1. "Load time" is first a function of disk I/O performance, not so much RAM. This is the part where you need to match your work flow to the bottleneck to the hardware. If your work flow has you loading one image every so often then even an SSD won't benefit you that much even though benchmarking software shows it's loading that image 20x faster (or whatever it may be). But, if you're loading 20-30 images at a time.. now you'll see a big performance benefit. And there will be a point where as you load those 20-30 images where you'll saturate your RAM and now RAM makes a difference.
Why don't you demonstrate, with a video on youtube, how using a fast disk can make loading 20-30 images faster than using a slow disk?
Or provide some timing results and document what you do so that others can do the same test as yourself to verify?
2. My point was that if you ARE running more than Lightroom as most tend to do, then more RAM most definitely will increase performance.
Will increasing my RAM from 12GB to 16GB make a big difference?
What about going from 16GB to 32GB?
3. I think it can for very specific work flows.
Can you provide some specifics on those work flows that benefit from huge amounts of RAM or faster disk?
4. No, that wasn't my counterpoint at all. And me building a web page on my work flow, or a specific work flow, won't help others understand their own work flow..
But it will help us understand how you work and how you benefit from faster disks and more RAM. Don't you think that this is useful information?
My counterpoint was partly yes, the SSD is demonstrably faster so I/O functions will be faster.
Note that it has been previously established that for large images, CPU time far outweighs I/O time.
With a work flow which loads one image at a time this will hardly be noticeable, but with a work flow that loads/saves multiple images at a time it will indeed be noticeable.
To the best of my knowledge, LR doesn't work on multiple images at a time - that is unless you've got multiple tasks running which I've found to be a bad idea of PCs in general.
So when you decide to do X with 20 or 30 images, they each get processed in serial. That means that LR does what it needs to do with one image, throws away the temporary data and starts on a new one.
Or if you're building a catalog/previews from an SSD, and I pointed out that perhaps the biggest gain is in the catalog.. doing searches and moving through libraries is light years faster than working off a normal mechanical drive.
So let me get this straight, you're advocating using SSD because it is faster but you're not even sure where the biggest gain is?
Have you actually timed any of your work flows to show that SSD is quicker when used for part X?
Not everyone will derive the same benefit from a specific work flow which is why I found that article flawed.
Thus far, you haven't provided any solid information, never mind anything that is any less flawed than that article.
It tested some, but not nearly all the functions where increased I/O performance is beneficial. The functions they checked would definitely show a performance increase in the I/O area, but if you're only using that function for a fraction of your work flow then it will appear it's not doing much at all.
Can you go into more details, please?
I'd very much like you to benchmark some actual work flows and show how and where benefit from SSD can be had.
Given that you know what to do, this shouldn't be very hard for you to demonstrate.
Afterall, people are adaptable and if there are better and more efficient ways for humans to work, I'm sure will adapt (or try to!)