Re: AW: Re: [qubes-users] Qubes 4rc3 :: 50% reduced battery runtime compared to Qubes 3.2 on Lenovo X230
On Fri, December 22, 2017 8:55 am, '[799]' via qubes-users wrote: > > This I also what I assumed as there must be a good reason why Qubes Team > has switched to HVM instead of using PV VMs. Still I'd like to learn more > about the vulnerabilities, so I can make a decision risk vs. runtime. And > as we can easy switch the Virtualization Mode via qvm-prefs, I could use > a script to do so: - shutdown VMs > - change virt_mode > - restart VMs See https://www.qubes-os.org/news/2017/07/31/qubes-40-rc1/ > If I switch to disposable VMs, I assume the risk would be reduced. > Can this be done for the sys-vms? I remember some discussion of allowing that but not the conclusion! -- You received this message because you are subscribed to the Google Groups "qubes-users" group. To unsubscribe from this group and stop receiving emails from it, send an email to qubes-users+unsubscr...@googlegroups.com. To post to this group, send email to qubes-users@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/qubes-users/10db6fa2795009771a845d7f340ddd7c.squirrel%40tt3j2x4k5ycaa5zt.onion. For more options, visit https://groups.google.com/d/optout.
Re: AW: Re: [qubes-users] Qubes 4rc3 :: 50% reduced battery runtime compared to Qubes 3.2 on Lenovo X230
Another partial workaround, you can change virt_mode from hvm back to pv for trustworthy VMs, like vault, etc... The bug: https://github.com/QubesOS/qubes-issues/issues/2849 For me it amounts to an overhead of 10% x cpu core per hvm, according to xentop. -- You received this message because you are subscribed to the Google Groups "qubes-users" group. To unsubscribe from this group and stop receiving emails from it, send an email to qubes-users+unsubscr...@googlegroups.com. To post to this group, send email to qubes-users@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/qubes-users/XkuZmIdoTkJz9S65a06Lsc9PjGpvgr8AfkSwRfAU7-hLiyLDWe2fiwGcAOSQ9FLUAgqHSYdBRiDm7_BVfEKRoa8ZfeQnB-mHAVuUZAStNV8%3D%40protonmail.com. For more options, visit https://groups.google.com/d/optout.
Re: AW: Re: [qubes-users] Qubes 4rc3 :: 50% reduced battery runtime compared to Qubes 3.2 on Lenovo X230
On 12/13/2017 07:30 PM, '[799]' via qubes-users wrote: Hello Chris, Original-Nachricht An 13. Dez. 2017, 19:15, Chris Laprise schrieb > Increased CPU usage is a known issue. > You can see it in the 'xentop' listing. > This may be one of the core tradeoffs > when moving to R4.0 I'll run the test on Q4rc3 and Q3.2 which will hopefully help investigating this problem. On which level is the increased CPU generated? If this is happening on the AppVM level. This would mean that I should get a better runtime when running only dom0, correct? Of course this is only meant for investigating this issue. Try playing with xenpm Using: 'xenpm set-scaling-governor powersave' I win between 20min and 1h depending on the CPU load (with an average runtime of 3-4h) -- You received this message because you are subscribed to the Google Groups "qubes-users" group. To unsubscribe from this group and stop receiving emails from it, send an email to qubes-users+unsubscr...@googlegroups.com. To post to this group, send email to qubes-users@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/qubes-users/9bca29b7-6896-7bae-0135-ee719f10e758%40riseup.net. For more options, visit https://groups.google.com/d/optout.
Re: AW: Re: [qubes-users] Qubes 4rc3 :: 50% reduced battery runtime compared to Qubes 3.2 on Lenovo X230
On 12/13/2017 01:30 PM, [799] wrote: Hello Chris, Original-Nachricht An 13. Dez. 2017, 19:15, Chris Laprise schrieb > Increased CPU usage is a known issue. > You can see it in the 'xentop' listing. > This may be one of the core tradeoffs > when moving to R4.0 Thanks for the hint. Honestly I can't believe that Qubes 4 comes with such a big trade-off. Keep in mind that the X230 has a very big battery and reduced performance that's why I can squeeze out ~10hrs battery runtime on Windows and Qubes 3.2. A "normal" laptop would run out after ~2 hrs of time. To test this I'll run my x230 connected to a power meter with batteries out and take a look at how much power is needed. I'll run the test on Q4rc3 and Q3.2 which will hopefully help investigating this problem. On which level is the increased CPU generated? If this is happening on the AppVM level. This would mean that I should get a better runtime when running only dom0, correct? Of course this is only meant for investigating this issue. [799] I believe the increase is mostly in the stub domains (stubdoms). You can see them with xentop. There may also be some lack of power tuning in places like xenpm, sys-net and sys-usb. -- Chris Laprise, tas...@posteo.net https://github.com/tasket https://twitter.com/ttaskett PGP: BEE2 20C5 356E 764A 73EB 4AB3 1DC4 D106 F07F 1886 -- You received this message because you are subscribed to the Google Groups "qubes-users" group. To unsubscribe from this group and stop receiving emails from it, send an email to qubes-users+unsubscr...@googlegroups.com. To post to this group, send email to qubes-users@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/qubes-users/8984bb42-8f76-7839-4b3b-1341fe83be47%40posteo.net. For more options, visit https://groups.google.com/d/optout.