344 -432
WIRE 1440 -432 1376 -432
WIRE 1568 -432 1568 -608
WIRE 1568 -432 1504 -432
WIRE 1184 -400 1184 -432
WIRE 1376 -400 1376 -432
WIRE 784 -272 720 -272
WIRE 944 -272 944 -432
WIRE 944 -272 848 -272
WIRE 1040 -272 944 -272
WIRE 1184 -272 1184 -336
WIRE 1184 -272 1104 -272
WIRE 1248 -272 1184 -272
WIRE 1376 -272 1376 -336
WIRE 1376 -272 1248 -272
WIRE 1568 -272 1568 -432
WIRE 1568 -272 1520 -272
WIRE 1184 -240 1184 -272
WIRE 1568 -240 1568 -272
WIRE 944 -224 944 -272
WIRE 0 -208 -96 -208
WIRE 144 -208 64 -208
WIRE -96 -112 -96 -208
WIRE -16 -112 -96 -112
WIRE 144 -112 144 -208
WIRE 144 -112 64 -112
WIRE 240 -112 144 -112
WIRE 288 -112 240 -112
WIRE 496 -112 288 -112
WIRE 720 -112 720 -272
WIRE 720 -112 576 -112
WIRE 784 -112 720 -112
WIRE 944 -112 944 -160
WIRE 944 -112 848 -112
WIRE 1088 -112 944 -112
WIRE 1184 -112 1184 -176
WIRE 1184 -112 1088 -112
WIRE 1264 -112 1184 -112
WIRE 1312 -112 1264 -112
WIRE 1568 -112 1568 -160
WIRE 288 -64 288 -112
WIRE 496 -64 496 -112
WIRE 576 -64 576 -112
WIRE -96 -48 -96 -112
WIRE 144 -48 144 -112
WIRE 1088 -32 1088 -112
WIRE 1184 -32 1184 -112
WIRE -96 64 -96 32
WIRE -16 64 -96 64
WIRE 144 64 144 32
WIRE 144 64 48 64
WIRE 288 64 288 0
WIRE 336 64 288 64
WIRE 496 64 496 16
WIRE 496 64 416 64
WIRE 576 64 576 16
WIRE 576 64 496 64
WIRE 624 64 576 64
WIRE 656 64 624 64
WIRE 1088 96 1088 32
WIRE 1184 96 1184 48
WIRE -96 112 -96 64
WIRE 576 112 576 64
WIRE 944 112 944 -112
WIRE 144 144 144 64
WIRE 352 192 272 192
WIRE 480 192 432 192
WIRE 528 192 480 192
WIRE 944 240 944 192
WIRE 944 240 720 240
WIRE 16 256 -96 256
WIRE 352 256 272 256
WIRE 480 256 432 256
WIRE 576 256 576 208
WIRE 576 256 480 256
WIRE 576 288 576 256
WIRE -96 304 -96 256
WIRE 944 304 944 240
WIRE 304 320 272 320
WIRE 576 400 576 368
WIRE -96 416 -96 368
WIRE -16 416 -96 416
WIRE 144 416 144 368
WIRE 144 416 64 416
WIRE 944 432 944 384
WIRE 304 448 304 320
WIRE 480 448 304 448
WIRE 720 448 720 240
WIRE 720 448 480 448
WIRE 144 464 144 416
FLAG 144 464 0
FLAG 576 400 0
FLAG 1088 96 0
FLAG 1184 96 0
FLAG 1264 -112 HV1
FLAG 944 432 0
FLAG 240 -112 IN
