Similar to an old saw about paranoia, just because the risk is small for most users doesn't mean that Microsoft and Apple won't push a Critical Security Update that negatively affects performance for the whole inventory of machines. All you Windows 10 users out there report on your success at blocking that patch.
Apple may have already patched it. My iPad can't seem to get out of its own way since the iOS 11 update and the maintenance that followed.
I'm more concerned about the performance hit than the security risk for any of my personal information.
Looks like x264 and ffmpeg are within margin of error for impact.
Someone told me that encoding itself is unlikely to be too bad as long as the encoder is leveraging the GPU/somewhat competently programmed. It shouldn't need to be doing much out of kernel space, and that switch between user and kernel space is where this patch places a performance tax.
Maybe now is a good time to take a system image. If the slowdown is too much one can simply roll back and put then block the 'critical patch' going forward.
Popular GFX Card webforum Guru-of-3d did benchmarks of before and after patching this vulnerability and found very minimal (if any) performance hit. I highly doubt you'll notice a difference.
(Last page summation but best to read over the article):
Preliminary conclusion
Given what I am currently seeing, desktop users and PC gamers should not be too worried about significant performance drops. Most test results do show a negative effect on performance overall, but we're really talking in a realm of 2% differentials here. Now my remark here needs to include this, there probably will be more firmware updates and perhaps new patches for all protections to kick in, accumulated and activated these all can have an effect on performance. If that turns out to be the case, we'll have to re-do and revisit all tests.
Update: The file IO tests have been updated with a new ASUS BIOS 1003 we do see a drop off in certain 4K performance workloads, sometimes very significant. For all other CPU/SYSTEM/MEM benchmarks, you are looking at 1% maybe 2% differentials that could even be tagged as your typical statistical anomalies.
Currently, however, if you have a reasonably modern PC and IF this patch is all there is to it, you'll be hard-pressed to notice any difference, if at all, even with an NVMe SSD and a 4K drop-off. Again I would like to re-iterate that the effect on older dual and quad-core processors with a lower frequency could be far worse, the truth here is that I do not know the effect on that just yet. But on your average modern PC, at this time, this doesn't seem to be that worrying at all. That said - I'll need to test older processors, if there's a need performance differences wise, we'll certainly report back on that. This article is aimed at gamers and Windows 10 desktop users, the results in the server / datacenter segment might look completely different, that however is not our cup of tea in expertise or demographic. In the end, please do get yourself patched up in software and motherboard system BIOS / firmware okay?
We'll update this article if applicable at any time, the result sets remain 'preliminary'.
Former user
wrote on 1/6/2018, 5:51 PM
In the end, please do get yourself patched up in software and motherboard system BIOS / firmware okay?
I wouldn't. It's all theoretical, there are no virus's yet. Video editors are hard hit, especially 4k video. Initial patches are heavy handed and cause the most reduction in performance. It's a stop gap measure to do the most possible to stop the bad guys. More refined patches that don't hit the system heavily will come out later. That's what i'll wait for. Also the patch is causing blue screen of death on AMD chip computers.
They say it's most dangerous for data centres where CPU's can read the data from other CPU's from other customers.
From what I remember, that 4K impact was from 4k (as in kilobit) read/write operations on a SSD, not 4K video.
Former user
wrote on 1/6/2018, 6:54 PM
hmm. that is a little coincidental. I was typing as I was listening to a live interview about the situation. You and knowledgeable Joseph agree about 4k IO.
"A firmware update from Intel is also required for additional hardware protection, and those will be distributed separately by OEMs. It’s up to OEMs to release the relevant Intel firmware updates, and support information for those can be found at each OEM support website. If you built your own PC you’ll need to check with your OEM part suppliers for potential fixes. "
I am worried that boot up issues or hard crashes will occur on older 5+ years CPUs without a BIOS update that will accommodate the OS patches.
Former user
wrote on 1/7/2018, 5:45 AM
People are seeing slowdowns of up to 33% on nvme ssd drives & 17% with Sata ssd with the bios patches for spectre. I am using the 2 drives this guy benchmarks (evo 1tb rather than 2tb) . I say no thankyou to bios patch.
