Testing Windows 10 Performance Before and After the Meltdown Flaw Emergency Patch

[ZackL04]

Ryzen has been proven to not be affected by Meltdown at all.
As for Spectre, Variant One (Bounds Check Bypass) affects all CPUs but will receive a software patch with no real kernel overhead issues. Variant Two (Branch Target Injection) might affect AMD, but until now researchers haven't been able to do it (it was done on Intel and some ARM CPUs).

As far I understand, the reason AMD isn't that concerned about Spectre is because of how it works in relation to their architecture. From the white paper it seems that it exploits the cached data from a branch mispredict. That will rarely contain sensitive data and is highly dependent on the target app.

Hmm.. well this quote is straight from Techspot 1 day ago “So far Meltdown has only been proofed on Intel processors”

 

[Puiu]

I woke up today to a 32% reduction in performance after Amazon patched one of the AWS instances we use. This is severely impactful for IO Heavy processes which are commonplace in a microservice oriented organization.

My team and I will have to spend multiple workdays scaling our systems and reforecasting hosting costs due to this issue.

So no, it's not just synthetic tests. There are multiple real world scenarios that show crippling decreases in performance. I surely hope no onr ever has to wake up their car suddently grtting 30% less miles per gallon.

Can you tell us more about this? It is so very hard to get good server data because of the nature of the workloads being so different for everybody.

 

[Adobo]

Can you tell us more about this? It is so very hard to get good server data because of the nature of the workloads being so different for everybody.

This one case that had like 32% less performance was a Redis Nose that handles around 300 connections and receives about 7k request per second. The CPU usage peak spiked from 48% to 80% and then steadied at around 70%. Response times for the service that depends on this redis node went up from 4ms to 7.8ms.

We're still assessing the damage on our infrastructure since Amazon is slowly rolling this out for compute instances (servers) so that response time increase could be even higher once the service that depends on redis gets the patch too.

 

[danwat1234]

Very interested in potential performance decrease for core 2 duo (Penryn) CPUs as this exploitation to supposedly valid in CPUs that are 10 years old even. I will probably skip the update on my old Penryn laptop if the performance decrease is significant.

 

[noname]

Amazon AWS forums started flooding with complains of degraded performance. Just 1 example : https://forums.aws.amazon.com/thread.jspa?threadID=269858&tstart=0&tag=httpwwwtechsp-20

Also it seems the biggest performance impact will be after people update the bios : https://www.reddit.com/r/pcmasterrace/comments/7obokl/performance_impact_of_windows_patch_and_bios/

In the next weeks will have more info of performance degradation once everybody has also the latest bios.

I expect Techspot to redo all the benchmarks once their system have the BIOS with new CPU firmware available.

 

[Puiu]

This one case that had like 32% less performance was a Redis Nose that handles around 300 connections and receives about 7k request per second. The CPU usage peak spiked from 48% to 80% and then steadied at around 70%. Response times for the service that depends on this redis node went up from 4ms to 7.8ms.

We're still assessing the damage on our infrastructure since Amazon is slowly rolling this out for compute instances (servers) so that response time increase could be even higher once the service that depends on redis gets the patch too.

Please keep us informed. When you have time do some followup comments.

 

[noname]

I bet many people with Intel systems will still be left vulnerable even after applying windows update since they need to get bios update from their motherboard vendors.

https://newsroom.intel.com/news-releases/intel-issues-updates-protect-systems-security-exploits/
https://security-center.intel.com/advisory.aspx?intelid=INTEL-SA-00088&languageid=en-fr

And that wont happen since many motherboards are discontinued or simply lack any support if u think of oem prebuilt systems sold in supermarkets.

An example : Medion Erazer X67015 with i7-8700 and GeForce GTX 1070 sold by Aldi grocery store in Germany : https://www.computerbase.de/2017-12/medion-erazer-x67015-aldi-gaming-pc-test/
Will they issue a recall to update the bios ? Or they expect people who bought that with some bread and onions in same chart to do it ? : >

Do u think average Joe or parents, grandparents even know how to update the bios by themselves ?

That's why embargo was suppose to last till 9 january 2018 ( info found here : https://insights.ubuntu.com/2018/01/04/ubuntu-updates-for-the-meltdown-spectre-vulnerabilities/ ) so everybody would be out of holiday at work, CES 2018 news flooding the media , keynote of Intel CEO Brian Krzanich already done on 8 January 2018 ( https://www.ces.tech/Conference/Keynotes/Intel-Keynote/Brian-Krzanich ) so no funny question to him then. And in all this busy week benchmarks popping up showing no real performance decrease since no bios update. Nice PR but then again its also a filter for good tech websites and bad tech websites.

