Very Bad Ram & Cache performance within VM vs Good Performance on Bare Metal.

[sonuyos]

So i am running a VM in unraid using my physical disk, as i want to retain as much power as possible,

My CPU performance is 7% low, thats kinda expected as it is 16 thread, and 2 threads are probably reserved by unraid.

However, my Memory is performing worse.

 

In VM - https://i.imgur.com/Ki6qe1g.png

In Bare Metal - https://i.imgur.com/YR8l52x.png

 

Can anyone help me?

As i am running off SSD on my VM as well as bare metal, the enviornment is same.

 

VM CPUZ Cache - https://i.imgur.com/VTHETSe.png

BM CPUZ Cache - https://prnt.sc/qjqpwn

[jonp]

Benchmarks don't always reveal a problem in tests like these. What I'd be curious about is what apps / games you are running and how they compare in performing similar tasks.

Benchmarks definitely push the raw hardware to the limit, but don't often take into account other bottlenecks that would appear when using actual apps.

What is your intended use case for the VM?

Sent from my Pixel 3 XL using Tapatalk

[sonuyos]

7 minutes ago, jonp said:

Benchmarks don't always reveal a problem in tests like these. What I'd be curious about is what apps / games you are running and how they compare in performing similar tasks.

Benchmarks definitely push the raw hardware to the limit, but don't often take into account other bottlenecks that would appear when using actual apps.

What is your intended use case for the VM?

Sent from my Pixel 3 XL using Tapatalk
 

Thanks for you Reply.

 

I have ran some gaming benchmark, as i have seens 1-2% low fps, that can be attributed very well to error margin, Cinebench diff between VM and BM is of 7%, which is truly expected.

 

My main use for this is Gaming + Work (Browser mostly, hence ram is of big use).

[jonp]

6 minutes ago, sonuyos said:

as i have seens 1-2% low fps

 

9 minutes ago, sonuyos said:

Cinebench diff between VM and BM is of 7%, which is truly expected.

 

Your benchmarks for the apps/games seem fine.  I don't see a problem here.

[sonuyos]

1 minute ago, jonp said:

 

 

Your benchmarks for the apps/games seem fine.  I don't see a problem here.

Why is there a problem with cache tho? I am still not sure about that one, i am just curious about it.

 

Also i would like to report a bug with Ray Traced Games I see a lot of artefacts while running Shadow of Tomb Raider. However after a reboot it was fixed, but i will keep my eyes open for it.

[jonp]

Artifacts in a game, unless you can consistently reproduce, can't be assumed as a bug. Artifacts can appear due to a faulty gpu, drivers, heat, and a number of other factors. If you could consistently reproduce artifacts in a VM that don't appear on bare metal with the same hardware, we might have something to investigate, otherwise I'd chalk this up to a non VM but that caused it.

With respect to the cache, what gets reported in a tool and what is actual are two very different things sometimes. E.g. lots of GPUs show up at 1x speed when using tools like GPUz in a VM, but that does not mean the GPU is really running at 1x. That's just what's being "reported" to the VM due to the underlying virtual hardware used for emulation.

This is one of the many reasons why we ask folks to look at how their apps perform and not what benchmarks report, because benchmarks and other reporting tools won't always display the right information when running inside a VM.

Sent from my Pixel 3 XL using Tapatalk