Microsoft Details Performance Impact of Spectre and Meltdown Mitigations on Windows Systems (microsoft.com) 237
Microsoft's Windows chief Terry Myerson on Tuesday outlined how Spectre and Meltdown firmware updates may affect PC performance. From a blog post: With Windows 10 on newer silicon (2016-era PCs with Skylake, Kabylake or newer CPU), benchmarks show single-digit slowdowns, but we don't expect most users to notice a change because these percentages are reflected in milliseconds.
With Windows 10 on older silicon (2015-era PCs with Haswell or older CPU), some benchmarks show more significant slowdowns, and we expect that some users will notice a decrease in system performance. With Windows 8 and Windows 7 on older silicon (2015-era PCs with Haswell or older CPU), we expect most users to notice a decrease in system performance.
Windows Server on any silicon, especially in any IO-intensive application, shows a more significant performance impact when you enable the mitigations to isolate untrusted code within a Windows Server instance. This is why you want to be careful to evaluate the risk of untrusted code for each Windows Server instance, and balance the security versus performance tradeoff for your environment.
For context, on newer CPUs such as on Skylake and beyond, Intel has refined the instructions used to disable branch speculation to be more specific to indirect branches, reducing the overall performance penalty of the Spectre mitigation. Older versions of Windows have a larger performance impact because Windows 7 and Windows 8 have more user-kernel transitions because of legacy design decisions, such as all font rendering taking place in the kernel.
With Windows 10 on older silicon (2015-era PCs with Haswell or older CPU), some benchmarks show more significant slowdowns, and we expect that some users will notice a decrease in system performance. With Windows 8 and Windows 7 on older silicon (2015-era PCs with Haswell or older CPU), we expect most users to notice a decrease in system performance.
Windows Server on any silicon, especially in any IO-intensive application, shows a more significant performance impact when you enable the mitigations to isolate untrusted code within a Windows Server instance. This is why you want to be careful to evaluate the risk of untrusted code for each Windows Server instance, and balance the security versus performance tradeoff for your environment.
For context, on newer CPUs such as on Skylake and beyond, Intel has refined the instructions used to disable branch speculation to be more specific to indirect branches, reducing the overall performance penalty of the Spectre mitigation. Older versions of Windows have a larger performance impact because Windows 7 and Windows 8 have more user-kernel transitions because of legacy design decisions, such as all font rendering taking place in the kernel.
Planned Obsolescence (Score:5, Interesting)
Re: (Score:2)
Re: (Score:3)
Except Intel changes the socket format so frequently, I've almost always been forced to by a new motherboard... and new memory.
Re: (Score:2)
Doesn't always work if you have a laptop (many CPUs are soldered in these days), and it does involve some cost on your own, especially if you don't do the work yourself.
Re: (Score:2)
That also assumes you can upgrade your CPU. My main computer runs an i7-3770K which I believe is the fastest CPU ever offered on Socket 1155. The other computer is also Socket 1155 and runs an i7-2600K. While that could be upgraded, the 2600K isn't really that much slower than the 3770K. I suppose I could overclock them as they are unlocked, though I haven't bothered because it hasn't seemed worth it.
Re: (Score:2)
Indeed...translation; "Intel fucked up, and we fucked it up even more in our crappy design decisions, but we are both going to make huge profits from selling everyone new hardware and software, so all is good."
Re:Planned Obsolescence (Score:5, Funny)
Yes, that's older. It's how time works. Look into it when you have some.
Re: (Score:2)
It took Intel 2 CPU generations to equal that single core speed (with only the 6700k in Aug 15), then the 7700k finally beat it (just) @ 4.2GHz a year ago - Jan '17.
The 7740X @ 4.3 followed in June '17 (and pulls 25% extra power to get that less than 10% increase...).
The 8 series tops out at 3.7Ghz.
They are all still running the 4 cores.
Sure it's "older". It's certainly no-where near obsolete.
Re: (Score:3)
What's a GHz? Is that some irrelevant performance metric from the 90s?
In other news The Pentium 4 had more GHz than the competition. Ever wondered why it was laughed at?
Re: (Score:2)
Well, now that raw processor speed is suddenly more important, the GHz numbers could start being relevant again.