FLAG -96 112 0
FLAG 1568 -112 0
FLAG 1248 -272 HV2
FLAG 1344 -432 HV3
FLAG 1520 -272 HV4
FLAG 624 64 QD
FLAG 480 192 QG
FLAG 480 256 QS
FLAG 480 448 FB
SYMBOL PowerProducts\\LTC3803 144 256 R0
SYMATTR InstName U1
SYMBOL ind2 480 -80 R0
WINDOW 0 -17 36 Right 2
WINDOW 3 -8 71 Right 2
SYMATTR InstName L1
SYMATTR Value 100ľ
SYMATTR Type ind
SYMATTR SpiceLine Ipk=3.0 Rser=0.01 Cpar=1pF
SYMBOL ind2 592 32 R180
WINDOW 0 -22 77 Right 2
WINDOW 3 -12 42 Right 2
SYMATTR InstName L2
SYMATTR Value 100ľ
SYMATTR Type ind
SYMATTR SpiceLine Ipk=20 Rser=1m Cpar=2pF
SYMBOL nmos 528 112 R0
WINDOW 0 114 48 Left 2
WINDOW 3 75 84 Left 2
SYMATTR InstName Q1
SYMATTR Value Si9420DY
SYMBOL res 560 272 R0
WINDOW 0 66 54 Left 2
WINDOW 3 48 88 Left 2
SYMATTR InstName R1
SYMATTR Value 100m
SYMBOL res 128 -64 R0
WINDOW 0 -50 33 Left 2
WINDOW 3 -52 64 Left 2
SYMATTR InstName R2
SYMATTR Value 1K
SYMBOL res 928 208 M180
WINDOW 0 -97 78 Left 2
WINDOW 3 -114 42 Left 2
SYMATTR InstName R3
SYMATTR Value 7Meg
SYMBOL res 928 400 M180
WINDOW 0 -74 64 Left 2
WINDOW 3 -80 31 Left 2
SYMATTR InstName R4
SYMATTR Value 20K
SYMBOL cap 1072 -32 R0
WINDOW 0 -61 25 Left 2
WINDOW 3 -64 58 Left 2
SYMATTR InstName C1
SYMATTR Value 66n
SYMATTR SpiceLine Rser=10m
SYMBOL res 80 400 R90
WINDOW 0 7 110 VBottom 2
WINDOW 3 29 111 VTop 2
SYMATTR InstName R5
SYMATTR Value 20K
SYMBOL cap -80 368 R180
WINDOW 0 -55 64 Left 2
WINDOW 3 -62 31 Left 2
SYMATTR InstName C2
SYMATTR Value 10n
SYMBOL voltage -96 -64 R0
WINDOW 123 -220 84 Left 2
WINDOW 0 50 34 Left 2
WINDOW 3 50 66 Left 2
WINDOW 39 0 0 Left 2
SYMATTR InstName V1
SYMATTR Value 24
SYMBOL res 1168 -48 R0
WINDOW 0 63 38 Left 2
WINDOW 3 64 71 Left 2
SYMATTR InstName R250
SYMATTR Value 250K
SYMATTR SpiceLine pwr=12
SYMBOL res -32 -96 R270
WINDOW 0 43 61 VTop 2
WINDOW 3 -9 58 VBottom 2
SYMATTR InstName Rmon
SYMATTR Value 1m
SYMBOL cap 64 -224 R90
WINDOW 0 -47 29 VBottom 2
WINDOW 3 -40 30 VTop 2
SYMATTR InstName C4
SYMATTR Value 50m
SYMBOL diode 784 -96 R270
WINDOW 0 -31 30 VTop 2
WINDOW 3 -32 31 VBottom 2
SYMATTR InstName D2
SYMBOL diode 960 -160 R180
WINDOW 0 56 41 Left 2
