Jump to content

What kind of workloads does this intel security bug effect

EzTech

I heard that processing performance can be reduced by 35% for certain workloads, but I do t know specifically what it effects. (Sorry I'm a bit of a noob at this)

Link to comment
Share on other sites

Link to post
Share on other sites

It affects mostly workloads that use a lot of IOPS, so stuff like databases. 

 

On the desktop it doesn't really affect much. 

 

idk

Link to comment
Share on other sites

Link to post
Share on other sites

I'd assume that the 30% remark was towards server loads which take up all the resources that a CPU can give it.

From what I understand, gaming performance may go down (a bit) but nowhere near as much as the server tasks.

 

As a disclaimer - I'm not too deep into CPUs. If somebody sees that I have said something wrong - please correct it.

Fan Comparisons          F@H          PCPartPicker         Analysis of Market Trends (Coming soon? Never? Who knows!)

Designing a mITX case. Working on aluminum prototypes.

Open for intern / part-time. Good at maths, CAD and airflow stuff. Dabbled with Python.

Please fill out this form! It helps a ton! https://linustechtips.com/main/topic/841400-the-poll-to-end-all-polls-poll/

Link to comment
Share on other sites

Link to post
Share on other sites

Thanks for the replys. Also, will these cpus ever return to their original performance?

Link to comment
Share on other sites

Link to post
Share on other sites

This entire 'Meltdown and Spectre' is really Frustrating. Imagine those who literally just bought an 8700k only to realise that they will be loosing 5% on its performance. 

Link to comment
Share on other sites

Link to post
Share on other sites

The only way to combat this issue, is to overclock. That way, atleast you would be gaining performance, even if its not as much before the kernel patch.

Link to comment
Share on other sites

Link to post
Share on other sites

5 hours ago, EzTech said:

Thanks for the replys. Also, will these cpus ever return to their original performance?

Not if you want to keep them secure.

You have two options, and it will remain that way forever (for the current processors):

1) Don't apply the patch. This means you keep the original performance, but you're vulnerable to attacks.

2) You apply the patch and lose some performance.

 

Future CPUs will most likely have these issues fixed.

Link to comment
Share on other sites

Link to post
Share on other sites

6 hours ago, JustAStrangeGeek said:

This entire 'Meltdown and Spectre' is really Frustrating. Imagine those who literally just bought an 8700k only to realise that they will be loosing 5% on its performance. 

That just happened to me... My MoBo didn't arrive yet, but I already have performance reduction. Being kind of salty, should've considered Ryzen 1800x instead of 8700k. 

Link to comment
Share on other sites

Link to post
Share on other sites

11 minutes ago, MementoMori said:

That just happened to me... My MoBo didn't arrive yet, but I already have performance reduction. Being kind of salty, should've considered Ryzen 1800x instead of 8700k. 

AMD and ARM are also affected, so it wouldn't have mattered. Hopefully this issue should be resolved by 9th gen.

 

My methodology for testing the meltdown is to stress test the CPU before and after the Windows kernel patch using cinebench to see how much 'spectre' actually affects your processor.

Edited by JustAStrangeGeek
Link to comment
Share on other sites

Link to post
Share on other sites

5 hours ago, JustAStrangeGeek said:

AMD and ARM are also affected, so it wouldn't have mattered. Hopefully this issue should be resolved by 9th gen.

 

My methodology for testing the meltdown is to stress test the CPU before and after the Windows kernel patch using cinebench to see how much 'spectre' actually affects your processor.

That's not a good way of testing things because the patch will have a different impact on performance based on the type of workload. Cinebench might show a 1% difference even though the average in all the tasks you do might be 5%. Or vice versa. Cinebench might show a 10% difference while the average impact on you might only be 2%.

Link to comment
Share on other sites

Link to post
Share on other sites

7 hours ago, JustAStrangeGeek said:

AMD and ARM are also affected, so it wouldn't have mattered

AMD handles their stuff differently, you would need physical access to the machine to exploit the bug where as Intel chips can be exploited remotely 

Link to comment
Share on other sites

Link to post
Share on other sites

5 hours ago, Evanair said:

AMD handles their stuff differently, you would need physical access to the machine to exploit the bug where as Intel chips can be exploited remotely 

LOL, almost exact words from Jayztwocents, love it.

Link to comment
Share on other sites

Link to post
Share on other sites

6 hours ago, LAwLz said:

That's not a good way of testing things because the patch will have a different impact on performance based on the type of workload. Cinebench might show a 1% difference even though the average in all the tasks you do might be 5%. Or vice versa. Cinebench might show a 10% difference while the average impact on you might only be 2%.

Yes, but it gives you an idea.

Link to comment
Share on other sites

Link to post
Share on other sites

5 hours ago, Evanair said:

AMD handles their stuff differently, you would need physical access to the machine to exploit the bug where as Intel chips can be exploited remotely 

Ehm, source on needing physical access for the AMD exploit?

Link to comment
Share on other sites

Link to post
Share on other sites

33 minutes ago, LAwLz said:

Ehm, source on needing physical access for the AMD exploit?

http://chronicle.augusta.com/news/business/2018-01-04/who-s-affected-computer-chip-security-flaw

 

https://www.theregister.co.uk/2018/01/02/intel_cpu_design_flaw/

 

https://lkml.org/lkml/2017/12/27/2

 

Quote

AMD processors are not subject to the types of attacks that the kernel
page table isolation feature protects against.  The AMD microarchitecture
does not allow memory references, including speculative references, that
access higher privileged data when running in a lesser privileged mode
when that access would result in a page fault.

 

Link to comment
Share on other sites

Link to post
Share on other sites

31 minutes ago, LAwLz said:

Ehm, source on needing physical access for the AMD exploit?

I have also read that you need physical access to the computer but I don't remember where I saw it. but AMD users don't need the patches at the moment.

http://www.tomshardware.co.uk/meltdown-spectre-exploits-intel-amd-arm-nvidia,news-57627.html This explains it better than I can.

 

Link to comment
Share on other sites

Link to post
Share on other sites

Just now, Evanair said:

None of those sources validates the claim that AMD processors are only vulnerable to exploits with physical access to the computer.

Also, I am fairly sure you don't understand what the quote you posted says.

Hint: It does not say AMD processors are not vulnerable to Spectre, and it does not say you need physical access to the computer.

 

2 minutes ago, Witchart said:

I have also read that you need physical access to the computer but I don't remember where I saw it. but AMD users don't need the patches at the moment.

http://www.tomshardware.co.uk/meltdown-spectre-exploits-intel-amd-arm-nvidia,news-57627.html This explains it better than I can.

There is only one patch released, and it fixes the Meltdown exploit. AMD processors does not need that patch, but what they do need are patches for Spectre. It's just that no patch fixing that has been released yet.

The Intel specific issue has been fixed.

The issues affecting all processors (including AMD, Intel and ARM) have no fix yet.

Link to comment
Share on other sites

Link to post
Share on other sites

2 hours ago, Evanair said:

AMD are actually still affected. They don't encrypt the 'Bounds Check Bypass' in the kernel. So yeah

Link to comment
Share on other sites

Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

×