New PC build for editing 4k in Davinci Resolve studio

Hi I'm new to the forum so bear with me....I'm building a new PC windows 10 based editing computer for Davinci resolve studio and I'm looking for some advice on the best components to buy in the UK.
1. Is it better to use an Intel i7 or i9 or the AMD Ryzen R9 processors?
2. Which GPU cards do people recommend, The new RTX 3080 or better?
3. Are there any Motherboards that perform better for video editing?
4. I will be using SSD drives M.2 and the Samsung EVO both 1TB, Any suggestions?.
5. Memory, I'm looking at 32 gig of DDR4 3200 or even 64 gig....Any suggestions on make ?

Hope you can help and thankyou in advance for any replies.

Regards Greeny1080

My choice would be Intel LGA 2066 CPU and a 3000 series RTX card.

I would also include enough drives for the following setup to begin with, which could be expanded upon later.

System (M.2)
Database (M.2 or SSD)
Export (4+ TB Spinner)
Gallery/Cache/Optimized Media/Proxies (2+ TB SSD)
Media (10+ TB Spinner)

Again, Jim? Your drive allocations make no sense, if you don't know how other people work and their particulal requirements. This Premiere left-over is maybe best forgotten. And a dedicated drive for databases? Since when? And why?

Used to be for speed, now more for organization.

While we use NVidia cards, the GPU Out of Memory error has me seriously considering AMD cards. 6900 XT is looking very attractive.
For CPU - AMD or Intel with the most cores you can afford.
More RAM the better. RAM speed is not crucial.
A decent NVME SSD will work great as your OS/App drive.

1) The selling point for AMD CPU's over Intel's for me right now is mostly that AMD's support PCIe Gen4, which opens you up to very, very fast SSD's - 7GB/s (double the speed of Gen3's, 10-20x SATA SSDs, and let's not even talk about spinning disks). The huge number of cores of AMD's offerings also makes light work of a lot of encoding/transcoding tasks that aren't handled by the GPU. Intel will be releasing their first Gen4 desktop CPU's in March, but then it'll be mid-year before you can buy them...

2) As mpetech mentioned, NVidia GPU users are currently plagued by out-of-memory errors, mostly around Fusion and such workloads. But I'd still find it hard to go with an AMD GPU because of how fast NVENC makes exports and transcoding. Even my generation-old mid-range 2070 will encode at >4x the speed of the CPU. Hopefully NVidia and Blackmagic get their driver/Cuda/memory issues sorted soon.
Whatever you choose, I expect buying something with as much memory as possible would be money well spent. I have to say, my 8gb card really isn't cutting it - but that might change as the v17 beta cycle continues and memory issues get fixed.

3) If you go with an AMD CPU, make sure you get a X570 motherboard not X470. Otherwise you won't get PCIE Gen4. Other things to look out for are USB 3.1 Gen 2 Type-C on both the back panel and headers for the front panel (and a case with a front Type-C port!), which will eliminate many motherboards and most cases. Fans headers. You can never have enough fans for a high-end machine; things will get HOT!

4) For SSDs, as I said under CPU's, Gen4 NVMe SSDs are just insane. I have a Gen4 for OS and databases, a Gen3 for Cache/Optimized Media/assorted media, another Gen3 for project footage (using an external enclosure so I can keep the original data, but the perf hit of external is OK here), then an ancient NAS for backups.

5) Like the video memory, it depends on what kind of projects you'll be working on. If you want to use Fusion at all, 64gb is quite necessary in my experience. Again, betas and all that...

Says more than I could say: https://www.pugetsystems.com/recommende ... 87/Buy_214

I've seen rx 6800 (xt) murder 3080 and 3090 in TNR in the standard candle benchmark on Computer base website.

Also the NVENC encoder argument in favor of Nvidia is a bit missed since resolve supports AMD hardware encode. Just use the right parameters.

I'm not saying AMD is better (it's not) just saying that one may not need to pay the full "Nvidia price" for the similar performance, depending also on the use case.

Sent from my Mi 9T using Tapatalk