WINDOW 3 64 9 Left 2
SYMATTR InstName D1
SYMBOL cap 784 -256 R270
WINDOW 0 111 33 VTop 2
WINDOW 3 104 34 VBottom 2
SYMATTR InstName C6
SYMATTR Value 22n
SYMBOL cap 1200 -176 R180
WINDOW 0 -59 45 Left 2
WINDOW 3 -65 15 Left 2
SYMATTR InstName C7
SYMATTR Value 22n
SYMBOL diode 1040 -256 R270
WINDOW 0 72 33 VTop 2
WINDOW 3 67 33 VBottom 2
SYMATTR InstName D3
SYMBOL res 1552 -256 R0
WINDOW 0 -119 60 Left 2
WINDOW 3 -116 97 Left 2
SYMATTR InstName R1200
SYMATTR Value 600K
SYMBOL diode 1200 -336 R180
WINDOW 0 58 38 Left 2
WINDOW 3 63 6 Left 2
SYMATTR InstName D4
SYMBOL cap 1392 -336 R180
WINDOW 0 44 0 Left 2
WINDOW 3 36 -31 Left 2
SYMATTR InstName C5
SYMATTR Value 22n
SYMBOL diode 1248 -416 R270
WINDOW 0 -34 31 VTop 2
WINDOW 3 -37 31 VBottom 2
SYMATTR InstName D5
SYMBOL cap 1024 -416 R270
WINDOW 0 77 32 VTop 2
WINDOW 3 71 33 VBottom 2
SYMATTR InstName C8
SYMATTR Value 22n
SYMBOL diode 1392 -512 R180
WINDOW 0 -37 -6 Left 2
WINDOW 3 -27 -35 Left 2
SYMATTR InstName D6
SYMBOL cap 1440 -416 R270
WINDOW 0 -39 32 VTop 2
WINDOW 3 -49 30 VBottom 2
SYMATTR InstName C9
SYMATTR Value 22n
SYMBOL diode 1440 -592 R270
WINDOW 0 -31 30 VTop 2
WINDOW 3 -32 31 VBottom 2
SYMATTR InstName D7
SYMBOL cap 1248 -592 R270
WINDOW 0 -38 33 VTop 2
WINDOW 3 -43 34 VBottom 2
SYMATTR InstName C10
SYMATTR Value 22n
SYMBOL res 448 240 R90
WINDOW 0 68 52 VBottom 2
WINDOW 3 75 53 VTop 2
SYMATTR InstName R7
SYMATTR Value 5K
SYMBOL res 448 176 R90
WINDOW 0 -44 52 VBottom 2
WINDOW 3 -37 53 VTop 2
SYMATTR InstName R8
SYMATTR Value 33
SYMBOL res 432 48 R90
WINDOW 0 -40 54 VBottom 2
WINDOW 3 -33 55 VTop 2
SYMATTR InstName R6
SYMATTR Value 1K
SYMBOL cap 272 -64 R0
WINDOW 0 -44 24 Left 2
WINDOW 3 -47 58 Left 2
SYMATTR InstName C3
SYMATTR Value 47p
SYMBOL cap 48 48 R90
WINDOW 0 7 -4 VBottom 2
WINDOW 3 24 -2 VTop 2
SYMATTR InstName C11
SYMATTR Value 2ľ
TEXT 536 -160 Bottom 2 !K1 L1 L2 1
TEXT 216 -408 Left 2 !.tran 12ms startup
TEXT 472 -224 Left 2 ;DRQ127 1:1
TEXT 736 272 Left 2 ;0.8V
TEXT 992 -224 Left 2 ;MMBD5004
TEXT 112 -472 Left 2 ;HV SUPPLY J LARKIN June 30, 2015
TEXT 792 -312 Left 2 ;500V









