[concept_fuel]

AnTiLooP wrote:

I will try to disable to see what I get. Interesting though. Any theory why this is happening ?

I wouldn't worry about it.
It's normal for your base clock or bclk or any of the clks like rtc(real time clock) to have small variances as this depends on the quality of the crystal oscillator and voltages variances that control these oscillator circuits(VCXO, PLL, etc) on your mobo, similar to the quartz crystal found in watches i.e. the time variances between them.

[AnTiLooP]

concept_fuel wrote:

I wouldn't worry about it.
It's normal for your base clock or bclk or any of the clks like rtc(real time clock) to have small variances as this depends on the quality of the crystal oscillator and voltages variances that control these oscillator circuits(VCXO, PLL, etc) on your mobo, similar to the quartz crystal found in watches i.e. the time variances between them.

I get that... But after reading several articles... It seems it's a known fact that svm with hyperV causes the degradation mentioned.

Svm with virtual box, VMware or xencentre has no such behaviours.

Not an issue, I'll just move to VMware haha. I've got the licensed install media anyway... Problem is, resetting up my lab..

[concept_fuel]

don't use hyper-v but have you tried installing it without svm and then enabling it after hyper-v is installed?
From what I gather when installing hyper-v if the svm feature is detected then the whole Windows host OS is executed as a privileged guest. The performance drop is likely due to the virtualization overhead of running the system as a guest.

[AnTiLooP]

concept_fuel wrote:

don't use hyper-v but have you tried installing it without svm and then enabling it after hyper-v is installed?
From what I gather when installing hyper-v if the svm feature is detected then the whole Windows host OS is executed as a privileged guest. The performance drop is likely due to the virtualization overhead of running the system as a guest.

I've already mentioned the behaviour in my post at 835pm yesterday.

Anyway it's virtual box or VMware. Another decision to make.

[Proximus]

AnTiLooP wrote:

I've already mentioned the behaviour in my post at 835pm yesterday.

Anyway it's virtual box or VMware. Another decision to make.

Full agreed that is a tough decision to switch Hypervisor. IMHO, it is do make perfect sense in the chase of a couple of Mhz or even the performance increase < 20%. This taken in considersation on the issues you might encounter in type 2 hypervisor which you did not faced earlier. It is like opening up a can of worms.

Been there before when I make "stupid" decisions and spend the rest of the week fixing the issues.

[AnTiLooP]

Proximus wrote:

Full agreed that is a tough decision to switch Hypervisor. IMHO, it is do make perfect sense in the chase of a couple of Mhz or even the performance increase < 20%. This taken in considersation on the issues you might encounter in type 2 hypervisor which you did not faced earlier. It is like opening up a can of worms.

Been there before when I make "stupid" decisions and spend the rest of the week fixing the issues.

I know right.... Anyway I just disabled HyperV and ran some benches... yeah getting some extra mhz... like 20mhz. ... getting a full bus clock of 100mhz.

Then before clicking install on the new hypervisor, I was thinking you know what... I mean if I want the full performance I can simply just disable HyperV just like how i just did it.

alright, idea has been shelved. Not migrating over. I'll just switch modes if necessary - the thought of starting from scratch and building my lab, or even V2Ving is so much hassle that it isnt worth it.

in fact, if I need access to my lab environment, I'll just simply run a quick batch file to invoke bcedit to enable. LOL

note:
bcdedit /set hypervisorlaunchtype on
bcdedit /set hypervisorlaunchtype off

[d3fc0n]

yay my u12a is here!


still havent get to buy my strix 2080ti 11gb OC. the price like wont drop.. maybe wait till BF..

[ragnarok95]

Forget it la. Just go 3000.

[d3fc0n]

ragnarok95 wrote:

Forget it la. Just go 3000.

feels like v chor

[Asakura]

using thermalright

[Proximus]

Asakura wrote:

using thermalright

On which 3000 series and how it's the performance?

[AnTiLooP]

Proximus wrote:

On which 3000 series and how it's the performance?

Now all appear to be copying noctua. Lol

[aikishi86]

My RMA-return 3700X just came in today.... hmmm i was thinking of reselling or should i keep ?

[aikishi86]

Seem like my thermal paste had settled in well. The temps had more or less stabilised, was playing COD:WW2 last night, max cpu die temp was only 68degC, CPU Die temp also drop to 43 last night too. Using MX-4 (not the 2019 version) btw

[AnTiLooP]

aikishi86 wrote:

Seem like my thermal paste had settled in well. The temps had more or less stabilised, was playing COD:WW2 last night, max cpu die temp was only 68degC, CPU Die temp also drop to 43 last night too. Using MX-4 (not the 2019 version) btw

Undervolt to bring it lower haha.

Ur spare CPU can use for a new rig for dedicated prawnhub.com viewing.