Asrock wrx80 threadripper 5975 hitting a limit of about 30gb/s

Hi All,
I havent really scratched around / troubleshooted this yet but i thought id post this here so long.
I have a asrock wrx80 creator v1 board + threadripper 5975 cpu, have 128gb ddr4/3200 ecc reg memory

I have 2 x 2tb seagate firecuda 530 in raid0 (on motherboard)
then I have 4 x 2tb seagate firecuda 530 in raid0 on a highpoint ssd7505 controller

As seen in screenshot attached I was able to hit almost 38gb/s in combined benchmark. this would be running crystaldisk at the same time benching both devices.

So yesterday I added the following disks to the system,
1 x 800gb intel optane p5800x
1 x 960gb intel optane 905p

So now i open 4 x crystaldisks and run the benchmark at the same time, I noticed some of the
disks showing degraded seq performance (e.g p5800x showing 4xxx mb/s instead of 7400mb/s…
basically add all the bench up and it seems there is some sort of 30gb/s bottleneck

Where before just the 2 x raid0 volumes would hit almost 38gb/s combined.

Before everyone goes all nuts on me about the bench, I know seq isnt the be all end all
heck i know crystaldisk isnt the “authority” on benchmarking either.

Removing the 2 x u.2 (intel optanes) doesnt change this situation either.

wondering if its not a windows update somewhere along the line that changes something,
I should really get on with installing linux on this rig.

Any ideas?

From what I remember AMD raid really really blows and md raid is the way to go

I believe you can md raid on Linux then pass that through virtualization if you want windows

This topic was automatically closed 273 days after the last reply. New replies are no longer allowed.