--

John Larkin Highland Technology, Inc
picosecond timing laser drivers and controllers

jlarkin att highlandtechnology dott com
http://www.highlandtechnology.com

 

[DecadentLinuxUserNumeroUn]

On Fri, 03 Jul 2015 10:23:50 -0700, John Larkin
<jlarkin@highlandtechnology.com> Gave us:

so I need maybe 1400 from the
supply, no big deal, just tweak the feedback.

Or add one more multiplier stage, since you are well below any power
requisite. It would probably run even better.

 

[John Larkin]

On Fri, 03 Jul 2015 13:29:57 -0400, DecadentLinuxUserNumeroUno
<DLU1@DecadentLinuxUser.org> wrote:

On Fri, 03 Jul 2015 10:23:50 -0700, John Larkin
jlarkin@highlandtechnology.com> Gave us:

so I need maybe 1400 from the
supply, no big deal, just tweak the feedback.

Or add one more multiplier stage, since you are well below any power
requisite. It would probably run even better.

Either way, but fewer parts is better; it's looking lopsided already.
Adding another mult stage reduces my HV1 output, which is more than I
need anyhow.

Looks like 2 mA at 1400 volts will be plenty, so that cheap stock
Coiltronics transformer is still OK.

The real advantage, to me, of using Spice here, is evaluating the
magnetics. Magnetics tend to be a pain. Spice computes the peak and
RMS coil currents, something I prefer not to do any other way.

I could buy a potted brick proportional (unregulated) HV supply for a
bit over $100, but it wouldn't make HV1 and I might need to close a
feedback loop around it to nail the output voltage. And it wouldn't be
as much fun.


--

John Larkin Highland Technology, Inc
picosecond timing laser drivers and controllers

jlarkin att highlandtechnology dott com
http://www.highlandtechnology.com

 

[DecadentLinuxUserNumeroUn]

On Fri, 03 Jul 2015 10:45:15 -0700, John Larkin
<jlarkin@highlandtechnology.com> Gave us:

The real advantage, to me, of using Spice here, is evaluating the
magnetics. Magnetics tend to be a pain. Spice computes the peak and
RMS coil currents, something I prefer not to do any other way.

Their problem is that there are few, if not no models of xformers with
a single turn feedback winding in their libraries, and adding one
doesn't work, because it needs to be magnetically coupled. So even
their modeling structure needs work.

That is one of the main reasons I do not model my supplies with it, as
I *do* use a feedback winding.

 

[DecadentLinuxUserNumeroUn]

On Fri, 3 Jul 2015 11:44:55 -0700 (PDT), Lasse Langwadt Christensen
<langwadt@fonz.dk> Gave us:

Den fredag den 3. juli 2015 kl. 20.35.39 UTC+2 skrev DecadentLinuxUserNumeroUno:
On Fri, 03 Jul 2015 10:45:15 -0700, John Larkin
jlarkin@highlandtechnology.com> Gave us:


The real advantage, to me, of using Spice here, is evaluating the
magnetics. Magnetics tend to be a pain. Spice computes the peak and
RMS coil currents, something I prefer not to do any other way.

Their problem is that there are few, if not no models of xformers with
a single turn feedback winding in their libraries, and adding one
doesn't work, because it needs to be magnetically coupled. So even
their modeling structure needs work.

That is one of the main reasons I do not model my supplies with it, as
I *do* use a feedback winding.

you just add the inductors and tell it what the coupling between is

http://cds.linear.com/docs/en/lt-journal/LTMag-V16N3-23-LTspice_Transformers-MikeEngelhardt.pdf

-Lasse

Thanks. It was years ago I gave up on them. Perhaps now, they have
facilitated said couplings (obviously so). I give it a gander.

 

[John Larkin]

On Fri, 03 Jul 2015 14:35:30 -0400, DecadentLinuxUserNumeroUno
<DLU1@DecadentLinuxUser.org> wrote:

On Fri, 03 Jul 2015 10:45:15 -0700, John Larkin
jlarkin@highlandtechnology.com> Gave us:


The real advantage, to me, of using Spice here, is evaluating the
magnetics. Magnetics tend to be a pain. Spice computes the peak and
RMS coil currents, something I prefer not to do any other way.

Their problem is that there are few, if not no models of xformers with
a single turn feedback winding in their libraries, and adding one
doesn't work, because it needs to be magnetically coupled. So even
their modeling structure needs work.

My HV supply model has a transformer made from two coupled inductors;
I didn't use a transformer model. Just plop down a couple of inductors
L1 and L2 and add a Spice directive anywhere on the screen

K L1 L2 N

where N is the coupling coefficient. N=1 is perfect coupling. K<>1
does slow down the simulation.