Re: (Score:2)
My son's computer is a C2Q 9500 with Windows 10. There's no specific release date published on Intel's ARK, but C2Q 9550 released in Q1 2008 (10 years ago!), so it's at least that old. He plays Overwatch on it in Ultra Quality at a constant 60 FPS (limited by his display). That's a 10 year old computer (with a newer GPU) running games at full speed that you're "supposed" to buy multi-thousand dollar computer setup for.
Re: (Score:2)
The decrease is performance only happens if you manage to update the BIOS and there's fat chance of doing that on a Haswell motherboard! It'll have reached non-support EOL long ago.
My Haswell laptop's BIOS & FW was updated this morning. This gave me a good excuse to get more coffee while not worrying about the paranoid fantasies of those that live in a world of imagined facts.
TFA is quite informative... (Score:2)
Kudos to Microsoft and Terry Myerson - great article with excellent details...
as long as you're running an Intel processor. Inquiring minds want to know where in the performance-hit list a Ryzen shows up. Does it have the " refined...instructions used to disable branch speculation to be more specific to indirect branches, reducing the overall performance penalty of the Spectre mitigation."?
Re: (Score:2)
Kudos to Microsoft and Terry Myerson - great article with excellent details...
It's a well laid out article, but it's a shame they've shied away from posting any actual numbers, leaving us to guess what kind of impact they're talking about, and for everyone to have to run their own benchmarks. They could have said something like "typically xx-xx%, with workloads having heavy I/O most affected".
Time to upgrade (Score:3, Interesting)
Folks, your CPUs 2015 and older are obsolete ... how convenient for Wintel.. get that fancy new CPU and another copy of Windows 10, because what you have today is OEM version, so with a new CPU you'll need a new Windows copy.
Re: (Score:2)
"Obsolete"? My home PC just blue screened from outrage at such a statement. Its Deneb processor from 2009 is just fine, thank you, and even ran DOOM 2016 acceptably (well, it did after I stole my son's video card).
Methinks your definition of Obsolete is different than mine...
Re: (Score:2)
Not *my* definition of obsolete, just a small dose of sarcasm. It is definition that intel + Microsoft are going to be pushing.
Re: (Score:2)
you're right; I RTFA'ed, but didn't fully RTFC. I stopped after the first phrase. I hang my head in shame...
Re: (Score:2)
Re: (Score:2)
Unfortunately, that font will be rendered in kernel space, because there have never been any security vulnerabilities in fonts... [threatpost.com]
Who the hell decided that was a good idea?
Re: (Score:2)
You know they fixed that and now fonts are no longer rendered in the kernel. Or you would know that, if you RTFA.
Re: (Score:2)
I did know that. However, they still made the decision to do that at one point, and they didn't reverse that until the somewhat recent past.
So the question stands: Who the hell decided that was a good idea?
Re: (Score:2)
You can't talk about obsolete on Slashdot or you're inundated with people running Pentium I systems, ancient Nokia phones, etc. It's an audience that takes pride in wringing the last bits of life out of old hardware.
Re: (Score:2)
And your point is????
Re: (Score:2)
I bet your fun at partys.
I bet you're a hoot at grammar conferences.
Re: (Score:2)
Folks, your CPUs 2015 and older are obsolete ... how convenient for Wintel.. get that fancy new CPU and another copy of Windows 10, because what you have today is OEM version, so with a new CPU you'll need a new Windows copy.
I am on windows 7 using a 2011 made i5-2500K and after upgrading GPU and switching to SSD the machines performance is still good enough to run things like PlayerUnknowns Battlegrounds on Ultra setting(GTA 5 is a bit choppy on ultra but runs buttery smooth on high).. And I haven't even overclocked the CPU yet.
I am yet to install the patches, but if the performance degrades beyond tolerable I'll first crank the OC to max, maybe try to find a deal on a faster socket 1155 processor as there are couple..
B
Re: (Score:2)
Spectre is actually applicable to Intel, AMD, and the various ARMs (Samsung, Qualcomm, ...)
You're thinking of Meltdown, that's Intel specific.
