actually more horsepower isn't necessarily helpful - I was reading earlier today about a point cloud architecture application that preallocates larger jobs by g

actually more horsepower isn't necessarily helpful - I was reading earlier today about a point cloud architecture application that preallocates larger jobs by guessing the capabilities of the multiprocessor available. This isn't a low end application but a banner product for Leica Geosystems. the reviewer could only get it to slice the job into 8 sections despite many more cores available. in filmmaking, personal equipment rental companies like sendgrid have raised the level of everything now I can buy my own Red and lenses and learn my tools intimately and pay for them (when I'm not paid to bring them on set) by rental customers who are liberated by the freedom from having to deposit the replacement value with traditional rental houses which risk is now insured. nvidia should sell a licence to use the 3090 professionally and rent it out like you can with the cards formerly known as Quadro now RTX A* beginning RTX A6000 (48GB VRAM at under $3K list is incredible if you can get one 96GB at 1TB/s+ linked..) if they really want to take the visual graphics industry for good. (equally AMD can grab this business forever just as easily with a turn on a penny worth of engineering) Architecture and geographical systems are strongly and culturally averse to parallel programming. the complexity involved has only recently been significantly helped by multiple cores - the workstation on my desk in 2005 was dual dual core 3.9GHz xeon. 32GB RAMBUS RAM. early solid state swap. (and four FC drops from a SAN shared array with enough RAM cache to hold geometry models) and laptops can be made to run single threaded codes very quickly single core, satisfying the (self imposed limits of) potential for most "graphics heavy" programs. But it isn't at all a good situation. Ed removed confusingly placed sentence