You are not logged in.
damige wrote:Also : I have done some testing in the bios and have reached (in my case) another extra watt less power usage.
I think i have narrowed it down to disabling the sdcard in the bios. Could someone confirm this behavior? (Yes i am using TLP)Thanks for tip - switched off my SD card and dropped idle from abiut 3.9 to 3.5
Thanks, whats strange though is i do get a power gain disabling it in the bios but not if i power it down from linux.
Looks like to me that there are some measurable gains to be had here. Just checked the diff again, i could not get below 4w before, now i saw 3.28w !
XPS 15 7590 | i9-9980HK | 32GB RAM | 512G NVME | Intel AX200 | NVIDIA 1650 | OLED 4K
Offline
My MB with 6500 died , dell replaced it with 6560 - and now I am in the club , consuming 4W where it used to be 2.2 ..... Still can not gain less than 3.5 - I got 16MB and killer 1535 so those two can make the difference. Anyway opening firefox, thunderbird , skype and so on wake the CPU up so not much idle.
Offline
My MB with 6500 died , dell replaced it with 6560 - and now I am in the club , consuming 4W where it used to be 2.2 ..... Still can not gain less than 3.5 - I got 16MB and killer 1535 so those two can make the difference. Anyway opening firefox, thunderbird , skype and so on wake the CPU up so not much idle.
I'm confused. What is 6550 and 6560?
Offline
i7 skylake CPUs - one with HD520 graphics, another with Iris 540.
Offline
id be happy at 4W ;P can barely go below 7W with the QHD touchscreen + 6560 i7 hehe (also i use my sdcard reader ;-)
Offline
id be happy at 4W ;P can barely go below 7W with the QHD touchscreen + 6560 i7 hehe (also i use my sdcard reader ;-)
Yes, same configuration, same wattage here. I do get down to 5W by disabling all the things USB (including the SD card reader) in the BIOS and running a light desktop environment and switching wireless off, which is useful when I want to get 9/10h out of it on a plane for example.
I tried the new PCIE APST powersaving options from the new rc kernel (4.11) but I didn't see much of a difference.
Offline
kang wrote:id be happy at 4W ;P can barely go below 7W with the QHD touchscreen + 6560 i7 hehe (also i use my sdcard reader ;-)
Yes, same configuration, same wattage here. I do get down to 5W by disabling all the things USB (including the SD card reader) in the BIOS and running a light desktop environment and switching wireless off, which is useful when I want to get 9/10h out of it on a plane for example.
I tried the new PCIE APST powersaving options from the new rc kernel (4.11) but I didn't see much of a difference.
I have less then that (or similar) on my 9550 quad core with 32GB, something is wrong. Yes, I have the FHD, but still.
Hardware: 2016 Dell XPS15 - matte FullHD - i5-6300HQ - 32GB DDR4 - Nvidia GTX960M - Samsung 840EVO 250GB SSD - 56Wh
Software: Plasma 5 - rEFInd - linux-ck - preload - prelink - verynice - psd - bumblebee
Offline
I have less then that (or similar) on my 9550 quad core with 32GB, something is wrong. Yes, I have the FHD, but still.
What's your CPU? QHD is know to draw a lot more power, even on windows, but I agree something is weird as the extra power draw is seen even with the screen off... I'd love to know what's wrong though.
Offline
OdinEidolon wrote:I have less then that (or similar) on my 9550 quad core with 32GB, something is wrong. Yes, I have the FHD, but still.
What's your CPU? QHD is know to draw a lot more power, even on windows, but I agree something is weird as the extra power draw is seen even with the screen off... I'd love to know what's wrong though.
i5-6300HQ. Yes, it's an i5, but it also is a quad core and should theoretically draw significantly more power. With the screen and WiFi off I can very easily get to sub-6W territories, and sub-4W without Plasma and other stuff open, usually. Last week I left it in its sleeve for 2 hours, accidentally turned on in Plasma, with Chromium open and WiFi on, and the fans were not even on when I removed it from the sleeve, which is quite impressive.
Also note that according to my power measurement on my machine almost one entire Watt is taken by the 32GB of RAM by themselves, even at idle.
Last edited by OdinEidolon (2017-03-17 11:30:32)
Hardware: 2016 Dell XPS15 - matte FullHD - i5-6300HQ - 32GB DDR4 - Nvidia GTX960M - Samsung 840EVO 250GB SSD - 56Wh
Software: Plasma 5 - rEFInd - linux-ck - preload - prelink - verynice - psd - bumblebee
Offline
i7-6560 uses way more power When I had 6500 with all USB switched off and lowest screen brightness at idle (mate desktop, one terminal opened and nothing else) it was 2.18W. Same settings on i7-6560 gives 3.8W. So almost double. Anyway in real use when some apps are running it is more anyway.
