20k messages by GerdW

[UliB]

Today GerdW posted his 20000th message. This means that he has posted 5000 messages within 2 years.
Congratulations on this absolutely amazing achievement!

[LiliMcDonald]

Congrats, Gerd! We appreciate all of your contributions to the community!

[altenbach]

Gerd, I always admire the quality of your contributions, the quantity is secondary. 😄

We have very little time overlap and it always seems that when I go to sleep late at night, GerdW appears well rested and ready to take over. 😉

Thanks for all your fantastic contributions! I learned a lot!

[GerdW]

Thanks to all of you!

Hi Christian, yes, we live at a (timezone) distance of 9h. And usually I'm early at work… 😄

[Ben]

---

@UliB wrote:

Today GerdW posted his 20000th message. This means that he has posted 5000 messages within 2 years.
Congratulations on this absolutely amazing achievement!

---

I'll say!

From my perspective... well I can say I am still spitting out the dust from when he went flying past.

He did wave as he flew past which was nice. 

Congratulations Gerd!

Ben 

[GerdW]

Thanks Ben!

I'm waiting for you in the 20k+ club… 😄

For all of you: In the attachment is a (very) simple version of the image used for my avatar.

(I created my very first version of a Fractal software about 30 years ago. At that time I used BASIC…)

[JB]

Congrats from CH !

With only 3k posts in 18 years, I will never join this club.

[RavensFan]

Congratulations Gerd!  Please take your foot off the accelerator a bit.  Otherwise if I blink you'll zoom right by me and I won't see it!

[altenbach]

---

@GerdW wrote:

For all of you: In the attachment is a (very) simple version of the image used for my avatar.

(I created my very first version of a Fractal software about 30 years ago. At that time I used BASIC…)

---

Very nice! That must have taken a long time to compute 30 years ago. :o:D

Note that things like that can of course be parallelized if you pull the ramps out of the loop. On my 8 core ryzen, things drop from 2.5s to 0.3s seconds, for a >8x speedup. 😄 Also, the inner ramp needs to be calculated only once (and not 600 times!) of course, but that's a very small difference. Taking the absolute value to get r is more efficient, because theta is not needed here. The below will generate a binary identical result.

[GerdW]

Hi Christian,

the VI above is a very quick&dirty version of my Mandelbrot program, put together just for this post to generate that simple image. (Mainly to avoid any subVIs, queues, FGVs, and so on.)

In my own tool I do all of your comments too. The parallelization is done by starting several workers computing parts of the image in parallel: this allows easier display of the already calculated image parts also in parallel.

On those FOR loops calculating the ramps: when you zoom in near the limits of EXT numbers it's problematic to do consecutive addition of very small numbers (delta, divided by pixel count) with a larger one (start value) - I also did some improvements for that corner case…

That must have taken a long time to compute 30 years ago.

Sure. It was a computer built around a 6502 at 1.8MHz. It used a BASIC version allowing floats with 10 significant digits, but rather slow computation speed (compared to other homecomputers at that time). On the other hand "graphics" were rather easy with just 8 or 16 different colors, instead of 2^24 nowadays…

@RavensFan:

I don't think this will happen in the near future, you still have a comfortable lead… 😄