flat assembler
Message board for the users of flat assembler.
 Home   FAQ   Search   Register 
 Profile   Log in to check your private messages   Log in 
flat assembler > Heap > Windows run slower after Meltdown/Spectre update?

Author
Thread Post new topic Reply to topic
TmX



Joined: 02 Mar 2006
Posts: 813
Location: Jakarta, Indonesia

Windows run slower after Meltdown/Spectre update?

For you Windows user, MS already provided some updates to mitigate those issues.
Now the question is, do you notice if there is performance issue after installing it?

I think my PC feels a bit slower, e.g:
- Running VS 2017 takes longer time
- Compiling 32-bit FreePascal from source takes almost 15 minutes
(the same codebase takes about 5 minutes to compile on another Windows machine pre-Meltdown patch).

Yes, I've disabled Windows Defender and 3rd party antivirus. No antispyware program is running.

Anyone? Confused
Post 24 Jan 2018, 07:04
View user's profile Send private message Reply with quote
revolution
When all else fails, read the source


Joined: 24 Aug 2004
Posts: 15733
Location: Crossing the Cauchy horizon

It has been expected that some applications will be slower. Reportedly up to 30% slower depending upon what you are doing.

It has also been recently reported that the patches are causing reboots, and failure to boot. So there is that problem also.

My suggestion if one finds that slowdowns and reboots are a problem is to not install the patch(es) at all. Instead disable JS and set your firewall to whitelist mode. And of course don't run arbitrary code that you can't be sure where it came from.

For my current system in front of me I haven't downloaded patches from MS since they started including the adware GWX and the spyware keylogger etc. So no slowdowns for me. Also no JS for me. And no unexpected exfiltration of data, or infiltration of unaudited code, from unwhitelisted programs.


Last edited by revolution on 25 Jan 2018, 04:44; edited 1 time in total
Post 24 Jan 2018, 07:49
View user's profile Send private message Visit poster's website Reply with quote
Tomasz Grysztar
Assembly Artist


Joined: 16 Jun 2003
Posts: 6820
Location: Kraków, Poland

fasm(g) seems unaffected, unsurprisingly - as it mostly does its job on its own, very rarely having to call the system.
Post 24 Jan 2018, 09:10
View user's profile Send private message Visit poster's website Reply with quote
revolution
When all else fails, read the source


Joined: 24 Aug 2004
Posts: 15733
Location: Crossing the Cauchy horizon

Intel asks customers to halt patching for chip bug, citing flaw

https://www.reuters.com/article/us-cyber-intel/intel-asks-customers-to-halt-patching-for-chip-bug-citing-flaw-idUSKBN1FB2M9 wrote:
Intel Executive Vice President Navin Shenoy disclosed the problem in a statement on the chipmaker's website, saying that patches released after months of development caused computers to reboot more often than normal and other "unpredictable" behavior. (intel.ly/2DsL9qz)

Linus Torvalds declares Intel fix for Meltdown/Spectre ‘COMPLETE AND UTTER GARBAGE’
Post 25 Jan 2018, 04:38
View user's profile Send private message Visit poster's website Reply with quote
Furs



Joined: 04 Mar 2016
Posts: 1133

I am so confused about the whole thing. What Intel fix? The meltdown fix in the kernel was approved by Linus long ago, what's this "new" fix for?

These vulnerabilities and their patches are so fucking confusing because every moron publishes crap about them without knowing ANYTHING technical, and it's hard to find an article or document with actual facts without being a little way over my head (i.e. assuming you're a security expert).

Also I never got an answer to the question about Spectre, looks like I'm not the only one in doubt. Why does the kernel need patches against Spectre? Can Spectre read kernel memory? Or where is the vulnerability? (this applies to Windows too, even)

(if it is because of the Linux BPF JIT thing, it's disabled by default and I've no intention to enable it, which means I don't need to patch Spectre -- IF, but I can't find a proper answer)

I do have kernel patched against Meltdown though, I assume this isn't the "complete and utter garbage" Linus was talking about, or what?
Post 25 Jan 2018, 20:10
View user's profile Send private message Reply with quote
revolution
When all else fails, read the source


Joined: 24 Aug 2004
Posts: 15733
Location: Crossing the Cauchy horizon

AFAICT Linus is talking about the hardware fix that is proposed. Apparently Intel is proposing to make the Spectre fix opt-in, selectable by the BIOS. Linus calls Intel to task showing that Intel's words "we place security first" are bullshit, and the opt-in shows Intel don't really care about security, but instead care about benchmark results first.

So I gather from that that the proposed hardware fix for Spectre kills performance so much that Intel backed off from enabling it by default: Thus it is "complete and utter garbage" apparently. Not sure, but that is my understanding of it.
Post 25 Jan 2018, 20:20
View user's profile Send private message Visit poster's website Reply with quote
rugxulo



Joined: 09 Aug 2005
Posts: 2279
Location: Usono (aka, USA)

Although I don't know if it's true, I heard that newer cpus with PCID won't suffer as much slowdown.


Wikipedia wrote:

Since the 2010 Westmere microarchitecture Intel 64 processors also support 12-bit process-context identifiers (PCIDs), which allow retaining TLB entries for multiple linear-address spaces, with only those that match the current PCID being used for address translation.



