Windows 10 CPU Bug Fix Patch Benchmarks (updated)
Click here to post a comment for Windows 10 CPU Bug Fix Patch Benchmarks (updated) on our message forum
chispy
AsiJu
Thanks a lot! Not bad then, chispy is this in line with your SSD benches or did you see bigger hits?
chispy
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=00a5ae218d57741088068799b810416ac249a9ce
It seems on Ryzen systems the performance hit it's bigger than on intel , need to do much more testing. But same as Hilbert findings Storage performance it is where it is more pronounced this performance degradation.
Now let's hope AMD can work with Microsoft to make a new patch to exclude AMD cpu's from this BS and get back our performance. Linux users have already a patch available that excludes them from this performance degradation.
Here: Patch to exclude AMD hit the Linux kernel 5 or 6 hours ago. Hilbert Hagedoorn
Administrator
I am still doing some NVMe tests on another system as I had an issue with the Samsung 960 pre and post patch results. Will update once the numbers are done.
Turanis
Good to know,but will be better to test on normal SSD.You know,without fancy wheels like Optane,Nvme.
Not many users have Nvme and the speed on that is already "astronomical".
Hilbert Hagedoorn
Administrator
NVMe is MUCH heavier on file IO as it is 4 to 6 times faster. So any problem would be better visualized.
However, i have not patched the 2nd test system yet so will run a regular SSD before patching as well.
StewieTech
Good stuff boss.
Yxskaft
I really hope the older CPUs don't get much more affected than the new ones. i5 2500 CPUs are pushed pretty hard in today's games, and for people still not needing more than a Core 2 Quad or i5 750, even a 10% performance drop would be noticed.
KayserKay
No Problems here with an Oldie (but Goldie) i5 4460 and GTX 1070.
Tested some heavy on the CPU, realtime physics based games: BeamNG Drive, EuroTruck Simulator 2 and ATS, City Car Driving...all on Utra and no unknown FPS drops!
Performance with Audiointerface and ASIO Driver: still rock solid, no dropouts!
RzrTrek
Looks like the AMD people cried wolf, but as always, good work and thank you for your benchmarks.
Hilbert Hagedoorn
Administrator
Alessio1989
sykozis
Goiur
Already installed, waiting for something to blow up lol
H83
Nice to see Hilbert already posting some benchies about the patch. But i have to say i´m surprised with the results after all the talk that performance would tank and the difference is around 2% in the end... Did people overblown the issue or should we wait for further updates/patches that may cripple the performance severely???
Thanks for the quick nice work Hilbert!
tsunami231
im sorry but those test before and after look like there within "margin" of error and are nothing most people will ever notice, another thing made to look worse then it acutal is?, the actual security hole is actual bad like really bad, and considering I used VM for the purpose to doing things that may or may not result in malware/virus that i dont want on my system this flaw make me worried then again I been doing this VM thing for years now think i was fine.
Personally I was hoping these fixes could be done threw say MB FW updates, which i would think would minimize performance hits more then it being done by OS, but HW changes need to happen to plug it from hw stand point? which means $$$ for intel cause people will still buy intel, I know i will, i just wont be making new PC anytime soon, I already kinda miffed by the fact year and half after i got my 6700k I could got 8700k with 6core/12t and same ghz and same STP for same price. which has the same flaw anyway. maybe in 3/4 years time this flaw will be shorted out and fixed hw wise and 8c/16t will be the normal with the same or BETTER stp, yes stp still important to me more so then mtp which will probably have different security flaw that in it hw that would need sw/fw patch to fix that might hurt performance that will need new hw change meaning more $$$ for intel, none these "issue" would be much of a issue if they would not be made known to the world, then ever bad person with know how will try and exploit things thing vs just few that know about.
Security flaws should be reported yes and they should be fix, but they should NOT be made public and made worldwide knowledge. that is like governments say to the world hey our nuclear plants and stock pile of weapons and other vital stuff are unprotected and located here are desings and even info on what security and flaws we have come get us... and in world were cyber threat is just as bad real world why would people make such thing public knowledge. Now tell said people that response able for the flaw and fixing to fix it they have till so and so time to do so or this info will be released world if they don't fix it is another story, even then i dont think it should be done for the above reason, but all these "flaws" seem to get release in wild regardless now cause FB/Twitter and everything else cause people want there spot in lime light. regardless of info the releasing what damage it can cause, social media sites imo are one worst things happen, when comes to stuff like this.
jaggerwild
foxx1337
So I only performed one test on my system, and I would say that I get like 10% worse performance.
Core i7 2600K at 4 GHz, samsung 850 EVO ssd. The test is, in a ConEmu terminal under Windows, under Bash (from git-bash), I type this:
Weasis compilation:
Before patch: 53 seconds
After the patch: 1 minute, 5 seconds (so 65 seconds)
Weasis is a Java viewer for X-Ray, CT-scan and CAT-scan file formats. This ConEmu + MinGW + bash + Maven + javac scenario probably touches quite a few different aspects of the operating system, from file access, to lookup, writes, compilation, the (few) Weasis tests, etc. I tested by restarting the computer, opening ConEmu and running the clean + install cycle 5-6 times until it stabilized.
Hilbert Hagedoorn
Administrator
File IO tests have been updated, a SATA SSD was added as well. I am not seeing anything that worries me.
user1
good to see that the performance hit wont be very significant for the majority of users.