10.13.2 Supplemental Update IO impact

[casperghst42]

Does anyone know if there is a way to role back the 10.13.2 Supplemental Update?

I've lost 40-50% in raw IO on my early 2013 rMBP, where I used to get approxy 600mb/s I now get around 200-250mb/s when testing with Blackmagic design (I know that it's no real indication), but I need the IO for when running VM's ...

Cheers.

 

[leman]

Did you actually lose any real world performance with your VMs? The slowdown is likely to because of the increased isolation so that your system is less vulnerable to the recently infamous exploits. BTW, running stuff in a VM is not a guarantee of safety — these exploits make it possible for attacker to get your sensitive data from the host machine when running a malicious program inside the VM.

 

[chrfr]

Does anyone know if there is a way to role back the 10.13.2 Supplemental Update?

I've lost 40-50% in raw IO on my early 2013 rMBP, where I used to get approxy 600mb/s I now get around 200-250mb/s when testing with Blackmagic design (I know that it's no real indication), but I need the IO for when running VM's ...

Cheers.

I'm still seeing >600MB/sec writes and >700MB reads on a 512GB SSD in a late 2013 MBP running 10.13.2 with the supplemental update installed, so I don't think your problem is as a result of the supplemental update.

 

[casperghst42]

Did you actually lose any real world performance with your VMs? The slowdown is likely to because of the increased isolation so that your system is less vulnerable to the recently infamous exploits. BTW, running stuff in a VM is not a guarantee of safety — these exploits make it possible for attacker to get your sensitive data from the host machine when running a malicious program inside the VM.

My notebook is my portable lab .. vulnerability or not, I need my VM's to be able to do my work. But otherwise from that you're right, due to Spector and Meltdown everything we thought was secure back in December is not out of the window. But that still does not change the fact of the need for VMs.

I did loose IO performance, how much I don't know, but some is gone. Blackmagic is probably not the best indicator, and after a massive clean up it now is somewhat better ... don't know what it was.
[doublepost=1516448966][/doublepost]

I'm still seeing >600MB/sec writes and >700MB reads on a 512GB SSD in a late 2013 MBP running 10.13.2 with the supplemental update installed, so I don't think your problem is as a result of the supplemental update.

Good to know, but there is a difference between the early and late 2013 rMBP's - the latter has PCI ssd's and the early have sata - anyhow it's good to know that it's probably a local issue.

 

[chrfr]

Good to know, but there is a difference between the early and late 2013 rMBP's - the latter has PCI ssd's and the early have sata - anyhow it's good to know that it's probably a local issue.

I haven't seen any I/O changes of note on a 2012 MacBook Air either, so I don't think you can point to the update for this problem. To be sure, if severe slowdowns like yours were widespread, you'd have no trouble finding others with the same problem.

 

[SoyCapitanSoyCapitan]

I'm still seeing >600MB/sec writes and >700MB reads on a 512GB SSD in a late 2013 MBP running 10.13.2 with the supplemental update installed, so I don't think your problem is as a result of the supplemental update.

The CPU bugs don’t affect that kind of disk performance.

 

[h9826790]

You may interested in this thread.

https://forums.macrumors.com/threads/10-13-2-17c205-check-ssd-speeds.2100638/

And I still can't get it. How can a VM not able to work if the SSD sequential speed changed?

 

[shafez]

13" Retina Macbook Pro Late 2013, High Sierra 10.13.2

 

[casperghst42]

You may interested in this thread.

https://forums.macrumors.com/threads/10-13-2-17c205-check-ssd-speeds.2100638/

And I still can't get it. How can a VM not able to work if the SSD sequential speed changed?

Intersting thread ... I found another write up, which pushed me in the direction of the SSD being tired, I'm switching it next month (also need more space), and then I'll see if there is a difference.

A VM is not directly impacted, but if you run multiple then it can become a bore.