Just make the L1/L2 inductance ratio the square of the turns ratio.

You can name several inductors in the K statement, for more windings.

You can add coil ohms into each inductor, or just add resistors
outside of ideal inductors. Similarly, you can tweak N down to fake
leakage inductance, or add separate visible inductors.

I try to use stock surface-mount parts whenever I can. Those
DRQ-series dual inductors can be very handy. As can 4-winding ISDN
transformers.


--

John Larkin Highland Technology, Inc
picosecond timing laser drivers and controllers

jlarkin att highlandtechnology dott com
http://www.highlandtechnology.com

 

[krw]

On Fri, 03 Jul 2015 09:04:03 -0700, John Larkin
<jlarkin@highlandtechnology.com> wrote:

On Fri, 3 Jul 2015 07:29:00 -0700 (PDT), Lasse Langwadt Christensen
langwadt@fonz.dk> wrote:

Den fredag den 3. juli 2015 kl. 12.52.52 UTC+2 skrev Martin Brown:
If you want to sell your soul for maximum financial gain then
destabilising the global stock trading systems with sophisticated high
frequency trading algorithms is definitely the way to go.

One guy in the UK in his parents bedroom can allegedly do this:

http://www.bbc.co.uk/news/business-32415664

People seem to get upset if you are too good at it!


yeh, you have to be a member of "the money sucking parasite club" to
manipulate prices and steal money like that


-Lasse

A small tax on transactions, like 0.1%, would have a remarkable
damping effect. Maybe we can get that soon, after the next monster
worldwide crash.

Nah. They'll just bail out the culprets. As usual.

The problem with a small tax is that it won't stay small. How about a
1sec delay in reporting sale prices? Perhaps even dithering the
delay.

 

[John Larkin]

On Fri, 03 Jul 2015 20:00:08 -0400, krw <krw@nowhere.com> wrote:

On Fri, 03 Jul 2015 09:04:03 -0700, John Larkin
jlarkin@highlandtechnology.com> wrote:

On Fri, 3 Jul 2015 07:29:00 -0700 (PDT), Lasse Langwadt Christensen
langwadt@fonz.dk> wrote:

Den fredag den 3. juli 2015 kl. 12.52.52 UTC+2 skrev Martin Brown:
If you want to sell your soul for maximum financial gain then
destabilising the global stock trading systems with sophisticated high
frequency trading algorithms is definitely the way to go.

One guy in the UK in his parents bedroom can allegedly do this:

http://www.bbc.co.uk/news/business-32415664

People seem to get upset if you are too good at it!


yeh, you have to be a member of "the money sucking parasite club" to
manipulate prices and steal money like that


-Lasse

A small tax on transactions, like 0.1%, would have a remarkable
damping effect. Maybe we can get that soon, after the next monster
worldwide crash.

Nah. They'll just bail out the culprets. As usual.

The problem with a small tax is that it won't stay small. How about a
1sec delay in reporting sale prices? Perhaps even dithering the
delay.

The tax could be inverse on holding time, tapering to zero after, say,
5 years. That would change a lot.


--

John Larkin Highland Technology, Inc
picosecond timing laser drivers and controllers

jlarkin att highlandtechnology dott com
http://www.highlandtechnology.com

 

[DecadentLinuxUserNumeroUn]

On Fri, 03 Jul 2015 20:00:08 -0400, krw <krw@nowhere.com> Gave us:

On Fri, 03 Jul 2015 09:04:03 -0700, John Larkin
jlarkin@highlandtechnology.com> wrote:

snip (learn how)

A small tax on transactions, like 0.1%, would have a remarkable
damping effect. Maybe we can get that soon, after the next monster
worldwide crash.

Nah. They'll just bail out the culprets. As usual.

The problem with a small tax is that it won't stay small. How about a
1sec delay in reporting sale prices? Perhaps even dithering the
delay.