Free speed upgrades for Appdoze 10! (Score:5, Funny)
Re: (Score:2)
Re: (Score:2)
I can't believe you're donning your tinfoil hat in response to a bug that Microsoft had no control over.
When Microsoft released Windows NT 4.0, which as compared to NT 3.51 merged the Kernel and GDI memory spaces in the name of graphics performance, many of us complained. "Leave that kind of insecure cockamamie bullshit in the kiddie desktop OS, and leave our Server and Workstation OS with some privilege separation," we cried. But all unheeding, Microsoft proceeded apace down the path of compromising security in the name of gaming performance (literally!) and gee, look where we are now:
Re: (Score:3)
Dude, they agreed with this conclusion and moved font rendering out of the kernel.
I can't imagine being your coworker if every time someone admits they made a mistake and correct it, you harp about how wrong they were in the first place. We get that you are Very Smart and were in fact right along along, being ungracious about it is not making you seem smarter, it's making you seem just as smart but more of a jerk.
Re: (Score:2, Insightful)
Or better yet, don't run EOL software.
Retard alert! Win7 has more than two years before EOL.
Complexity unfortunately means Holes. (Score:2)
I don't think this is really that surprising. Modern CPU's are almost like mini-computers in themselves breaking down and reorganizing code on the fly internally. Fixing them means ugly workarounds which will usually cost a bit in performance. As time goes on, expect more of these issues rather than less. It's why most modern CPUs have a BIOS loaded table that makes workaround fixes in hardware although this problem is probably too big to fix there.
Re: (Score:2)
Re: (Score:2)
You could make it fully compatible with x86, no problem. That's not the issue at all.
The cost would be much more than 10% of performance. OOO and speculative execution would probably cost at least 50% of the performance (and, for mobile, this is perf/watt, which translates directly into usable battery life).
Think about it this way: the CPU hits a plain old branch. The branch has a condition variable that's in L3 or main memory. In your "brute force" model, the processor sits idle for 200-300 instructions wh
Re:Complexity unfortunately means Holes. (Score:5, Funny)
Genius!
Sounds like (Score:2)
Re: (Score:2)
Want to make it sound more impressive? Microsoft currently has 45 supported variants of Windows. They shipped patches for 41 of those versions.
Of course, it's crazy to support so many different variants. At the same time it's crazy to support Windows 7 for years after 10 comes out, but people will complain mightily if you EOL it and don't provide security patches.
And it's even more crazy that none of this was Microsoft's fault to begin with.
How to avoid the patch? (Score:2)
I've got a Windows machine just for playing games. I don't have any sensitive information in the machine, nor do I really care if I have to reinstall at any point (slightly annoying). So, how can I say to Microsoft "Thanks for looking out for me, but I'd rather the extra performance"?
Note that the machine on which I game is from 2015 so the "fix" would have a noticeable slowdown. I've already turned off automatic updates, but this would likely be classes as an emergency update which ignores the settings.
Re: (Score:2)
The thing about all this is: Why should I care?
Sure, this is a potentially big problem for people who run virtual server farms, but for my Windows game box (which runs Windows 7, and has had updates turned off since all that "gwx" bullshit), what is the point? Even if I did get it in malware (how? I don't do general web browsing on that computer, just one or two specific games, not even Steam, it's behind NAT, and I also disable a lot of useless services), they would do better to force ads on me... oh wait
Re: (Score:2)
Thanks.
I also saw another post that says to disable the update service, dosvc, and bits. So I've done that as well. A quick test showed that my games still work - so I might be safe.
Notice a change (Score:2)
we don't expect most users to notice a change because these percentages are reflected in milliseconds.
Most users don't even notice that percentages are dimensionless, so it makes no sense to translate them to milliseconds.Yes, I understand that he pretends to mean that the penalty for each instance of the problem only causes a delay of milliseconds, but still, the performance drop is there and many consecutive penalties aggregate into a noticeable slowdown.
Apart from this joke, this is a good move from MS.
I haven't noticed any slowdowns (Score:5, Informative)
I am running a Skylake Core i5-6500 with 8GB RAM and a 250GB Samsung SSD with Windows 7 Home Premium and I have the patch installed and haven't observed any slowdowns.
