Jump to content
Sign in to follow this  
Nemo

Intel/Windows 30% performance reduction

Recommended Posts

2 hours ago, FlightSimmer68 said:

Nobody knows *exactly* what any patch will do. So far it's just pure FUD which teh interweb will love and speculate about endlessly, even in spite of any facts that may emerge.

People do - the Linux kernel patches are starting to appear. They'll need to do a TLB flush when switching from user-mode code to the kernel, which will slow down that switch since they'll need to reload page table entries.

Bottom line on the effect is it depends on how much time your code spends in the kernel. You can probably take a look in Task manager to get a rough idea.

Cheers!


Luke Kolin

I make simFDR, the most advanced flight data recorder for FSX, Prepar3D and X-Plane.

Share this post


Link to post
Share on other sites

Every Intel CPU is affected. In normal use there will be a very small performance loss - not a big deal. BUT, the "up to 40% impact" is mostly measured on heavy IO loads, with is (again, as far as I know) a big deal for flight simmers. The sim reads a lot of data. So this could be a major issue for us. Time will tell, I hope Im wrong...

Share this post


Link to post
Share on other sites
11 minutes ago, Humpix said:

BUT, the "up to 40% impact" is mostly measured on heavy IO loads, with is (again, as far as I know) a big deal for flight simmers. The sim reads a lot of data.

No, it doesn't. It's primarily CPU and GPU dependent - this is why SSDs don't affect the frame rate much, if at all. In the MS world, you'll primarily see the impact in initial load.

Cheers!

 


Luke Kolin

I make simFDR, the most advanced flight data recorder for FSX, Prepar3D and X-Plane.

Share this post


Link to post
Share on other sites

Hmm, I already have troubles to load the autogen and terrain while flying, so I always thought that the SSD is the bottlenek. For sure, the task manager shows less load on the SSD, but it does not cover the IO load, only the overall MB/s load.

But anyway - this makes some hope. Hopfully im wrong. - we will know more next week when the patch is out...

Share this post


Link to post
Share on other sites
7 hours ago, Adrian123 said:

just downloaded the Intel test tool. yep, vulnerable. 7700k

 

 

This is for different bug and not relevant in this case. Perhaps the link could be removed as it appears to be misleading people?

  • Upvote 1

Share this post


Link to post
Share on other sites

No wonder I was holding off on a new system.

I have been ready to buy a "Mother of All PC's" for 8 months now, but for whatever reason I had a "red alert" to hold off and wait, but I did not know why. Nevertheless, my current PC is doing great until that red alert goes away.

Share this post


Link to post
Share on other sites

So bottom line with the verification tool is that this, risk assessment tool (Intel-SA-00086-GUI.exe ) is not checking for the current problem, Kernel Page Table Isolation (KPTI), there is no risk assessment tool for this yet.

(Intel-SA-00086-GUI.exe) It is checking the code for MEI and TXEI code issues, so your Verification Vulnerability reports showing green "not Vulnerable" output are only for the below vulnerabilities

Intel® Management Engine Interface (Intel® MEI) driver 
or
Intel® Trusted Execution Engine Interface (Intel® TXEI) driver

Its going to be a bumpy ride, and we will not like the medicine the boffins plan to treat this (KPTI) cancer with. It all amounts to Intel corner cutting for market supremacy over AMD. Intel's Speculative Code Execution shortcuts certainly speed up code execution but expose the Kernel to User code security violation. But we can all see now this was a big mistake and we must bear the performance loss with the fix they propose to close the hole. (@#$% cowboys).

 

 

Share this post


Link to post
Share on other sites

Disregard.


 i7-6700k | Asus Maximus VIII Hero | 16GB RAM | MSI GTX 1080 Gaming X Plus | Samsung Evo 500GB & 1TB | WD Blue 2 x 1TB | EVGA Supernova G2 850W | AOC 2560x1440 monitor | Win 10 Pro 64-bit

Share this post


Link to post
Share on other sites
9 hours ago, Nyxx said:

i7 4790K = NOT Vulnerable

Allelujah. Balking at early adoption pays off again!

Question. For all those vulnerable.  Does OC bypass the issue?


Frank Patton
MasterCase Pro H500M; MSI Z490 WiFi MOB; i7 10700k 3.8 Ghz; Gigabyte RTX 3080 12gb OC; H100i Pro liquid cooler; 32GB DDR4 3600;  Gold RMX850X PSU;
ASUS 
VG289 4K 27" Monitor; Honeycomb Alpha & Bravo, Crosswind 3's w/dampener.  
Former USAF meteorologist & ground weather school instructor. AOPA Member #07379126
                       
"I will never put my name on a product that does not have in it the best that is in me." - John Deere

Share this post


Link to post
Share on other sites

How many more times? The test tool linked here earlier is NOT for this issue. 

