link to zdnet article

so those of us that keep our systems bang up to date might get some performance back from mitigations now in place in (KB4482887).

stay safe 😉

Thanks, Win 10 just updated, rebooted and automatically installed the patch. It's only for Windows 10 release 1809. Also, the install put a scare into me, as when it rebooted, I got a message from the BIOS to hit F1 because my CPU and or RAM configuration had changed. I'm not sure what's that about, as the KB info  says the patch is at the OS level.

Also worth noting is that it doesn't work with Skylake or newer CPUs:

"However, this construct is only safe to use on processors where the RET instruction does not speculate based on the contents of the indirect branch predictor. Those processors are all AMD processors as well as Intel processors codenamed Broadwell and earlier according to Intel’s whitepaper. Retpoline is not applicable to Skylake and later processors from Intel."

https://techcommunity.microsoft.com/t5/Windows-Kernel-Internals/Mitigating-Spectre-variant-2-with-Retpoline-on-Windows/ba-p/295618

Intel blocked the Spectre exploit in microcode in the newer processors.

9 hours ago, jabloomf1230 said:

Intel blocked the Spectre exploit in microcode in the newer processors.

Only the 9xxxk series CPUs, apparently: https://www.digitaltrends.com/computing/intel-9-series-cpu-spectre/

Edited March 4, 20197 yr by vortex681

More info on Retpoline for Windows here:  https://www.techspot.com/news/79038-microsoft-introduces-drastic-performance-fix-spectre-variant.html#comments

Greg

On 3/3/2019 at 7:43 PM, vortex681 said:

Only the 9xxxk series CPUs, apparently: https://www.digitaltrends.com/computing/intel-9-series-cpu-spectre/

The net result of all this is a whole lot of nothing, as can be gleaned from the posts in this thread. Anything older than Skylake (6xxx) will get the patch, but only with Windows 10 1809 and only as the patch is rolled out gradually by MS. The 9xxx processors have microcode built-in already. Anything in the middle (6xxx, 7xxx) already has a different patch and the performance reducing impacts of the earlier patch on these processors is minimal for flight simulation. There might be some very modest impact on app load times, but that's about it. The reason that the Spectre exploit was considered as a huge crisis was that it had a big impact of the network and disk performance of older Xeon-based servers.

So if you're running a 4xxx or 5xxx CPU under Win 10 you can jump up and down about this, although even for these CPUs, the impact is going to be more on disk I/O, which is not a critical factor in determining flightsim frame rates.

1 hour ago, jabloomf1230 said:

The 9xxx processors have microcode built-in already.

Not all of them - see my previous post.

update link some gamers are suffering from stuttering or lag after this update. it hasn't affected me with P3Dv4.4.

Microsoft recommends those affected remove the update (KB4482887).

follow up:

on my system KB4482887 was removed by the update process on 3/12 and KB4489899 has replaced it. this update is supposed to fix the problem for some users (gamers) that had issues with KB4482887. 

as of yesterday, this has had the opposite effect! yesterdays flight stuttered and ground textures were abysmal. during a flight of <500nm I had two CTDs.

if the flight tonight is still bad then KB4489899 is getting uninstalled. the irony of KB4482887 being fine and the 'fix' unstable is not lost on me.

as of now microsoft has put procedures in place to automatically uninstall 'buggy' updates on a users PC. this relies heavily on the data collected by win10 during operation. while I am not one to 'bury my head in the sand' with regards to updates, I would much prefer less background data collection and more control given back to the user if they choose it.

stay safe, etc.

2 hours ago, jpf said:

this relies heavily on the data collected by win10 during operation. while I am not one to 'bury my head in the sand' with regards to updates, I would much prefer less background data collection and more control given back to the user if they choose it.

That's how automatic updates have always worked. Microsoft needs to know if you have missing updates or specific software which is incompatible with the pending update to prevent system crashes.

yes, but I think this is at a different level to uninstalling immediately post failure. it sounds more 'intrusive' to me.

link to zdnet article

On 3/13/2019 at 8:38 AM, jpf said:

if the flight tonight is still bad then KB4489899 is getting uninstalled. the irony of KB4482887 being fine and the 'fix' unstable is not lost on me.

CTDs again so KB4489899 has been uninstalled. see how it goes...

On 3/2/2019 at 6:32 AM, jpf said:

link to zdnet article

so those of us that keep our systems bang up to date might get some performance back from mitigations now in place in (KB4482887).

stay safe 😉

Cool, I always update when prompted...

Been suffering from CTDs over and over. I've uninstalled both KBs and they still persist. I'll post back if I find something that works...