Intel cpu bug. Os patch might reduce perf up to 30%

[x-voodoo]

Originally Posted by x-voodoo

This patch is out already? I just tried checking Windows Updates and it says I'm up to date (last patch installed was 12/17)

Looks like Avast was preventing me from getting this latest security update. On my laptop I uninstalled Avast, restarted, checked for windows updates, and VOILA -- downloading KB4056892!

Now I am debating on whether or not I will reinstall Avast......

[howard stern]

Avast must not have created the compatibility registry key maybe they have not completed their testing yet.

If you are using the free version of Avast you can download Avira free they are compatible and very good.

[spike700]

Avast is compatible

Public spread sheet listing compatibility of many AV's as of Jan 5.

https://docs.google.com/spreadsheets...aring&sle=true

[bill dennison]

Originally Posted by howard stern

Samsung Magician results before and after the patch.

pre-patch
sequential (MB/s)
Read - 3,342
Write - 1,983

Random (IOPS)
Read - 285,876
Write - 230,443

Post patch
sequential (MB/s)
Read - 2,435
Write - 1,825

Random (IOPS)
Read - 148,129
Write - 131,629

I removed the patches and the performance stayed at post patch levels.
I have a Ryzen 1800X with a Samsung 960 pro running on pci-e gen 3 x 4

it's a virus
so now this intel sh*t is going to kill my ssd speed also on a amd 1800x and 1900x


I blocked updates for now

where do I sign up for the intel class action

[{WOR}lock]

Originally Posted by howard stern

Samsung Magician results before and after the patch.

...

I removed the patches and the performance stayed at post patch levels.
I have a Ryzen 1800X with a Samsung 960 pro running on pci-e gen 3 x 4

This is exactly why Microsoft shouldn't be patching AMD systems for Meltdown. They aren't affected by it, why should they take any kind of hit?

More to the point, if it's a blanket activation of the fix what happens when Intel eventually release a chip that doesn't have this security flaw?

Is this because it's a kernel level change that cannot change flags after compilation?

[howard stern]

Good question {WOR}lock I do not know the answer to that but it appears to me anyway once installed its does not reverse itself when removed. Like Bill said its a virus

[badsykes]

This is going too far ...

[@md_Guy]

https://twitter.com/m/status/949251143363899392
Effect on hypervisor..

Expect things to get worse when firmware updates go out

Sent from my SGP561 using Tapatalk

[badsykes]

Quick Facts about Meltdown and Spectre



https://www.hardocp.com/article/2018...ltdown_spectre

What to do with the information we have today:

Average Desktop User (Intel): At this point your best and only option is to apply the Microsoft KAISER patch when they become available. As this attack is also reported to have delivery via web-browser via .js, it may be plausible to block .js execution from the browser as well.

Average Desktop User (AMD): Hold tight. AMD is adamant that these exploits do not affect their architecture. If anything changes, we are actively tracking and will alert you.

The Gamer (Intel): Performance impact benchmarks have not been done yet, so we really have no idea how much of a performance hit your favorite video games will take. You could risk it and keep Windows from updating, but we would not recommend that currently.

I wish all Ryzen users Happy Camping ...

[Megaman]

Anyone who bought a 8700k will received Intels next generation chip for free as of the end of Q1 2018

Intel offers free chips to consumers

[badsykes]

Originally Posted by Megaman

Intel offers free chips to consumers

Thx for letting me try my vulnerability ... I clicked on your link without thinking ... The biological AV was bypassed flawlessly ... Maybe it's "meltdown-ing" and i didn't know ... I have to patch myself with "high brain usage" ?

[Megaman]

[Megaman]

NVE copped a serious beating. Regular SSDs not so much.

[kboye]

ESXi info on reddit
https://www.reddit.com/r/vmware/comm...n_and_spectre/

Seems both hypervisor and guest OS needs an update.

[x-voodoo]

So it seems those with NVE SSD's are the ones having performance issues, while others experience negligible performance inpacts.

Honestly I can't tell a difference post patch on my PC (I haven't run benchmarks, I just use it).

[bill dennison]

by a few things he says AMD's only vulnerability requires physical access to my system

if the hacker is in your house you have more problems than your PC .

so why rush to patch AMD now with the same generic windows fix

some of these class action lawsuits need to hit Microsoft


........

1800x



not bad but no ssd tests

[metroidfox]

Let me start off by saying that I'm a kernel developer. Let me also say that AMD's response is misleading, and kind of confusing. Meltdown seems to affect their processors less, but Spectre (the really bad one) is a big deal on their hardware too. In fact, it's a big deal on almost every out of order processor ever.

Do NOT avoid these patches.

[Treeckcold57]

Originally Posted by bill dennison

by a few things he says AMD's only vulnerability requires physical access to my system

if the hacker is in your house you have more problems than your PC .

so why rush to patch AMD now with the same generic windows fix

some of these class action lawsuits need to hit Microsoft


........

1800x


not bad but no ssd tests

Actually, I heard someone ran the powershell tool on his Ryzen and it's disabled on Ryzen even with a patch. I've rebenchmarked mine and didn't see any differences.