Just kicked of a full compile (in VS 2017) of a large (~2100 files) project I have here and I saw no noticeable slowdowns compared to how fast the thing compiled before the patch. And such a thing would be highly I/O bound (reading all the input source files and things, writing out compiled obj and other files, reading toolchain binaries etc) and likely making a lot of kernel-user transitions.
I have no games on here that are demanding enough to show any observable difference between old and new so I cant test those.
Re:Really? (Score:5, Insightful)
What — exactly — do you expect Microsoft to do? They didn't make the hardware and they can't fix this in software. I believe the fact that they're offering the choice (on Server at least) of suffering the performance impact or not shows an extraordinary degree of pragmatism. I honestly can't imagine a better response; how would you have them change it?
Re:Really? (Score:5, Funny)
Actually, this is a software (more precisely, a compiler) problem. The problem is that the modern compilers
make it difficult for these modern CPUs since they don't clearly instruct the CPU on how to proceed.
So the CPU has to speculate (a.k.a. guess) what instruction it should do next. If the compilers produced
better code in a more organized fashion, then the CPUs wouldn't have to be guessing all of the time, amiright?
(Hey, you know this is the next step in the blame game - watch somebody make a serious thread of this.)
CAP === 'spells'
Re: (Score:3)
Hey, you know this is the next step in the blame game - watch somebody make a serious thread of this.
Too late; Itanium engineers did this two decades ago. ;)
Re: (Score:3)
Actually, this is a software (more precisely, a compiler) problem. The problem is that the modern compilers
make it difficult for these modern CPUs since they don't clearly instruct the CPU on how to proceed.
So the CPU has to speculate (a.k.a. guess) what instruction it should do next. If the compilers produced
better code in a more organized fashion, then the CPUs wouldn't have to be guessing all of the time, amiright?
(Hey, you know this is the next step in the blame game - watch somebody make a serious thread of this.)
CAP === 'spells'
Actually, no. You're NOT right.
Speculative branch prediction happens because there is NO WAY for ANYONE to know whether a Branch based on DATA will have to be taken or not. Otherwise, there wouldn't have to be CONDITIONAL branches AT ALL!
I see software development is NOT your forte, amirite? ;-)
Re: (Score:2)
I just submitted a patch to gcc so it makes more predictable code.
The patch:
if you write x=x+100;
This will now translate into:
for (int i=0; i<100; i++) {
x++;
}
It will then be easier for the CPU to guess the next instruction is x++
Alternatively, one may use the brainfuck programming language:
https://en.wikipedia.org/wiki/... [wikipedia.org]
Re: (Score:3)
Re: (Score:2)
I expect them to give me a choice to install this patch or not.
I want a highly-visible ON/OFF switch in Control Panel.
For most servers it should default to OFF.
Re: (Score:2)
I expect them to give me a choice to install this patch or not.
Yep. "We expect most users to notice a decrease in system performance" is corporate-speak for "you're fucked if you install this".
Most users wouldn't notice a slowdown until after the 30th Internet explorer toolbar.
Re: (Score:2)
They didn't make the hardware and they can't fix this in software.
I imagine Intel actually can fix this with a chip firm/software update, but won't because they don't want the responsibility of dealing with anything that goes wrong across the many, many varied CPUs affected. (Just a very cynical guess though.)
Re: (Score:2)
The fix requires permanent disabling of performance enhancing functionality, and is required a lot faster than BIOS updates could bring - hence the OS inclusions.
It will take over a year before Intel has new silicon out that does not have this flaw baked into their CPUs.
Re: (Score:2)
Makes a lot of sense to me. Imagine a classified or close network. Those often require physical access, and often those ports are managed, which means you need to be logged in, so it's not even an issue if it isn't internet accessible.
Re: (Score:2)
No, Microsoft is the victim of this bug, not the perpetrator. They just try to mitigate the impact. The fact that they give customers this option implies that the performance impact is huge.
Re: (Score:2)
8.1 was worse than 10 for sure. I could get if you were defending 7, as that was a sweet spot.
Re: (Score:2)
Windows 8.1 is actually pretty good in some aspects. Noticeably faster and more stable than Windows 7 (and Windows 10 for that matter). It's also the last version of Windows without all the forced telemetry and forced updates. That makes it really the last version of Windows you can use without handing control over to Microsoft.