 

[Skrekkur]

Its quite strange that there is little or no difference. Make me think that the one of the following (or more ) must be true:
Windows patch has less impact than the linux one
The workloads here are not affected, or somehow in these benchmarks the fix isn´t properly applied.
(this link shows quite a rough degrade in compilation and database performance)
https://www.extremetech.com/computi...fix-intel-security-flaw-hits-performance-hard

Honestly I really hope this is correct and desktop workloads are relatively unchanged, but the effect on servers in the cloud are already pretty brutal and the total effects on that (including lag on servers and stuff) are yet to be seen.

p.s. my company applied the update on one of our clusters with 100% more cpu load post update
https://twitter.com/CCP_SnowedIn/status/948980181577875456

 

[Axle Grease]

The patch doesn't do anything until you update the BIOS to one containing the new Intel microcode.That will allow the patch to kick in. There's a big performance drop in random 4k read and writes, up to 30%. Everything else sees very little or no performance hit.

 

[noname]

Guess we will have to wait for BIOS updates for the motherboards used in the 8700k benchmark ( https://www.techspot.com/review/1497-intel-core-i7-8700k/ ) that is linked in this article

Asrock Fatal1ty X299 Professional Gaming i9 : https://www.asrock.com/mb/Intel/Fatal1ty X299 Professional Gaming i9/index.asp#BIOS
Asrock Z370 Taichi : https://www.asrock.com/MB/Intel/Z370 Taichi/index.asp#BIOS
Asrock Z270 Taichi : https://www.asrock.com/MB/Intel/Z270 Taichi/index.asp#BIOS

In the meantime Techspot should update this article with the output of the Get-SpeculationControlSettings powershell script that Microsoft is providing. The other tech sites already done this : https://pics.computerbase.de/8/1/1/6/0/6-1080.2265386108.png
In that screenshot its obvious that protection is not enabled and bios update is needed if only the windows patch is installed.

 

[Puiu]

@Skrekkur oh damn, anything but Eve Online! Intel please don't mess with my favorite MMO and its servers!

for those who want to test their PC and see if it still is vulnerable here's the powershell script and instructions:

1. Press the Windows key and type PowerShell.
2. Right click the PowerShell shortcut and select Run as Administrator.
3. Type Install-Module SpeculationControl and press Enter.
4. If you are prompted to install the NuGet provider, type Y and press Enter, and repeat if you are warned about installing from an untrusted repository.
5. With the installation complete, type Import-Module SpeculationControl and press Enter.
6. Type Get-SpeculationControlSettings and press Enter.

I ran into a Windows PowerShell execution policy issue when doing step 5. If that happens to you then run the following command: Set-ExecutionPolicy -ExecutionPolicy AllSigned
After you are finished with the test you can set it back to the default one using: Set-ExecutionPolicy -ExecutionPolicy Restricted

If you have the windows and bios patches the results should look like this:

Speculation control settings for CVE-2017-5715 [branch target injection]

Hardware support for branch target injection mitigation is present: True

Windows OS support for branch target injection mitigation is present: True

Windows OS support for branch target injection mitigation is enabled: True

Speculation control settings for CVE-2017-5754 [rogue data cache load]

Hardware requires kernel VA shadowing: True

Windows OS support for kernel VA shadow is present: True

Windows OS support for kernel VA shadow is enabled: True

Windows OS support for PCID optimization is enabled: True

 

[Charles Olson]

Some graphs are difficult to understand as it's not clear if they are measuring seconds or throughput. Also "Measured in seconds, Higher is better" appears on a few graphs - I'm pretty sure that's not the case
But good to see the overall message is that casual users / gamers are ok. Enterprise users might be concerned though.
Not that I care either way as I'm on Ryzen now.

It still might effect you if you're a gamer and the game you want to play is hosted on Intel severs and then the 30 or 35 percent difference might be noticeable to you the end user !

 

[AQUATARKUS]

Ive noticed that whenever u do tests, u dont test music production software, I use cubase and my projects often push the hardware to its limits, so anything that causes a drop in performance is most unwelcome.

 

[Mike M]

Well I installed the patch yesterday and the only drama seen on both my desktop (i7-7700k) and laptop (i7-6600U) was a few minutes at 100% and temperature spikes (75C for desktop and 100C on laptop). CPU usage eventually came back down to normal, as did temps. I'm not noticing anything different performance wise. If it weren't for all the media attention most people wouldn't notice any difference between this Windows patch and any prior...that also affect performance. As for the industry (Intel, AMD, Apple, ARM, Qualcomm, etc.) making any performance promises, I've only seen it for specs when chip is working normally...clock speed, temp threshold, cores/threads, cache, etc. It is Internet and magazine articles that are setting expectations that people are now trying to argue are manufacturer promises. There will always be vulnerabilities (including many unknown...or many that shouldn't be published until fixes are in place) and patches that may or may not impact performance in some way or another.

 

[Julio Franco]

In the meantime TechSpot should update this article with the output of the Get-SpeculationControlSettings powershell script that Microsoft is providing. The other tech sites already done this : https://pics.computerbase.de/8/1/1/6/0/6-1080.2265386108.png
In that screenshot its obvious that protection is not enabled and bios update is needed if only the windows patch is installed.