The term is 'culprits'. Maybe you should delay sends and proofread
for a few seconds. Oh... that's right... you thought that was the
correct spelling.

 

[Robert Baer]

John Larkin wrote:

http://www.itworld.com/article/2694378/college-students-learning-cobol-make-more-money.html

The revival of Basic is next.



....and for dynamic graphing, use Visual Cobol.

 

[Martin Brown]

On 03/07/2015 17:33, rickman wrote:

On 7/3/2015 8:38 AM, Martin Brown wrote:
On 02/07/2015 23:07, rickman wrote:
On 7/2/2015 6:03 PM, John Larkin wrote:

We can always buy faster CPUs.

That is literally the stupidest thing I've ever seen come from you.

I don't often defend Johns comments but on this he does have a point.
For most home and office kit these days the CPU power available is so
huge that efficiency literally does not matter in end user code.

The notable exceptions are video editing and 3D gaming which really do
push the performance envelope. Word processing and general stuff you can
trade a few percent of speed for safety without any problems.

Fast CPUs are cheap and getting cheaper in line with Moore's law whereas
good software engineers are rare, expensive and getting more so.

TBH I am amazed that Moore's law has held good for as long as it has.

Moore's law has not been holding as of late when you consider the speed.

Moore's law never made any claims about speed. It was specifically about
the number density of transistors on a given area of silicon.

For a long while it did translate to speed gain as features got smaller
with lower capacitance, higher frequencies and lower working voltages.

Over the last 10 years or so they have been focusing on eliminating
the massive heat sinks required to continue on the earlier path (there
have been CPUs dissipating as much heat as a 100 watt lightbulb) with
the result that CPU clock speed has not improved at all and CPU
performance only trickles ahead each year. So process improvements are
no longer a salvation.

The time to triple the performance of bleeding edge CPUs has gone from
about 2-3 years in the old days to about 5-6 years now. The absolute
performance of most desktop PCs is now well beyond what any normal user
will ever need apart from for video editing, gaming and running spice.

In some gaming PCs the video card has vastly more compute power than the
main CPU and for the right problems there are libraries to exploit it
directly like NVidias CUDA.

http://www.nvidia.com/object/what-is-gpu-computing.html

The real issue is that it is silly to think there is no cost to using
ever increasing amounts of CPU performance. John's own project used a

Although you could interpret what he said to mean that. I took it to
mean the much more sensible position that modern CPUs are already so
fast that you can afford to lose some performance to gain *reliability*.

CPU/FPGA which was designed in, only to find he unexpectedly required a
bloody CPU fan over the chip. Yeah, he still has room to replace that
with an x86 and a GPU or even a bank of GPUs. But at some point
Samaritan will decide the computer is a threat and have Decima kill
John. So it is a slippery slope.

The singularity is coming - beware. We are already letting computers
design new bigger chips that no individual human can fully comprehend.

Deterministic number crunching computers are safe enough but when they
have neural net hardware or a software simulation of it on a sufficient
scale to exceed the human brain connectivity all bets are off.

--
Regards,
Martin Brown

 

[Jasen Betts]

On 2015-07-03, Phil Hobbs <pcdhSpamMeSenseless@electrooptical.net> wrote:

I agree, but that's a separate issue. Outlawing C strings, for
instance, outlaws good code as well as bad. Those of us who know that,
for instance, strncpy() doesn't append a null if it runs out of space,
know enough to unconditionally put the null in there. (Yes, it's a
stupid design, but alternatives are available. The auto industry's
standards are actually pretty useful.)

it's using the wrong function strncopy is for writting to null padded
records, where maximal strings are unterminated.

if you want a length limit and a nul at the end use sprintf

sprintf(dest,"%.*s",len-1,src);

--
umop apisdn

 

[DecadentLinuxUserNumeroUn]

On Sat, 04 Jul 2015 09:19:35 +0100, Martin Brown
<|||newspam|||@nezumi.demon.co.uk> Gave us:

The singularity is coming - beware. We are already letting computers
design new bigger chips that no individual human can fully comprehend.