Of course, the huge flaw is the UI. Though there are some fixes for that (Classic Shell), and I never really minded it that much, though I agree it is still a step back.
The other
Re: (Score:2)
Yes, but Windows 8.1 is Windows 7 + rearranging the UI so you cant find anything + attempting to force you to use lame walled Garden apps ( aka Metro ) + lame home page "applets" ( aka Metro ) sucking all kinds of CPU and memory to tell you all kinds of things you cant see at the moment and really dont care that much about probably.
Windows 7 is infinitely superior to Windows 8.1
Re: (Score:3)
WTF is any operating system doing rendering fonts in ring 0? This was one of the reasons for the MS and IBM divorce, MS wanting to run stuff (video drivers) in ring 0 for performance and IBM refusing for security.
Re: (Score:3)
It's pretty fucking obvious that MS are trying to take advantage of the situation to get more users to finally switch over to Spyware 10 by fearmongering.
Not at all.
Since Vista, Microsoft has been on a very deliberate path of reducing dependencies, streamlining interfaces, and stripping legacy cruft within Windows. They have been modularizing the Windows kernel and kicking more functionality into user space for over a decade. This isn't news. I'm sure they're glad that it finally paid off though.
You are conflating the architectural improvements with the intrusive telemetry. They are two entirely different issues. In fact, I wouldn't be surprised if the peopl
Re: (Score:2)
""It's pretty fucking obvious that MS are trying to take advantage of the situation to get more users to finally switch over to Spyware 10 by fearmongering.""
"You are conflating the architectural improvements with the intrusive telemetry. They are two entirely different issues. In fact, I wouldn't be surprised if the people who argued in favor of the current kernel design turned around and argued against the always-on telemetry. "
Where does the poster conflate them? The post seemed to be complaining about
Re: (Score:2)
We have lived with this "flaw" for more than two decades without a single person noticing it or being affected by it.
Sure, but in 2018 every bad person in the world suddenly knows about it (and how to use it).
Re: (Score:2)
So adorable!
Re: (Score:2)
Re: (Score:3)
*Some* AMDs are bricked, older chips.
Re: (Score:2)
AMDs are bricked, didn't you read the earlier news?
Except no PCs were bricked. An unspecified subset of computers needed the OS reinstalled.
Re:How is 2015 old? (Score:5, Interesting)
Re: (Score:2)
I know! I'm with you. My 2 year old still needs diapers. As for my i7 quad-core from 2010 I might need to rethink upgrading... or putting that Registry key and preventing the update from downloading. It already is slow because it has a spinning HD.
But now I'm mad because "I have to" upgrade and pay them money for their mistake. However, proper CPUs don't exist yet. So I want to wait. And suffer in the meantime?
Grumble grumble.
Re: (Score:2)
I know! I'm with you. My 2 year old still needs diapers. As for my i7 quad-core from 2010 I might need to rethink upgrading... or putting that Registry key and preventing the update from downloading. It already is slow because it has a spinning HD.
But now I'm mad because "I have to" upgrade and pay them money for their mistake. However, proper CPUs don't exist yet. So I want to wait. And suffer in the meantime?
Grumble grumble.
That mechanical drive of yours is a much bigger bottleneck than any of these patches will be if your system is otherwise well equipped. Also try overclocking before buying new anything.
Re: (Score:2)
I am sorry but how is 2015 called old? Most of 2015 was barely 2 years ago.
" older" is the word they used, not "old". Are you of the belief that something from 2015 is not older than something from 2017?
Re: (Score:2)
" older" is the word they used, not "old". Are you of the belief that something from 2015 is not older than something from 2017?
Of course not, it's just less new.
Re: (Score:2)
I have an "older" Broadwell 14 Core Xeon and that thing kicks the ass of 99% of any "newer" chips. Performance drop or not.
Re: (Score:2)
Actually, most of 2015 is closer to 3 years ago.
Re: (Score:2)
No mention of AMD but :
From TFA:
In general, our experience is that Variant 1 and Variant 3 mitigations have minimal performance impact, while Variant 2 remediation, including OS and microcode, has a performance impact.
