Anyone with a Threadripper/X399 setup can you do me a favor?


TType85

Recommended Posts

I'm looking for someone who is running a X399 setup with a gaming VM on windows 10 who can run benchmark or two for me.  Specifically I am looking for CPU-Z single core and Cinebench R15 Single Core, along with what speed/type of ram you are using.

 

I am in the process of gathering parts for 1920x build to replace my E5-2667 v2 build. Gaming wise, I mostly play World of Warcraft when I game and it plays decent on my current build (4c/8t isolated for VM) but it is very single core speed reliant.  Between bare-metal and in a VM I am seeing frame rates halved in busy areas (60ish vs 30ish). In less busy areas it is fine >100fps most of the time.

 

Thanks in advance.

Link to comment
  • 2 weeks later...

Just as a follow up, I got the 1950X, I get a few oddities with my original VM.  I swap the core assignments and I get the same results (380ish on the old install and 410ish on the new install)  I have a few NVME drives on the way that will hold the VM's so I will reinstall and try from there.  

 

EDIT: Changed the windows power plan settings from High Performance to Balanced and now it's up to 409-410. Change it back to High Performance and it goes back to 380... Odd.

 

Windows 10 Install that was carried over from the E5

image.png.34cc0c7c9613c560c0f3896561ab06ca.png

 

Fresh Windows 10 install

image.thumb.png.f193ebb9968231c8fac7b26cfe62d43c.png

Edited by TType85
Link to comment

Based on the findings in another thread, it looks like your utilising the physical and SMT threads from opposing cores.

Try allocating in "columns" instead of across and see if that helps.

 

On 8/25/2018 at 8:58 PM, bastl said:

Threadripper_1950x_core_numbering.jpg.8c3bd0a2b67ff5e2c8a1c764baf8eb54.jpg

 

I'm hoping to finish setting up my gaming VM this weekend, so I'll try and get you some benchmarks.

Link to comment

Hey guys,

 

that picture from the other thread where my initial guesses what the core pairings should be shown by unraid after i did a couple benchmarks. Later in the thread i corrected myself. The cores 0-7 and 16-23 are on one die and the bottom half and the second die. In the case you're using cores from 2 dies you automatically using the memory channels from both dies and the bandwith is higher in that case. But you add a bit of latency to it. That also affected my Cinebench scores. You should also check how your memory interleaving is setup in the BIOS. That also affected my scores. Maybe you have a look into the thread where i posted all my testings.
 

 

Edit:

 

I did a quick single core bench on my current Win10 gaming VM with assigned cores 8-15 and 24-31 i never did before. I guess Cinebench should only use one core in that case, right? The load jumps from core to core during the bench. Weird

Edited by bastl
Link to comment

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Restore formatting

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.