In the end, please do get yourself patched up in software and motherboard system BIOS / firmware okay?
I wouldn't. It's all theoretical, there are no virus's yet. Video editors are hard hit, especially 4k video. Initial patches are heavy handed and cause the most reduction in performance. It's a stop gap measure to do the most possible to stop the bad guys. More refined patches that don't hit the system heavily will come out later. That's what i'll wait for. Also the patch is causing blue screen of death on AMD chip computers.
They say it's most dangerous for data centres where CPU's can read the data from other CPU's from other customers.
Ditto on the blue screen of death on AMD computers. Mine was a victim.
I suppose this isn't much of a solution, but lately I've been thinking that we will all have to figure out a way for our machines to live off-line. I realize cloud services like Adobe are making this impossible, but if your computer is not online, then there should be no way for it to get infected. Perhaps you only go online with one system and just online for brief times when required, but you largely function off the grid so to speak. It just seems the vulnerabilities and attacks are endless. I remember the pilot to Battlestar Galactica and the only way the ship survived the Cylon attack was because it was an older ship and off line to the network. IDK, it just seems safer to minimize your machine's exposure if at all possible.
I have an Asus Z270 motherboard that now has the BIOS patch which I haven't installed yet, plus I have the same Samsung SSD they benchmarked with.
However the problem we have is we don't know if it is the benchmarking program running less optimally due to the patches (maybe they use some lower level tricks to flood the write/read queues which is impacted), or it is an actual read/write performance over everything that is taking a hit. Also the Asus firmware has a specific note of "Fixed SAMSUNG M.2 device issue" for my board and not the 370 they tested, so perhaps that has fixed the performance issues and the 370 will get another update shortly.
Either way we do know these updates are not having a positive effect on performance.
It's all FUD really, certainly for home users, as for anyone to exploit these bugs you would still need to download and run a malicious program, and even then it's hit and miss what data they will obtain. It would be easier to exploit Wi-Fi, or have the malicious program do something else more fruitful. The main issue is cloud data centres and virtual machines, as one customer can access data for another customer.
Well so far what I get out of this because my computers are 5+ years old is new motherboard, processor, bios, operating system update, memory, and maybe a slug of other new stuff to match the new new stuff.
Interesting article on this, mainly from the POV of a web developer, but applies to us all with regard to things like setting up 2-factor authentication and not doing banking on shared wifi networks: https://www.freelock.com/blog/john-locke/2018-01/spectre-meltdown
This brought performance back up to what it was before. I'm not convinced home users need this fix so I'm happy leaving it disabled.
Regards
Phil
Former user
wrote on 1/18/2018, 8:15 AM
This app will remove the meltdown & specture patches. At the moment it's preferable to not have the patches as they cause such a performance hit & there's not been any occurrences of an actual attack in the wild. It's all just theoretical
Due to some hardware problems I had with my ‘old’ system, I build up a new one based on an i7-8700k. Without over clocking I could see a slight improvement of the UHDp30 Intel HEVC (QSV) render speed and with my AMD R7 used for video acceleration on internal preview, I could reach the full 50fps frame rate in an UHDp50 Vp15 B261 project with UHDp50 footage from a GH5.
Caused by the MAGIX online activation there were some uncompleted MS Updates pending that I now allowed to complete – with a disastrous result !!! Now the system only reaches about 47 to 48fps and this with an increased CPU load of ~63% where I had only 33% before.
I already deinstalled the latest KB4058043 + KB4054517 (Flash Player 4053577) and the KB 4054022 can not be deinstalled, but I can not get the performance of the system before the MS update – may be it had an µcode-update for the CPU on board, that can not be undone by deinstalling the Kbxxxxx.
Now I have no clue how to get back the original i7-8700k Vp15 performance ????????
I really wish that Vegas 14/15 and following will allow to be activated offline to allow the systems to work without net-connection.
Former user
wrote on 1/19/2018, 7:06 AM
Did you try app I posted?
My older intel computer changed from 'good performance' to 'poor performance' with windows meltdown patch. But after I used app to uninstall patch & restarted computer , app reports 'good performance' again and unpatched.