Spectre is Variant 1 and 2. Meltdown is Variant 3.
AMD original response is that there is a "near zero" risk of exploit for variant 2 and a "zero" risk for variant 3. Notice the difference.
And from a link page on the Microsoft website:
Microsoft is aware of a new publicly disclosed class of vulnerabilities that are called “speculative execution side-channel attacks” that affect many modern processors and operating systems, including Intel, AMD, and ARM.
So it is very likely that performance impacting mitigations are in place for AMD CPUs too. In fact, they are currently working with AMD on the issue, so it is probably too early to tell.
Re: (Score:2)
AMD original response is that there is a "near zero" risk of exploit for variant 2 and a "zero" risk for variant 3
Claims about "near zero" or "zero" risk of exploit are traditional CYA effluent, after all...
Even AMD has claimed (for a different CPU errata) that problems "would not occur during normal desktop usage and we've never encountered it during our testing." The claim was funny, as all you had to do was fire up a VM... or the right benchmark... It didn't speak well to AMD's testing at the time.
Re:Sandy Bridge processors info (Score:5, Informative)
Your performance will be fucked six ways to Sunday if your workload does a lot of user to kernel mode switches. Unless you fancy waiting for Intel to release fixed chips and then buying a new CPU and a new motherboard to put it in.
Ironically Intel forcing people to buy new motherboards to switch between very similar CPU generations coupled with the fact that any Intel CPU you buy now still has the bug means that its just as easy to buy an AMD CPU and motherboard than an Intel one.
Then again a new Intel chip has Process Context ID support which means the workaround for the bug is relatively low impact.
Re: (Score:3)
Not worth supporting is more like it. It may sound like this is going to spell losses for Intel, and I'll grant that they've seen a momentary dip in stock prices, but you can bet good money that this will ultimately result in a rush on new Intel hardware to replace "bad" hardware. And people will just throw piles of cash at them, because reasons.
There is no new Intel hardware available for sale to replace "bad" hardware. There *IS* however new AMD hardware.
Re: (Score:2)
Not worth supporting is more like it. It may sound like this is going to spell losses for Intel, and I'll grant that they've seen a momentary dip in stock prices, but you can bet good money that this will ultimately result in a rush on new Intel hardware to replace "bad" hardware. And people will just throw piles of cash at them, because reasons.
Give it another week or two to settle, then buy Intel stock. Watch and see.
I am still on 2011 made i5-2500K and and with upgraded GPU new games still run at highest settings; If I have to upgrade both mobo and memory because of these shenanigans my next machine won't be an intel system, thats for certain
Still running at stock speeds thought, will try overclocking before buying anything new.
Re: (Score:2)
AFAIK, most of the issues people have with Windows 10 are about policies and the GUI. And on these aspects, is is reasonable to consider Windows 7 superior.
On the core technical aspects however, most people seem to agree that Windows 10, and even 8 are superior to 7.
Re:Honest reasons why 'Windows 7' isn't good enoug (Score:5, Informative)
Windows 10 sucks (quantify that)
1. Keyloggers that send your keystrokes to Microsoft to serve targets ads in the OS. *cough* I mean 'Telemetry'.
2. Inability to control patches / updates
3. Reduced control over a system I *own*
There's a good start.
Re: (Score:2)
>>> 3. Reduced control over a system I *own*
I find this the most fascinating part of modern computing - "Here, pay me $700 for this smartphone so I can see everything you do, know everyone you know, and sell that information to anyone who'll pay me a dime for it. And, no, you're not allowed to stop me, or to even find out what information I'm collecting". Windows 10 moved that smartphone concept down to the PC.
This whole concept of paying large amounts of money for something that isn't yours (bec
Re: (Score:2)
And you let all these things happen.
Android phones can be re-imaged. Nobody forced anyone to use quicken or any software that uploaded everything to the could. I drive older cars not because I am broke but for how solid they are built, and they don't report shit.
If it's new, 'Free' has 'consequences'. For now, it's still my choice.
Re: (Score:2)
That used to be the case, but it's completely changed now. Now you pay big money, and you are still the product being sold. In fact there's absolutely no way to not be the product being sold with most new products these days. I miss the days when you could say that only the free services were selling you as the product, now you have to assume that every service does that. No matter how expensive.
