The sharp eyed will have spotted that the spreadsheet @GM5ALX uploaded needed an fix/update (hidden columns causing issues). These are the final results:
Furthest Summit 2 Summits:
Summit 1 Ref
Name
Summit 2 ref
Name Summit 2
ARC Distance (with summit Height and Bulge)
Rank
ARC Distance with Earth Bulge
Rank
EA1/OU-006
O Xistral
ZL3/CB-101
Mount Una
20043.17
1
20037.06
7
EA1/OU-006
O Xistral
ZL3/TM-011
ZL3/TM-011
20042.59
2
20036.63
32
EA1/ZA-003
Moncalvo
ZL3/MB-085
ZL3/MB-085
20042.28
3
20036.36
63
EA1/OU-015
Peña Nofre
ZL3/TM-022
Mahanga Range
20041.96
4
20036.62
35
EA4/CR-066
Rayo
ZL1/MW-002
Mount Ngauruhoe
20041.88
5
20036.80
22
EA1/OU-001
Peña Trevinca
ZL3/MB-161
ZL3/MB-161
20041.75
6
20036.44
53
EA1/LE-267
Las Colinas
ZL3/CB-293
Snowflake
20041.59
7
20036.29
69
EA1/LU-019
Faro
ZL3/CB-406
The Nelson Tops
20041.57
8
20037.12
6
CT/TM-004
Armada
ZL3/WC-454
Lyell Range
20041.49
9
20037.25
5
EA5/MU-009
Gato
ZL1/BP-039
ZL1/BP-039
20041.45
10
20037.37
2
As you can see a CT summit makes the list CT/TM-004
Here is the list of the Most Antipodean Summits to Summits - Directly opposite each other:
Summit 1 Ref
Name
Summit 2 ref
Name Summit 2
ARC Distance (with summit Height and Bulge)
Rank
ARC Distance with Earth Bulge
Rank
EA7/GR-068
Cabras
ZL1/WK-217
ZL1/WK-217
20040.50
50
20037.42
1
EA5/MU-009
Gato
ZL1/BP-039
ZL1/BP-039
20041.45
10
20037.37
2
EA5/MU-011
Gigante
ZL1/GI-084
ZL1/GI-084
20040.71
35
20037.30
3
EA7/JA-090
Cerro de Gontar
ZL1/BP-141
ZL1/BP-141
20040.64
39
20037.25
4
CT/TM-004
Armada
ZL3/WC-454
Lyell Range
20041.49
9
20037.25
5
EA1/LU-019
Faro
ZL3/CB-406
The Nelson Tops
20041.57
8
20037.12
6
EA1/OU-006
O Xistral
ZL3/CB-101
Mount Una
20043.17
1
20037.06
7
CT/TM-038
Rebordondo
ZL3/TM-181
Blue Cliffs Ridge
20039.89
111
20037.02
8
EA7/JA-051
El Majalón
ZL1/BP-185
ZL1/BP-185
20040.81
32
20037.01
9
EA1/CR-031
Alto de Fernandiña
ZL3/CB-159
Hinge Peak
20040.98
21
20036.97
10
EA7/GR-068 Cabras is only 47m off the centre of summit ZL1/WK-217. Both of these look reasonably accessible to EA and ZL hams and might make for a interesting challenge.
New Zealand to EA & CT dominates the list by virtue of these being antipodean and the only SOTA regions that are. Many countries would be, but are not yet SOTA regions. Its is not until entry 642 (assuming Furthest) or 57 (Assuming most antipodean) do you get Japan to Brazil !