We did this hours ago, not in this performance article but in the "what you need to know":
https://www.techspot.com/news/72576-massive-cpu-flaws-get-name-meltdown-spectre-what.html

Also here's our article update from earlier this evening:

When we tested and published this article, the Windows 10 emergency fix had been out for just a few hours and no microcode or firmware updates had become available. For laptops and brand name desktops, that means you also have to update your system with new firmware that will come from your respective OEM. For builders, that means you have to wait for an update from your motherboard manufacturer.

As of writing, a handful of companies have started to offer these updates on some of their product lines: Dell, Lenovo/ThinkPads, Asus, Microsoft Surface, and some others. When our test platform receives a security firmware update we plan to add those results to this article. Users around the web have started to post benchmarks of their systems with both patches applied and the result appears to be slightly worse performance than when they had only applied the OS patch.

It's our understanding that the Windows patch and other OS-level updates cover only one of three known vectors for exploiting the flaws. The combination of OS and firmware should cover all three, although in some scenarios application-level patches will also be required to be fully secure. To have a better understanding of Meltdown and Spectre we suggest you read our condensed what you need to know or read the whitepapers.

 

[James00007]

It was pointed out by another tech channel that there is a chance that the upcoming patch could be reverse engineered by hackers. So performance drop is the least to worry about if that would happen. It's Y2K all over again.

Yea and to take that a step further and get really tin foil hat, there really isn't a better way for a global government spying network to get its hooks into nearly every system on the planet than this. This kind of thing is alarm bells galore in so many ways.

 

[sharky777]

It wasn't really fear-mongering, just a reaction to early benchmarks on server applications. You can see this in the random 4k reads in this article, which is an important metric in many server workloads. We've know from the start that regular desktop applications would see at most a 5% reduction and that games should have no issues.
For example PostgreSQL seems to be hit quite hard by this patch. Older chips are also hit harder than the newer Kaby/Coffee Lake.

We just need to wait for Steve to finish the rest of the tests, although as he said he won't do server stuff since it is outside of his expertise.

The way articles I saw were titled and other things were done poorly then (not saying TS did). A lot didn't really mention servers up front, and definitely would lead less tech-savvy people into believing it would significantly affect them.

Now I get to look forward to a minimal 30% slowdown.

Earlier this week, some security researchers said any fix -- which would need to be handled by software -- could slow down computer systems, possibly by 30 percent or more.

The good news is that there is a fix. But unfortunately the security patch will slow down processing power by as much as 30 per cent, hitting personal computers and cloud services around the world.

But a few of the quotes I found off of FB news (yes FB news, because I get to see how stupid the uninformed reactions are lol)...

I wonder, if there is going to be a lawsuit against Intel?
Nearly in all cases nowadays, someone ends up being suited.

There are already lawsuits that have been filed class action, there it is for you to read, calif for one,
Class action lawsuits have been filed in California, Indiana, and
these case should be interesting to say the least, you may need to copy and paste, seems tech-spot doesn't do much with the URLs .cept allow you to post them here, they don't seem to make them clickable, but I may be wrong, wont be the first time!

Oregon.https://arstechnica.com/gadgets/201...ts-regarding-meltdown-and-spectre/?comments=1

 

[noname]

Epicgames Epic Services & Stability Update :
https://www.epicgames.com/fortnite/forums/news/announcements/132642-epic-services-stability-update

That's a huge impact on performance looks like their cpu usage tripled from 20% to 60% after patch.

 

[Julio Franco]

Update (1/7): Following up to our initial testing, we're looking deeper into the matter by testing a patched desktop system for both Meltdown (OS-level patch) and Spectre (firmware/BIOS update):

https://www.techspot.com/article/1556-meltdown-and-spectre-cpu-performance-windows/

 

[mojoba]

Glad to hear the world's most powerful GPU is not effected. Maybe it's because there is so much spare power at 1080p? My 980ti is badly effected at 1440p on very high settings on Assassin's Creed Origins. I have a i7 4790k (4th gen) running at up to 90% usage. Even Lego Batman 3 has stuttering now at 1440p! These are the only games I've played so far. I have blocked Windows Update in Services and uninstalled this "emergency patch" and all is well again. I'd rather be hacked than have to go back to 1080p gaming.

Maybe try benchmarks where the GPU and CPU are being pushed to its limits? For a 1080ti try comparing Assassin's Creed Origins at 4k. Get it running on a stable benchmark pre patch at the maximum graphics settings you can have whilst getting a steady and smooth FPS, then install the Windows 10 patch and try again. Any stuttering where there was non before?

 

[Sentinal]

I was wondering if I was better off losing performance with an AMD laptop with rx gpu or to try my luck with a normal intel i5 with the usual 1060 and see if I loose to much performance