Re:Honest reasons why 'Windows 7' isn't good enoug (Score:5, Insightful)
Re: (Score:2)
Windows 8 was preschool children - it may as well have had the "tiles" outlined in crayon.
Windows 10 is at least 3rd grade UI.
Re: (Score:2)
Stick with it just stop pretending to be experts, telling others Windows 10 sucks (quantify that) and encouraging NOVICES who would benefit from architecture improvements to stay away from it because YOU don't like it.
Windows 10 fails to run quite a bit of legacy software developed for windows platform that functions just fine on windows 7 and earlier, this includes some very simple software using nothing but windows libraries.
Windows 10 fails to routinely respect user choice in things like allowed update install and reboot times.
On windows 10 the user is not in control of the machine even if logged in as administrator.
If you look on Nvidia forums theres a +100 page complaint thread about performance issues appe
Re: (Score:2)
Windows 10 fails to run quite a bit of legacy software developed for windows platform that functions just fine on windows 7 and earlier, this includes some very simple software using nothing but windows libraries.
People complained about this a lot going from XP to Vista/7. The culprit, in most cases, was a better security model for the operating system which broke old applications. I am perfectly fine with this kind of change.
Windows 10 fails to routinely respect user choice in things like allowed update install and reboot times.
This is largely resolved in v1709. You set your active hours, and you can override it when prompted if you're outside the active hours.
Or, for an unsanctioned fix, disable the Windows Update service until you want to install updates.
If you look on Nvidia forums theres a +100 page complaint thread about performance issues appearing over 2017 windows updates
There were all kinds of performance issues with XP, Vista, and
Re: (Score:2)
This is largely resolved in v1709. You set your active hours, and you can override it when prompted if you're outside the active hours.
Or, for an unsanctioned fix, disable the Windows Update service until you want to install updates.
I use windows 10 on my work laptop and in the up to date version I have installed on the option to change the active hours have been completely removed, futhermore microsofts own update prevention tool that lets you select individual updates doesn't work, at all, in my experience; If I let microsoft update the graphics drivers it completely breaks any gaming performance whatsoever, it also does the same every time theres a large patch and I have to remove all graphics drivers using DDU then reboot and hope
Re: (Score:3)
I find that statement interesting. I was playing GTA 5 on a i7-6700K last night and didn't' notice any performance impact from the game. I was running over goats just as well as I always have been. Do you have any benchmarks you can provide, before and after?
Re: (Score:2)
Here's a real world example of the performance loss from the patch:
https://www.epicgames.com/fort... [epicgames.com]
Looks like about 60% for that server workload. My guess is that Microsoft made it optional on server versions of Windows after seeing similar numbers.
Re: (Score:2)
Hate to break it to you, but that knocks you out of the running for "their best customers." Any small to medium business that refreshes machines every three years to stay within vendor support will buy more CPU's in a month than you'll buy in a lifetime. Intel manages to fleece gamers for some high-margin parts, but their real money is in the volume sales for mid-range plus Xeon's to business.
Re: (Score:2)
Re: (Score:2)
PCID is only supported on newer cpus
Having run across a nearly four-year old CPU having PCID (Haswell Xeon E3-1276), I'm less inclined to say that "newer" is entirely accurate; more that it was less common...
Re: (Score:3)
PCID is supported on Westmere (2010) and up.
Re: (Score:2)
Probably depends on your OS as well (I think Linux only added it in 4.14?).
Re:Only here (Score:5, Insightful)
Re: (Score:2)
You are right. You can't expect Microsoft to fix their older operating systems that are under support.
They are fixing them.
There are patches for Windows 7, 8, and 10. Because the kernel architecture has evolved over time, the performance deltas are different when patching each OS. Microsoft has been streamlining and modularizing the Windows kernel for a decade, so this performance difference should be expected.
E.g., Microsoft kicked font rendering out of the kernel going from Windows 7 to 8---this means fewer transitions between user and kernel execution when rendering web pages or Office documents. Since t
Re: (Score:2)
Re: (Score:2)
Windows 10 was a free update!
Re: (Score:2)
Probably. I assume any such reductions are already included in the overall performance metric.