But I haven't done any benchmarking myself, so I have no idea.
Post 07 Feb 2018, 20:26
View user's profile Send private message Visit poster's website Reply with quote
Coty



Joined: 17 May 2010
Posts: 540
Location: ␀


rugxulo wrote:
Although I don't know if it's true, I heard that newer cpus with PCID won't suffer as much slowdown.



From what I've read, For windows you'll suffer the least if you have both:

PCID - Introduced on some Sandy bridge chips
INVPCID - Introduced on Haswell chips.

From what I've seen from my windows Desktop (Haswell based), I've had no major impact on performance and actually don't even notice. (Although It is a very stout chip to begin with). Personally I didn't take the time to bench mark but some people on the PCMR subredit did and was only about 3% slower after.
Post 20 Feb 2018, 04:06
View user's profile Send private message Send e-mail Visit poster's website AIM Address Reply with quote
Coty



Joined: 17 May 2010
Posts: 540
Location: ␀


rugxulo wrote:

Phoronix (FreeBSD Finally Gets Mitigated For Spectre & Meltdown) wrote:

There is Meltdown mitigation for Intel CPUs via a KPTI implementation similar to Linux, the Kernel Page Table Isolation. There is also a PCID (Process Context Identifier) optimization for Intel Westmere CPUs and newer, just as was also done on Linux



This was interesting, I wasn't aware Westmere supported PCID. So I did more digging and found this:

https://www.sqlskills.com/blogs/glenn/checking-your-intel-processor-features-regarding-the-meltdown-exploit/

Basis was:
Some Westmere (initial product line release July 2010) have PCID.
Most Sandy bridge (January 2011) have PCD
Haswell (June 2013) or newer will have PCID and INVPCID

I'm going to go out on a limb and say the Westmere chips that support it are probably XEON and or i7 based. And the sandy bridge chips are probably XEON, i7 and or i5 based. and also this:


SQLKills wrote:
It is possible to for a processor to support PCID and not support INVPCID. My understanding from the Microsoft Powershell code is that you need both to get Windows OS support for the PCID performance optimization.

Post 20 Feb 2018, 04:43
View user's profile Send private message Send e-mail Visit poster's website AIM Address Reply with quote
rugxulo



Joined: 09 Aug 2005
Posts: 2279
Location: Usono (aka, USA)


Phoronix (FreeBSD Finally Gets Mitigated For Spectre & Meltdown) wrote:

There is Meltdown mitigation for Intel CPUs via a KPTI implementation similar to Linux, the Kernel Page Table Isolation. There is also a PCID (Process Context Identifier) optimization for Intel Westmere CPUs and newer, just as was also done on Linux

Post 20 Feb 2018, 05:00
View user's profile Send private message Visit poster's website Reply with quote
revolution
When all else fails, read the source


Joined: 24 Aug 2004
Posts: 15733
Location: Crossing the Cauchy horizon

I love the timewarp that happened above in those two messages Smile Coty predicted the future.


Last edited by revolution on 21 Feb 2018, 00:26; edited 1 time in total
Post 20 Feb 2018, 08:06
View user's profile Send private message Visit poster's website Reply with quote
Coty



Joined: 17 May 2010
Posts: 540
Location: ␀

Time warps are weird man... The interesting thing is that the post is broken, both of my posts were made after his last one... o.O

maybe from the website hickup?
Post 20 Feb 2018, 17:43
View user's profile Send private message Send e-mail Visit poster's website AIM Address Reply with quote
revolution
When all else fails, read the source


Joined: 24 Aug 2004
Posts: 15733
Location: Crossing the Cauchy horizon

The CoLo lost a UPS and when the site came back about 30 hours later the clock was set in the future by about 8 hours. So when that was fixed we then got duplicate times. You might also notice that the post numbers are in sequence, just that the time went backwards, and the sorting for the forum is based upon the post time, not the sequence number.
Post 21 Feb 2018, 07:08
View user's profile Send private message Visit poster's website Reply with quote
Coty



Joined: 17 May 2010
Posts: 540
Location: ␀

That sounds like the UPS was very bad at it's job Laughing Or was this a total power failure?
Post 22 Feb 2018, 04:31
View user's profile Send private message Send e-mail Visit poster's website AIM Address Reply with quote
revolution
When all else fails, read the source


Joined: 24 Aug 2004
Posts: 15733
Location: Crossing the Cauchy horizon

Yeah, the UPS was interrupted. It became an IPS.
Post 22 Feb 2018, 04:40
View user's profile Send private message Visit poster's website Reply with quote
Coty



Joined: 17 May 2010
Posts: 540
Location: ␀

And here I thought IPS was a desirable trait (in monitors), Now I know the truth. Wink
Post 22 Feb 2018, 05:09
View user's profile Send private message Send e-mail Visit poster's website AIM Address Reply with quote
Display posts from previous:
Post new topic Reply to topic

Jump to:  


< Last Thread | Next Thread >

Forum Rules:
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
You can attach files in this forum
You can download files in this forum


Main index   Download   Documentation   Examples   Message board
Copyright © 2004-2018, Tomasz Grysztar.
Powered by rwasa.