I’m on a crusade to understand what is going on with Excel 2016 performance.
In summary its bad, like more than 10 times slower than Excel 2010 bad.
ie BAD.
I am only just scratching the surface, so I am sure I’ll find areas where 2016 is better … at least I hope I will…
My little cell selection test gave Excel 2016 2.6secs, Excel 2010 0.14secs.
Which I appreciate is indeed more than 10x, I could calc it exactly but it would take a couple of life times in Excel 2016,
But of course these are different machines so I need to understand how much is due to the machine. so I wrote a C++ console app (so limited windows, or .net influence).
It just adds some (vaguely) random numbers, so a bit dependent on that library, but in principle I don’t see a glaring reason why it can’t give me an approximation of the difference in speed between two machines.
This is it:
for (int n = 0; n < 1000 * 1000 * 1000; ++n) {
result += -70 + (std::rand() % (60 – 25 + 1));
}
It runs 14 seconds on my XP box and 15 on Win 10, so I think you could say the machines are comparable, roughly.
So the 10x slowness is pure Win10/E2016 then?
Well its a bit early to say, I think there is a video/graphics element so I’m going to develop some more tests.
I mean as it stands Excel is so so bad you could develop in VSTO without noticing a performance drop (unless its worse too).
I’ll keep you posted, ping me if you want this console exe to benchmark your machine.
cheers
simon
Leave a Reply