Do The Meltdown and Spectre Patches Affect PC Gaming Performance? 10 CPUs Tested
-
Page 1:Gaming With Meltdown And Spectre
-
Page 2:Exploiting The Unexploitable
-
Page 3:Test Setup
-
Page 4:VRMark, 3DMark & AotS: Escalation
-
Page 5:Civilization VI Graphics & AI, & Dawn of War III
-
Page 6:Far Cry Primal, Grand Theft Auto V & Hitman
-
Page 7:Shadow of War, Project CARS 2 & PUBG
-
Page 8:Game On...For Now
Test Setup
Our methodology for this story is fairly simple in practice. First, we performed a round of benchmarks on our Windows 10-based systems. Then, we applied the relevant patch and repeated the tests. Although the Variant 2 operating system patch we're using was deactivated via Windows, it should represent worst-case performance for now. We did not use a BIOS with Spectre mitigations because none are currently available (they were pulled). We'll follow up with more comparative testing as firmware updates arrive.
We tested each processor at its stock settings, including representatives from the Core i3, Ryzen 3, Core i5, Ryzen 5, Core i7, Ryzen 7, and Pentium line-ups. Naturally, overclocked systems will fare better. And as always, we disable Enhanced Multi-Core Turbo on our test systems.
In order to avoid variance from GPU Boost as our GeForce GTX 1080 heats up, we use multiple runs from each benchmark in quick succession. We select the median value from the last recordings, so, in many cases, the patched/unpatched results could land higher or lower than each other due to the tight variances you'll see today.
MORE: CPU Security Flaw: All You Need To Know About Spectre
MORE: Best Gaming CPUs
MORE: Intel & AMD Processor Hierarchy
Last page, third paragraph:
"so it's possible that the impact on older CPUs could be minor as well (game testing on those is in-progress)."
3/4 of all personal computers in the world today are NOT running windows 10. I don't know the exact percentage of gaming systems that are NOT running windows 10, but surely it is substantial.
Why wasn't the performance hit measured on the operating system running 3/4 of all PC's in the world today published immediately? To date, it appears that these numbers are being withheld from the public; the only reason has to be that the performance hit is absolutely massive in many cases.........Oh, and out of the total number of PC's used world wide, "gaming" PC's are a very small percentage, again begging the question of why tests are only being published for windows 10....
I have a question.
I read originally that AMD Zen architecture had near-immunity to Spectre variant 2 because a CPU specific code (password if you will) (that changes with each CPU) was required in order to exploit the CPU. Which is why AMD was claiming that Zen was almost immune to Spectre variant 2. Is this not the case?
AMD continues to insist that Spectre 2 is difficult to exploit due to CPU architecture. You left this out, and you continually lumped AMD with Intel with respect to Spectre 2 vulnerability.
This is misleading to your readers, and portrays a bias towards Intel.
https://www.amd.com/en/corporate/speculative-execution
Global OS penetration for Win 10 and Win 7 is effectively tied.
From the AMD page (which you linked)
And...
AMD hasn't released the microcode updates yet, but to its credit, it's probably better to make sure it is validated fully before release.
Do you know anything about the AMD Spectre variant 2 relating to the CPU code which is required beforehand in order to exploit AMD CPU's using variant 2?
I'm gathering that the answer is no.
As AMD didn't mention why they think they're less susceptible to variant 2 in the white paper, I'll assume that what I had was old news and no longer applies. I'm still curious though.
Win7 is dying out so soon that putting efford to win10 is something that is just sensible. Win7 is popular, so Yep. But end of the support is next year, so most likely not much efford is put in that Front.
'Nearly immune'? You mean the same as 'not immune'? 'Difficult to exploit' is PR for 'exploitable'.
This is AMD PR pushing a 'Spectre/Meltdown don't affect us... Much!' People wanted to hear that, so they read it in a positive light, rather than the practical light that AMD was also impacted. The simple fact they are releasing patches as well should tell you *everything* you need to know.
PR versus reality, AMD was impacted too. They are not immune, as they have said.
That 3/4 estimate is close for gamers though.
According to the most recent Steam Hardware and Software surveys,
56% of gamers using Steam are still on Windows 7 64Bit, and 35.4% of users are on Windows 10 64Bit.
http://store.steampowered.com/hwsurvey/
Tests should really be repeated with CPUs that lack PCID and/or INVPCID flags. I recall these flags and some newer commands are leveraged heavily to protect against Spectre and Meltdown while minimizing performance impact. Not having these CPU features means a more significant impact in gaming. As an i7-4930k user, I've noticed a huge hit to Civilization VI.
I'm pretty sure the performance difference between Windows 7 and actual Windows 10 are negligible.
That being said, I'm still on Sandy Bridge (2600k) and I too, would like a comparison of older generations, even though they "think" there's only a negligible effect for older gens too.
If they get performance and instability impact, so be it...is their fault using the hacked mods.
To try and turn this into a binary issue where every product is either "impacted" or "immune" is a drastic oversimplification of an extremely complex issue. After months of study the top experts still don't fully understand the full range of possibilities with Spectre, so outside of CPUs without branch prediction, no one can responsibly claim absolute "immunity". You'll note that Intel's early round of patches were released with the promise of "immunity" to Spectre but plenty of independent experts argued that they only actually reduced the vulnerability. The patches made the vulnerability much more difficult to leverage, but theoretical exploits remained. This is the problem with claiming "immunity", you only need one example, however unlikely or infeasible in the real-world, to disprove your claim.
In just about all forms of security there are degrees of vulnerability. For example, a home network running an unsecured wifi network is "exploitable", while the network at an NSA office is (I assume!) "difficult to exploit". There will doubtless be an array of security measures in place on the NSA network to make it extremely difficult to exploit... but history tells us the network and their data is not immune to being compromised. To characterise both networks as "exploitable" might be technically correct, but it's not an accurate representation of the situation.
Press Releases should absolutely be scrutinised and I'm not suggesting we just take AMD's word for this. It may well turn out to be PR nonsense and AMD every bit as vulnerable as Intel. But that needs to be determined by experts through testing. This issue is much more complicated than a simple "vulnerable/not vulnerable" equation.
AMD is preparing an OPTIONAL microcode update, that will insert some instructions to keep branches "in their own lanes" so to speak, as a precaution. But, they are NOT going to implement the main instruction that has the performance hit to the Intel chips. Maybe, someday, someone may get Spectre 2 to run on Zen, but as of now, it's a non issue. To imply that there are proof of concept programs in the wild affecting the Zen cpus by omission is just not right. Near zero risk of Spectre 2, and a vulnerability to both Specter variants are NOT different statements. Near zero means a remote chance of vulnerability. Still, since there are NO examples......
Intel makes quick references to memory. so it's easy to mark a set of memory you want to look at. AMD doesn't.
so if the information you want is the eiffeltower. on AMD you start in the middle of spain with a set of binoculars ducttaped to your eyes. also you don't have a map. with Intel you still start with the binoculars, but you can pretty much choose your spawn location making it alot easier to get the data.