 

Share this post


Link to post
Share on other sites
10 minutes ago, fppilot said:

Allelujah. Balking at early adoption pays off again!

The tool that was linked and all results from it are not relevant for this latest bug, so you are definitely not in the clear.

Most reports seem to implicate all Intel CPUs in the past 10 years, and possibly much older. While AMD systems don't seem to be vulnerable, ARM processors appear to be (think Android phones).

https://security.googleblog.com/2018/01/todays-cpu-vulnerability-what-you-need.html

https://arstechnica.com/gadgets/2018/01/whats-behind-the-intel-design-flaw-forcing-numerous-patches/

Share this post


Link to post
Share on other sites
30 minutes ago, goates said:

The tool that was linked and all results from it are not relevant for this latest bug, so you are definitely not in the clear.

Most reports seem to implicate all Intel CPUs in the past 10 years, and possibly much older. While AMD systems don't seem to be vulnerable, ARM processors appear to be (think Android phones).

https://security.googleblog.com/2018/01/todays-cpu-vulnerability-what-you-need.html

https://arstechnica.com/gadgets/2018/01/whats-behind-the-intel-design-flaw-forcing-numerous-patches/

I am attempting to process all of this. Meanwhile my system churns along according to specs, without outside influence....


Frank Patton
MasterCase Pro H500M; MSI Z490 WiFi MOB; i7 10700k 3.8 Ghz; Gigabyte RTX 3080 12gb OC; H100i Pro liquid cooler; 32GB DDR4 3600;  Gold RMX850X PSU;
ASUS 
VG289 4K 27" Monitor; Honeycomb Alpha & Bravo, Crosswind 3's w/dampener.  
Former USAF meteorologist & ground weather school instructor. AOPA Member #07379126
                       
"I will never put my name on a product that does not have in it the best that is in me." - John Deere

Share this post


Link to post
Share on other sites

Hi Frank,

1 hour ago, fppilot said:

Meanwhile my system churns along according to specs, without outside influence....

Yes your current Intel CPU will continue to churn along happily passing potentially damaging code intercepted by malware or hacker from User directly at the Kernel then back to the User, unbeknown to the user because Kernel memory is supposed to be protected from such security risk. The problem comes with Intel's speculative approach to the exchange, by design a critical flaw un-noticed until the Linux guys questioned the process because of some unusual data exchange between secure Kernel and User (not as secure) pre-emptivley caching data that may or may not be correct according to an algorithm (which is usually correct about what to cache and send on back to the user), but it is exactly this vulnerability which can be easily exploited.
Between User Mode and Kernel Mode and back to User Mode, parsed stamped and patted on the backside goes the data, back and fourth with nary a security check from Intel's speculative process, on all Intel CPU's for the last decade.
The problem for the PC user, comes later on with the OS patch side. To fix the gaping security hole they must nobble your processor by creating a man in the middle Kernel to User isolation cache (PTI) (Page Table Isolation) as a workaround for current Intel processors. The fix provided for the Operating System is where we may see performance drop up to 30% simply because of the extra time it will take the (PTI) man in the middle cache, to exchange the data between both parties.

It appears that the next generation Intel CPU's will not have this design flaw, and will not need the K(PTI) man in the middle OS software patch.

The only thing about the whole debate in the linked articles on the interweb, at least for me is the constant reference to (x86) 32bit process I'm assuming this also refers to x86 CPU's  rather than a x64 CPU's 64bit process. hmmmmm... I am wondering if we might be ok afterall.

  • Upvote 1

Share this post


Link to post
Share on other sites

Another way to look at this would be that this "flaw" is a method to improve performance, and it was probably introduced by Intel in good faith a long time ago.

What is bugging me is, that we have to spend money, time, disk space and quite a few processor cycles just because some people don't understand the difference between right and wrong. Imagine how well a computer could be humming along if we didn't have to care about some word not allowed breaking it.

Best regards


LORBY-SI

Share this post


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
Sign in to follow this  

  • Tom Allensworth,
    Founder of AVSIM Online


  • Flight Simulation's Premier Resource!

    AVSIM is a free service to the flight simulation community. AVSIM is staffed completely by volunteers and all funds donated to AVSIM go directly back to supporting the community. Your donation here helps to pay our bandwidth costs, emergency funding, and other general costs that crop up from time to time. Thank you for your support!

    Click here for more information and to see all donations year to date.
×
×
  • Create New...