Summit 1 Ref
Name
Summit 2 ref
Name Summit 2
ARC Distance (with summit Height and Bulge)
ARC Distance with Earth Bulge
JA6/KG-177
Ontake
PY3/PA-010
Morro Cantagalo
20037.55
20035.83
Thanks to Alex @GM5ALX for the supercomputing to enable this. Apparently the sun was actually shining for a change in Scotland and it was solar powered
For reference: We used the formula to derive the “Circumference Segment” calculating the effective radius at each point (factoring bulge and summit height), the central angle between the two points and then the calculating the circumference segment (arc length). Is you want to know more, ask ChatGPT…
The whole subject is thought provoking for two reasons. One is my long engineering career has paid the bills through 75-80% software 20-25% electronics. Yet it’s a long time since I’ve written specific parallel processing software, 25 years to be precise. You tend to forget, or push to the back of you mind, a lot of things you learn writing for big parallel arrays of DSPs. The second is despite spending about 20% of my current programming time using Python, I’ve never need to do most of the tasks this problem needs. Never needed multiprocessing in Python or any floating point. It’s always been take a file in format A, massage it to format B. Or callbacks in a simulation system to configure specific simulation performance. The rest of the time is C++ code and again you never end up with more than a handful of threads running at one time. The whole discussion is, to my mind, wonderfully thought provoking. You program for a living, think you know what’s what and then a whole new challenge comes along.
Which brings me to the next point. Once you know the longest paths etc. the urge to arrange with other idiots, sorry activators the other side of the planet, attempts to S2S these ultimate paths. It does look like a holiday to Southern Spain is in order
Finally once you have a list of the longest short paths for all the summits, people who live where their longest paths are not that long can always try to arrange S2S schedules where they can demonstrably prove they used the long path
The Cray-1 is 50 years old this year. They were so awesomely fast when released they still stick in the minds of old giffers like me who were impressionable school kids at the time.
i5 9500 single core = 422000 megaFLOPS (floating point operations per second)
Cray -1 = 160mega FLOPS
The i5 9500 has 6 cores so in theory that’s 2.532 teraFLOPS but you normally hit memory/CPU bandwidth limits before you reach that.
That’s a shame. Nothing in the new max distance list above Faro-Nelson Tops looks like a nice place to be in the dark pursuing grey line.
ZL3/MB-085 and ZL3/MB-161 are realistic evening activations with a torchlight descent to a valley camp. But their EA pairs look, well … challenging.
ZL3/CB-293 Snowflake is ok in daylight, but questionable descent in the dark for the sake of an extra 20m! So would mean a ‘dry’ (carry water) tops camp on rock/scree.
In the late 1980s, I bought a copy of Borland TurboC for my first PC.
Using the now legendary “Numerical Recipes in C” book published in 1988, I was able to rustle up the code to solve n equations in n unknowns, using the “LU Decomposition” algorithm.
I recall that for 50 equations with 50 unknowns, the PC took a few seconds to come up with the solution. Things were helped along because I had ‘pushed the boat out’ and bought an 80287 floating-point coprocessor. As such, the maths was done largely in hardware rather than with a nasty slow emulator.
My pal worked at the Council, and had access to their IBM 3081 mainframe. He crafted a version of my program in FORTRAN; the Council didn’t need, so didn’t have, a C compiler.
Long story short, my PC beat the IBM mainframe into a ‘cocked hat’ when it came to floating-point. As the Council workload was essentially Data Processing, I don’t think they had any (optional and expensive) floating-point hardware.
But horses for courses. My PC would certainly have struggled with 500 logged-on Council users trying to process everybody’s Poll Tax
Not really! I went to university intending to be a scientist but it didn’t work out that way. What I (and my department) actually did was much closer to engineering than science. We didn’t call it that for essentially political reasons.
Quite so. I am very happy with approximations. But the problem here is that we’re trying to establish a ranking order rather than determine the actual distance. If you want to assert that something is the largest of its class, the answer is either right or wrong.
Indeed. One thing I’ve relatively recently come across is PostGIS. If the SOTA database were PostgreSQL with the PostGIS extensions installed, all of these calculations on geographic objects would be right there at the SQL level. It is very impressive.
Years ago, when I was a younger Engineer, I worked at a company that designed and built Burn-in/ test equipment for Cray. I seem to remember it was for ECL SRAM devices.
We also did boards for IBM, ICL, Phillip’s and Motorola components.
Andy
MM7MOX
The real problem is not the extensions are there but that the data isn’t exported for others to play with. The level playing field is the summitslist.csv file, everyone gets access to that.
Ignore the initial data, error with hidden column in the spreadsheet means that’s from the Greenwich meridian. Use these links:
Looks like a very accessible peak ZL1/WK-217 on the Coromandel peninsula that is the most antipodean summit! @ZL1THH . But then again I’m not local - maybe an active volcano, have DoC controls or like wall to wall bush !