Offline
So I updated to BIOS 1.4.14 and now my battery constantly shows it's at 100%. What the hell? Can anyone else confirm this?
Offline
My bios is also 1.4.14 - but battery percentage reported ok.
Offline
So I updated to BIOS 1.4.14 and now my battery constantly shows it's at 100%. What the hell? Can anyone else confirm this?
Unplug the battery for a few minutes, this will reset the bios and probably fix it.
Hardware: 2016 Dell XPS15 - matte FullHD - i5-6300HQ - 32GB DDR4 - Nvidia GTX960M - Samsung 840EVO 250GB SSD - 56Wh
Software: Plasma 5 - rEFInd - linux-ck - preload - prelink - verynice - psd - bumblebee
Offline
kgizdov wrote:So I updated to BIOS 1.4.14 and now my battery constantly shows it's at 100%. What the hell? Can anyone else confirm this?
Unplug the battery for a few minutes, this will reset the bios and probably fix it.
You're joking right. I need to disassemble the laptop to unplug the battery.
I was able to fix it by restoring the BIOS defaults.
Last edited by kgizdov (2017-03-21 16:53:32)
Offline
unscrew 9 bolts and remove cover - not a joke . Its easy and does not void warraty - but you right it is better tio try simple things like reset bios defaults first.
Offline
Great news,
commit a79fa23c82a1b7ee3726ea348336fbe874130a7c of kernel 4.10.5 seems to fix openvpn not being able to set routes correctly on newer kernels
AND
This patch seems to fix the issue with HDMI on the DA200 !!!
I have compiled a 4.10.5 kernel version with the DA200 & nvme patches, you can get it from here if you want:
linux-nvme-4.10.5-1-x86_64.pkg.tar.xz 78975d9bfcddbb5b4604c4a18a9400e5780d47b47d7c42f1f652f5fbe88c76b5
linux-nvme-docs-4.10.5-1-x86_64.pkg.tar.xz 389eadf1467a640c3dd24a077b0eb70f52d990477f1562b98a00cfc79b9362a9
linux-nvme-headers-4.10.5-1-x86_64.pkg.tar.xz 97032dc9ca6e50c752f3174e132f92daeffb89c7ac94c71168385a4fa1df2aa1
Warning from intel regarding the DA200 patch
The patch is currently in review. It's a workaround, and it impacts all DP sinks, not just the problematic ones. Thus, we need to be careful about merging it, and extra careful about backporting it to stable kernels.
If we agree on the approach of the patch, I expect it would be included in v4.11. I expect the release will be in about a month or so.
Last edited by kubrick (2017-03-23 14:47:24)
Offline
New version of DA200 patch is out - tested with 4.10.5, it works fine.
Offline
New version of DA200 patch is out - tested with 4.10.5, it works fine.
Here's the final version: https://cgit.freedesktop.org/drm/drm-ti … 701ef54c42
Does anyone know if Dell's USB C to HDMI adapter uses the same chip as the DA200?
Offline
Dell USB C to HDMI uses different chip, not the same as DA200. It worked perfectly without a patch from the beginning, and works after patching. It is the DA200 which is a big fat hairy bug, not the USB-C - HDMI.
Offline
-- Reboot --
Apr 19 13:08:31 xps13 kernel: IP: drm_dp_get_mst_branch_device+0xa2/0x110 [drm_kms_helper]
Apr 19 13:08:31 xps13 kernel: BUG: unable to handle kernel NULL pointer dereference at (null)
my life lately when connecting to the WD15 dock on 4.10.10 and 4.11rc7 ;-)
not sure what changed yet, though obviously there's a bug somwhere in the drm or i915 code
(also i guess i might try to use HDMI then lol)
Last edited by kang (2017-04-19 20:52:28)
Offline
Now that 4.11 is out, is there any reason to run a patched kernel? Are there any relevant NVME or Intel graphics power saving patches that have not been merged upstream?
Offline
Now that 4.11 is out, is there any reason to run a patched kernel? Are there any relevant NVME or Intel graphics power saving patches that have not been merged upstream?
The only relevant patch that I know of that is of interest is the DP training patch if you want the DA200 to work. NVMe APST is now part of mainline (and was actually the cause for the one week delay...)
Offline
New bios update.. and hmm new error:
sed_opal:OPAL: Error on step function: 0 with error 8194: Unknown Error
Offline
New bios update.. and hmm new error:
sed_opal:OPAL: Error on step function: 0 with error 8194: Unknown Error
Self Encrypting Devices (SED) is new in 4.11, are you sure it's the BIOS upgrade that triggered this message, not the kernel upgrade?
Offline
kang wrote:New bios update.. and hmm new error:
sed_opal:OPAL: Error on step function: 0 with error 8194: Unknown Error
Self Encrypting Devices (SED) is new in 4.11, are you sure it's the BIOS upgrade that triggered this message, not the kernel upgrade?
you're right
Offline