pyrojim wrote in post #17569568
Most applications are NOT optimized for multithreaded processors. Additionally, unless you are programming IN CUDA, a new GPU isn't going to save you.
Regardless of thread use, I'm running multiple applications at the same time, and most support CUDA acceleration too. Think I'm running half the Adobe suit while rendering and a VM and some other stuff.
You must take a darn-large number of photos.
Not a lot. Just taking photos is something I've stopped doing a while ago.
I still have trouble maxing out ONE core while running a simulation...
Depends on the simulation I guess? I thought that kind of stuff likes multi-threading and GPU acceleration, not that I really know...
Also why on earth use network attached storage? Do you really want to use a network connection to transfer your claimed work load? Now you are intentionally doing things the slow way!
Oh no, it's not about speed, but volume, I'm fed up with juggling a bunch of labelled disks to get data. There's no real way to have tens of terabytes of unified storage without building a RAID5/6 array, but I obviously can't configure an internal one via LSI card on an iMac.
Modern NASes have port linking, so once I do get a PC, I'll be sure to get a motherboard and network switch that support port linking and 10Gb, and I should be able to get up to 350MB/s speeds, which is more than enough for my purposes. For short-term storage speed I'm obviously going to need a bunch of SSDs.