What's new

Intel slowdown

gman713

Member
I don't know if anyone brought this up yet but are the current SB2's impacted by the security flaw? I guess it is too soon to know the impact to performance after the patch.
 
Well there is a vulnerability tool out there from Intel. It will tell you if your device is affected and my Surface Book said it was affected. As far as the fix MS will be adding to Tuesday security updates next week.
 
The Patches are already released. Some antivirus programs may prevent it from being delivered though as there is a dependency on a compatible AV. Then there is the usual mysterious not getting the update for no apparent reason and you just have to keep waiting.

I just got it on my SP3... haven't noticed any impact yet.
 
Here's a run of Passmark with the patch applied.
upload_2018-1-4_10-47-35.png


These scores are actually better than the scores I got before with Passmark 8.0. We probably shouldn't compare different versions of the Test software so if someone can run it on an i5 SP3 without the patch and post it we would have a direct comparison.

Here ae the prior results
Forum member hands on reviews SP4

Passmark 8.0 1874
CPU 3964
2D 464
3D 532
Mem 1585
Disk 2400
 
Ran the tool and it says that my 15 inch SB2 is not vulnerable to this particular intel chip design flaw.
I guess the chip was designed after they became aware of the flaw.

EDIT: I actually don't know how this tool works. It could not be detecting vulnerabilities because the windows patch is working and prevents the tool from detecting an exploit. If somebody who hasn't updated in a while wants to run it to compare results I'd like to see them.

KUgkzW0.png
 
Last edited:
I uninstalled the patch and ran Passmark 3 times then reinstalled it and ran 2 more times for 3 runs with and without.

The averages come out to a slight advantage without the patch but the highest run with the patch were higher than the lowest run without the patch.


Passmark compare.PNG
 
I just had a thought- would MS still push a patch out to the SB2 even if the chip is not vulnerable?
 
I suggest reading some newer stories on this issue. First, it's not an Intel CPU issue, but rather a pair of exploits based on how all modern CPUs work. It applies to AMD and ARM processors as well. Second, the performance impact is actually negligible for most people. It's limited to some very specific and not terribly common workflows.

In short, this is a story that was essentially incorrect as first reported and has since been clarified into something completely different. The exploits are real and important, but the response isn't going to slow down our SB2s either much at all or any different than if we were running AMD- or ARM-based systems.
 
I suggest reading some newer stories on this issue. First, it's not an Intel CPU issue, but rather a pair of exploits based on how all modern CPUs work. It applies to AMD and ARM processors as well. Second, the performance impact is actually negligible for most people. It's limited to some very specific and not terribly common workflows.

In short, this is a story that was essentially incorrect as first reported and has since been clarified into something completely different. The exploits are real and important, but the response isn't going to slow down our SB2s either much at all or any different than if we were running AMD- or ARM-based systems.

Apparently there are two different bugs:

"The flaw, now named Meltdown, was revealed on Wednesday and affects most processors manufactured by Intel since 1995.

Separately, a second defect called Spectre has been found that also exposes core memory in most computers and mobile devices running on chips made by Intel, Advanced Micro Devices (AMD) and ARM Holdings, a unit of Japan’s Softbank. "

How a researcher hacked his own computer and found 'worst' chip flaw
 
Right. Meltdown applies primarily to Intel CPUs and is both the easiest to exploit and the easiest to fix. Spectre is harder to exploit but also much harder to fix. The bottom line is that the impact of fixing the exploits likely won't matter directly to us end-users.

Now, the impact in data centers and for Linux web servers is likely to be more significant, as I understand it. But unlike what was implied by some earlier stories based on the original The Register report, we're not going to see our Intel-based PCs suddenly slowed down by 30%.
 
Back
Top