Slow light technology will usher in a simple 4 bit optical computer
that puts them all to shame.

 

[Tom Del Rosso]

Martin Brown wrote:

TBH I am amazed that Moore's law has held good for as long as it has.

Moore's law will last forever if people keep re-defining it. The number of
transistors doubled every year, like Moore said it would, from 1959 to about
1982. Then we hit the first wall. Carver Mead re-wrote the book on VLSI
and it continued to double every 1.5 years until about 2000. Since then it
has doubled every 2 years. When it doubles every 20 years people will still
call it Moore's law. It is continued progress, but it isn't Moore's law.

 

[Tom Del Rosso]

John Larkin wrote:

What I want is a 1000x speed improvement, so I can move sliders and
see waveforms change instantly, just like a breadboard with pots and a
scope. N-dimensional iteration at 5 minutes per trial is not
intuitive, but then 1 minute isn't either.

Build an analog computer.

 

[E. Kappos]

"John Larkin" <jlarkin@highlandtechnology.com> wrote in message
news:egcdpal3j5j8en0ead82236nm8g28pim5b@4ax.com...

On Fri, 3 Jul 2015 07:29:00 -0700 (PDT), Lasse Langwadt Christensen
langwadt@fonz.dk> wrote:

Den fredag den 3. juli 2015 kl. 12.52.52 UTC+2 skrev Martin Brown:
If you want to sell your soul for maximum financial gain then
destabilising the global stock trading systems with sophisticated high
frequency trading algorithms is definitely the way to go.

One guy in the UK in his parents bedroom can allegedly do this:

http://www.bbc.co.uk/news/business-32415664

People seem to get upset if you are too good at it!


yeh, you have to be a member of "the money sucking parasite club" to
manipulate prices and steal money like that


-Lasse

A small tax on transactions, like 0.1%, would have a remarkable
damping effect. Maybe we can get that soon, after the next monster
worldwide crash.

but most likely would destabilize the ...government!

 

[Tom Del Rosso]

Martin Brown wrote:

Moore's law never made any claims about speed. It was specifically
about the number density of transistors on a given area of silicon.

I roll my eyes when I hear "Moore's law" and the "computing power of a chip"
in the same sentence. He stated his law when a chip had 4 transistors.
Since you can't make a computer with 4, it makes no sense to speak of the
computing power of a chip.

What we need is a breakthrough in 3D structures. In 2D we're limited to a
few connections per transistor, and a few per gate. It's
connections-per-element that will make HAL possible.

 

[Lasse Langwadt Christense]

Den lřrdag den 4. juli 2015 kl. 15.56.00 UTC+2 skrev krw:

On Sat, 4 Jul 2015 16:25:10 +0300, "E. Kappos" <mike@net.gr> wrote:


"John Larkin" <jlarkin@highlandtechnology.com> wrote in message
news:egcdpal3j5j8en0ead82236nm8g28pim5b@4ax.com...
On Fri, 3 Jul 2015 07:29:00 -0700 (PDT), Lasse Langwadt Christensen
langwadt@fonz.dk> wrote:

Den fredag den 3. juli 2015 kl. 12.52.52 UTC+2 skrev Martin Brown:
If you want to sell your soul for maximum financial gain then
destabilising the global stock trading systems with sophisticated high
frequency trading algorithms is definitely the way to go.

One guy in the UK in his parents bedroom can allegedly do this:

http://www.bbc.co.uk/news/business-32415664

People seem to get upset if you are too good at it!


yeh, you have to be a member of "the money sucking parasite club" to
manipulate prices and steal money like that


-Lasse

A small tax on transactions, like 0.1%, would have a remarkable
damping effect. Maybe we can get that soon, after the next monster
worldwide crash.


but most likely would destabilize the ...government!

...or at least the politician proposing it.

politicians are not going to bite the hand that feeds them

-Lasse