MX 17/18 Repository: The Liquorix Kernel Thread

Message
Author
User avatar
stsoh
Posts: 183
Joined: Sun Aug 20, 2017 10:11 am

Re: MX 17 Repository: The Liquorix Kernel Thread

#31 Post by stsoh »

asqwerth wrote:but even with performance benefits, updates may still bork other parts of your system because of your
...upgrade gcc-7 and cpp-7 from buster repo
Does that mean every time it breaks, you will reinstall your system and then install the latest and greatest kernel from a .deb file again, with the necessary importing of gcc and cpp from buster repo?

That's my understanding of your comments:
...anyway i use minimal space for mxlinux (16gb of my 120gb ssd), freshly install with every beta/new release. fiber broadband take only a few minutes to download, upgrading with deb files takes about a couple of minutes instead of two hours for make install from kernel tar files
no, does not break. no problems at all, even update r normal. i do fresh installation because it is much easier and fast than iso.zsync.
MX-17.1_x64 Horizon, G41M-P33 Combo (MS-7592), Pentium E5400 (2706 MHz), 8Gb RAM (984 MT/s),
Intel 4 Series Integrated Graphics, Realtek PCIe Fast RTL8101/2/6E, PCI Gigabit RTL8169 Ethernets.
Accepted Linux when i found MX-Linux in 2016.

User avatar
timkb4cq
Developer
Posts: 3186
Joined: Wed Jul 12, 2006 4:05 pm

Re: MX 17 Repository: The Liquorix Kernel Thread

#32 Post by timkb4cq »

Those results are essentially identical - certainly within the margin of error for the tests.
HP Pavillion TP01, AMD Ryzen 3 5300G (quad core), Crucial 500GB SSD, Toshiba 6TB 7200rpm
Dell Inspiron 15, AMD Ryzen 7 2700u (quad core). Sabrent 500GB nvme, Seagate 1TB

User avatar
stsoh
Posts: 183
Joined: Sun Aug 20, 2017 10:11 am

Re: MX 17 Repository: The Liquorix Kernel Thread

#33 Post by stsoh »

benchmark r marginally close, in real life usage, it is better.
MX-17.1_x64 Horizon, G41M-P33 Combo (MS-7592), Pentium E5400 (2706 MHz), 8Gb RAM (984 MT/s),
Intel 4 Series Integrated Graphics, Realtek PCIe Fast RTL8101/2/6E, PCI Gigabit RTL8169 Ethernets.
Accepted Linux when i found MX-Linux in 2016.

User avatar
stsoh
Posts: 183
Joined: Sun Aug 20, 2017 10:11 am

Re: MX 17 Repository: The Liquorix Kernel Thread

#34 Post by stsoh »

updated liquorix 4.14.6, benchmark still as close as to 4.13.13. no breakage as everyone excepted to have, synaptic packages update as per norm.

Code: Select all

$ inxi -F
System:    Host: mx1 Kernel: 4.14.0-6.1-liquorix-amd64 x86_64 bits: 64 Desktop: Xfce 4.12.3
           Distro: MX-17.RC1_x64 Horizon December 10, 2017
Machine:   Device: desktop Mobo: MSI model: G41M-P33 Combo(MS-7592) v: 7.0 serial: N/A
           BIOS: American Megatrends v: V32.13 date: 11/06/2014
CPU:       Dual core Pentium E5400 (-MCP-) cache: 2048 KB
           clock speeds: max: 3145 MHz 1: 3145 MHz 2: 3145 MHz
Graphics:  Card: Intel 4 Series Integrated Graphics Controller
           Display Server: x11 (X.Org 1.19.2 ) drivers: modesetting (unloaded: fbdev,vesa)
           Resolution: 1920x1080@60.00hz
           OpenGL: renderer: Mesa DRI Intel G41 version: 2.1 Mesa 17.2.5
Audio:     Card-1 Intel NM10/ICH7 Family High Def. Audio Controller driver: snd_hda_intel
           Card-2 GYROCOM C&C driver: USB Audio
           Sound: Advanced Linux Sound Architecture v: k4.14.0-6.1-liquorix-amd64
Network:   Card-1: Realtek RTL8101/2/6E PCIE Fast/Gigabit Ethernet controller driver: r8169
           IF: eth0 state: down mac: 44:8a:5b:91:69:6c
           Card-2: Realtek RTL8169 PCI Gigabit Ethernet Controller driver: r8169
           IF: eth1 state: up speed: 1000 Mbps duplex: full mac: 64:70:02:00:71:f9
Drives:    HDD Total Size: 4120.8GB (73.1% used)
           ID-1: /dev/sda model: TOSHIBA_Q300 size: 120.0GB
           ID-2: /dev/sdb model: TOSHIBA_DT01ACA2 size: 2000.4GB
           ID-3: /dev/sdc model: TOSHIBA_DT01ACA2 size: 2000.4GB
Partition: ID-1: / size: 16G used: 5.5G (38%) fs: ext4 dev: /dev/sda2
           ID-2: swap-1 size: 0.42GB used: 0.00GB (0%) fs: swap dev: /dev/zram0
           ID-3: swap-2 size: 0.42GB used: 0.00GB (0%) fs: swap dev: /dev/zram1
Sensors:   System Temperatures: cpu: 43.0C mobo: N/A
           Fan Speeds (in rpm): cpu: N/A
Info:      Processes: 163 Uptime: 4 min Memory: 501.7/7948.2MB Client: Shell (bash) inxi: 2.3.53 
MX-17.1_x64 Horizon, G41M-P33 Combo (MS-7592), Pentium E5400 (2706 MHz), 8Gb RAM (984 MT/s),
Intel 4 Series Integrated Graphics, Realtek PCIe Fast RTL8101/2/6E, PCI Gigabit RTL8169 Ethernets.
Accepted Linux when i found MX-Linux in 2016.