[bill dennison]

Originally Posted by Treeckcold57

Actually, I heard someone ran the powershell tool on his Ryzen and it's disabled on Ryzen even with a patch. I've rebenchmarked mine and didn't see any differences.

I've still blocked it for the windows 10 35 day max and will wait a week or two and see what comes out

[Treeckcold57]

Originally Posted by bill dennison

I've still blocked it for the windows 10 35 day max and will wait a week or two and see what comes out

I'm sure AMD will sort it out for next round.

[Arrowhead]

Originally Posted by x-voodoo

So it seems those with NVE SSD's are the ones having performance issues, while others experience negligible performance inpacts.

Honestly I can't tell a difference post patch on my PC (I haven't run benchmarks, I just use it).

****. Literally just bought a 960 pro.

[Skynet]

Originally Posted by bill dennison

I've still blocked it for the windows 10 35 day max and will wait a week or two and see what comes out

Did the same hopefully by then Microsoft will pull their head out of their ass. Running Threadripper BTW too lazy to keep updating my system specs.

[pax]

This dev reported big hits on intel and swtitched to amd..

https://www.reddit.com/r/Amd/comment...my_8700k_to_a/

After a long love/hate relationship with Intel over the last few years and being frustrated with with what’s happened this year and last, my software dev and personal home machine have switched over to AMD.

I originally just about 2 weeks ago built an 8700k version of the exact PC I’ve listed below, but after patch and the performance hits I took in Virtualization, compile time, NVMe speed.. I’m done with Intel. It was enough of a hit for me to call up my local store and get an exception/return period bump so I could actually swap to AyyMD.

[–]Amdestroyer94AMD 30 points 4 hours ago
How is the performance when compared with your previous system. Also what was the degree of performance degradation with the patch that made you switch

[–]HugeIRLRyzen 7 1700 | 1080Ti | 1TB NVMe SSD | Software Developer[S] 34 points 4 hours ago*
The Ryzen 7 1700 is a little slower performance wise but mainly due to clock speed. It still does everything I need and the extra cores/threads outweigh the clock loss for me (I plan on hitting 4GHz soon, right now I’m cozy at 3.8/1.350v).

For me it felt drastic. I saw compile times drop significantly especially with Java, my NVMe drive was a touch slower too leading to my large transfers suffering.

I felt I easily hit the upper numbers of degradation (30ish%) especially with the server tasks/virtualization I was doing.

[shadow001]

Originally Posted by Megaman

I wonder if the same test will be done with an AMD setup and the same NVME drives to see if there's the same ~30% performance loss, and i can imagine some being ticked off by the longer game load times, and online server operators absolutely fuming with the current fixes that are seriously increasing CPU load on their servers, hence the much longer log in times for end users.

[shadow001]

Just did an SSD test using my 960 pro NVME drive, running off a 16 core thread ripper setup using the ATTO benchmark, and read speeds at high queue depths and transferring big files still hits 2000 MB/sec writes and 2750 MB/sec reads, but i'll try with Samsung's magician software and post again.

[Skynet]

^^ nice you're doing the work for me.

Do these flaws actually go back to 1995 on Intel? If so that's uh...staggering.

[shadow001]

Tried samsungs own software and the results are hard to repeat with consistent results, where the best run was around 2000 MB/sec in both read and write performance, with the other runs slightly lower than that but I have no idea on queue depths or the size of the file used for the test which the ATTO benchmark allows, so i take magician for what it is (synthetic benchmark test ).



All windows updates are installed as of today, 2h48 local so that includes the "fix" for the worst case scenarios as mentioned in this thread already, and i didn't notice any increased load times on anything so far in the last few days....

[shadow001]

Originally Posted by Skynet

^^ nice you're doing the work for me.

Do these flaws actually go back to 1995 on Intel? If so that's uh...staggering.

Since the flaw is related to how Intel implemented speculative execution, basically the CPU trying to guess what the likely result will be before it has all the data available, given that data access even from main memory is already slow enough as it is, with recent systems having 4 memory channels using DDR4 3200 only now going over 100 GB/sec, never mind NVME SSD's on M2 slots being considerably slower ( like 30 x slower... ), that speculative execution becomes essential in making the best of a bad situation to be in.



It could go that far back, but one would hope that even users that aren't on the cutting edge of hardware still upgrade their setups every 5 years, though it sucks that even their latest and greatest Intel processors suffer from it, so the real fix comes from more than software and a bios update, and also includes a design change within the CPU.....Takes time to design that change and validate that it works properly with all software out there....A couple of years at least.

[bill dennison]

do or will AMD MB's require a new BIOS for this also ?

as I've seen nothing new for this for my asus c6h or zenith extreme

[shadow001]

Originally Posted by bill dennison

do or will AMD MB's require a new BIOS for this also ?

as I've seen nothing new for this for my asus c6h or zenith extreme

If the problem also presents itself, it's likely there will be new bios releases from all makers but as of right now, i've only been running with the software patch installed and nothing else.


How AMD designed the speculative execution function and/or out of order execution features of their chips is the difference in if a BIOS update will be needed to further mitigate the problem.....