It's a pcie4.0 world, my friend. Or it will be in less than a year.
Given what I've seen from leaked pricing, the 570 boards will be $40-60 more than equivalent 470 boards. There is where I'd spend til it hurts for best lifetime from that expenditure. Today's polymer-electrolyte capacitors and the name-brand manufacturers' current tendency to over-build the VRM rather than under-build make this a MUCH safer investment than back in the
SOYO Dragon days.
I'd buy the best 570 board (minimum: 4 RAM Slots, 2 nvme-capable M.2 slots) and 2 modules/at least 16GB of the fastest known-compatible DDR4 RAM (DDR4-3200 minimum, I'd guess) I can afford first, and then the highest-on-the-foodchain 2
xx0
X Zen2 processor I can scrape the money together for after that. I'd even hold off on a nvme boot drive if I had any decent SSD to use; spend that money on better RAM/CPU. I'm expecting a new crop of larger MLC as opposed to TLC drives soon.
The demand for faster MLC SSDs as boot drives (with a larger, slower TLC as data storage in a 2nd M.2 slot) is fast becoming the standard arrangement for performance builds.Next on the agenda would be... as best price/capacity deals vs money come available... buy matching part # DIMMs to double up my existing RAM, and biggest MLC nvme SSD I can afford for a boot drive. Relegate existing SSD to data storage, then finally clone over & replace with a TLC nvme drive in the 2nd slot. EVEN the slowest TLC nvme drives now available are 5x-10x as fast as you can get from SATA; this gap will only embiggen
as pcie4.
x becomes the standard.
Anyways... this is generally the battle plan I've used for every major upgrade I've done in the last 3 decades: Buy the most advanced MB I can afford in the brand I prefer (usually ASUS/AMD), then most quantity/fastest RAM (Spend til it hurts on these two), then best bang/buck CPU available at the time that will run on that MB. Don't give a damn if you aren't running the RAM to its maximum speed; as CPU deals come along and upgrade money comes available,
then buy the next up on the food chain CPU until I maxx out the board. Use it until it no longer serves as a gaming/workstation, then replace the daily-driver web-surfing machine with it & start the next build.
It's true that getting 10 years out of my old PhenomII 1055T is a little long to stretch such an investment even for me; but the fact I could do that shows how well this strategy can keep you in a usable machine and truly get you both best performance you can afford AND best bang for the buck over time.
mnem
*Veteran of a Thousand Psychic Wars PC Builds*