User avatar
Stevo
Developer
Posts: 12776
Joined: Fri Dec 15, 2006 8:07 pm

Re: MX 17 Repository: The Liquorix Kernel Thread

#35 Post by Stevo »

We are now up to 4.14-17 (4.14-13.2) in the test repo. This incorporates up to the vanilla kernel 4.14.14 rc2 release.

By the way, it's now a lot simpler to adapt the kernel for building on MX, and mmikeinsantarosa has a new HP Xeon workstation that can crank them out in maybe ten minutes per kernel. Hopefully, he can find the time to help with these.

User avatar
stsoh
Posts: 183
Joined: Sun Aug 20, 2017 10:11 am

Re: MX 17 Repository: The Liquorix Kernel Thread

#36 Post by stsoh »

btw, pls remove liquorix 4.14.0-13.1 from mxpi test repo. it is a debug kernel, replaced by 4.14.0-13.2.
https://techpatterns.com/forums/about2643.html
MX-17.1_x64 Horizon, G41M-P33 Combo (MS-7592), Pentium E5400 (2706 MHz), 8Gb RAM (984 MT/s),
Intel 4 Series Integrated Graphics, Realtek PCIe Fast RTL8101/2/6E, PCI Gigabit RTL8169 Ethernets.
Accepted Linux when i found MX-Linux in 2016.

User avatar
Stevo
Developer
Posts: 12776
Joined: Fri Dec 15, 2006 8:07 pm

Re: MX 17 Repository: The Liquorix Kernel Thread

#37 Post by Stevo »

We should now be up to 4.14-22 (4.14-15.2).

When asked why Liquorix uses good ol' cpufreq instead of pstate like Debian for Intel processors, Steve Damentz replied:
When intel_pstate is used with MuQSS, intel_pstate is too aggressive at increasing frequency, leaving the CPU running at a higher voltage than necessary when idle or running mixed loads. The 'schedutil' frequency governor has the same problem as 'intel_pstate'.

And even if these frequency spikes were fixed in a later version of MuQSS or intel_pstate, MuQSS distributes its workload significantly more evenly than CFS, resulting in less opportunities to boost individual cores. With MuQSS, it's more important that all core turbo works correctly rather than single core.

User avatar
Stevo
Developer
Posts: 12776
Joined: Fri Dec 15, 2006 8:07 pm

Re: MX 17 Repository: The Liquorix Kernel Thread

#38 Post by Stevo »

We have been rolling in backports as they appear. We are now up to 4.15-4 in the test repo, which incorporates the 4.15.9 and 4.15.10-rc kernel patches. All major dkms packages should build with this kernel with the versions of those now in our main MX repo.

User avatar
entropyfoe
Posts: 591
Joined: Thu Apr 19, 2007 11:42 am

Re: MX 17 Repository: The Liquorix Kernel Thread

#39 Post by entropyfoe »

Stevo,

I am trying to understand these kernel numberings.
We are now up to 4.15-4 in the test repo, which incorporates the 4.15.9 and 4.15.10-rc kernel patches.
What is the 4.15-4? Where is the -4 coming from? Is that an MX number, because it seems it is actually from 4.15.9 and 10.
Where can you find the actual kernel numbers that correspond to these MX numbers.

I am currently running 4.15.5-antix.1. I thought that .1 was the internal revision number, and the .5 is the true kernel.org version.
Asus PRIME X470-PRO
AMD Ryzen 3600X (12 threads @ 3.8 GHz)
32 Gig DDR4 3600 (Crucial CL 16)
Nvidia GeForce GT 710
Samsung 970 NVMe nvme0n1 P1-3=MX-23, P4=testing
Samsung 980 NVMe =1TB Data, plus 2TB WD =backups
on-board ethernet & sound

User avatar
Stevo
Developer
Posts: 12776
Joined: Fri Dec 15, 2006 8:07 pm

Re: MX 17 Repository: The Liquorix Kernel Thread

#40 Post by Stevo »

Debian and Liquorix both use dual versioning for the kernel. The 4.15-4 means that it's the fourth version of the 4.15 kernel that Liquorix has released. The 4.15-1 first release was actually based on the 4.15.5 vanilla kernel release.

Example:

Code: Select all

$ uname -a
Linux mx1 4.15.0-7.1-liquorix-amd64 #1 ZEN SMP PREEMPT liquorix 4.15-3~mx17+1 (2018-03-07) x86_64 GNU/Linux
means that the third Liquorix version is based on the 4.15.7 kernel release.

AntiX kernels are built in a different manner, and don't have the same dual versioning scheme. You can see the Liquorix source packages at the bottom of this page: https://liquorix.net/debian/pool/main/l/linux-liquorix/

It's probably possible to adapt the antiX kernel builds to use the same framework as Liquorix. You just have to use the antiX config files, throw out the Liquorix patches in favor of the 4.15.X one and the antiX custom patches, and adjust the abi in debian/config/defines, as well as fixing the gcc version in that if necessary.

Locked

Return to “Package Requests/Status - MX 17/18”