Linux 5.4 vs. Liquorix Kernel Benchmarks For AMD Ryzen + Radeon Gaming On Ubuntu

[flower]

if someone wants to use it on arch it's called linux-zen there.
i just it for my gaming desktop

[Mario Junior]

Aye aye! I have a few more I use, but yes these are musts.

#https://github.com/graysky2/kernel_gcc_patch
echo "*** Copying and applying graysky's GCC 9 tweaks for Intel.. ✓"
if [ "${KERNEL_BASE_VER}" = "5.4" ]; then
cp $PATCH_PATH/graysky-gcc9-5.4.patch .;
patch -p1 < ./graysky-gcc9-5.4.patch;
else
cp $PATCH_PATH/graysky-gcc9-5.5+.patch .;
patch -p1 < ./graysky-gcc9-5.5+.patch;
fi

#https://github.com/sirlucjan/kernel-patches/blob/master/5.4/cpu-patches
echo "*** Copying and applying O3 patches.. ✓"
cp $PATCH_PATH/0001-cpu-patches.patch .;
patch -p1 < ./0001-cpu-patches.patch;

#https://cchalpha.blogspot.com/
echo "*** Copying and applying BMQ Scheduler patch.. ✓"
cp $PATCH_PATH/bmq_v5.4-r1.patch .;
patch -p1 < ./bmq_v5.4-r1.patch;

#https://lkml.org/lkml/2019/7/30/1398 and https://lkml.org/lkml/2019/7/30/1399
#https://gitlab.collabora.com/krisman/linux/commit/368eb7d8c86bd2a5db200567e82c477efc1a9502
#https://github.com/sirlucjan/kernel-patches/tree/master/5.4/futex-patches-sep
echo "*** Copying and applying Valve fsync/futex patches.. ✓"
cp $PATCH_PATH/futex-patches-sep/*.patch .;
patch -p1 < ./0001-futex-Split-key-setup-from-key-queue-locking-and-rea.patch;
patch -p1 < ./0002-futex-Implement-mechanism-to-wait-on-any-of-several-.patch;
patch -p1 < ./0003-futex-Change-WAIT_MULTIPLE-opcode-to-31.patch;

#https://github.com/sirlucjan/kernel-patches/tree/master/5.4/clearlinux-patches-v6-sep
echo "*** Copying and applying Clear Linux patches.. ✓"
cp $PATCH_PATH/clearlinux-patches-v6/0001-clearlinux-patches.patch .;
patch -p1 < ./0001-clearlinux-patches.patch;

#https://github.com/sirlucjan/kernel-patches/tree/master/5.4/bfq-patches-sep
echo "*** Copying and applying BFQ patches.. ✓"
cp $PATCH_PATH/bfq/$KERNEL_BASE_VER/0001-blkcg-Make-bfq-disable-iocost-when-enabled.patch .;
cp $PATCH_PATH/bfq/$KERNEL_BASE_VER/0002-block-bfq-present-a-double-cgroups-interface.patch .;
cp $PATCH_PATH/bfq/$KERNEL_BASE_VER/0003-block-bfq-Skip-tracing-hooks-if-possible.patch .;
patch -p1 < ./0001-blkcg-Make-bfq-disable-iocost-when-enabled.patch;
patch -p1 < ./0002-block-bfq-present-a-double-cgroups-interface.patch;
patch -p1 < ./0003-block-bfq-Skip-tracing-hooks-if-possible.patch;

(big shout out to sirlucjan for his excellent github repository and maintaining the patches!)

Thanks for the tip

PS: I'm using BMQ patch, not the PDS-mq patch.

[F.Ultra]

Take a look at total frame time then. CFS low latency is just awesome except if you're some awkward Google 'developer'.

But even total frame time should be a pure throughput problem and not a latency problem. Now this of course depends how the game logic is written with modern games having several threads all contributing to the frame time, but still as a general rule the time it takes to create a single frame should increase with a true realtime kernel.

edit: I think that I misread your post. Sorry about that.

[perpetually high]

>not compiling your minimal kernel with a patch for enable GCC optimizations for additional CPUs, PDS-mq schedule patch, fsync patch and some clear linux patchs

Aye aye! I have a few more I use, but yes these are musts.

#https://github.com/graysky2/kernel_gcc_patch
echo "*** Copying and applying graysky's GCC 9 tweaks.. ✓"
if [ "${KERNEL_BASE_VER}" = "5.4" ]; then
    cp $PATCH_PATH/graysky-gcc9-5.4.patch .;
    patch -p1 < ./graysky-gcc9-5.4.patch;
else
    cp $PATCH_PATH/graysky-gcc9-5.5+.patch .;
    patch -p1 < ./graysky-gcc9-5.5+.patch;
fi

#https://github.com/sirlucjan/kernel-patches/blob/master/5.4/cpu-patches
#https://github.com/sirlucjan/kernel-patches/blob/master/5.4/cpu-patches-sep
echo "*** Copying and applying O3 patches.. ✓"
cp $PATCH_PATH/0001-cpu-patches.patch .;
patch -p1 < ./0001-cpu-patches.patch;

#https://github.com/sirlucjan/kernel-patches/tree/master/5.4/futex-patches-v2 (single patch includes everything)
#https://lkml.org/lkml/2019/7/30/1398 and https://lkml.org/lkml/2019/7/30/1399
#https://gitlab.collabora.com/krisman/linux/commit/368eb7d8c86bd2a5db200567e82c477efc1a9502
echo "*** Copying and applying Valve fsync/futex v2 patches.. ✓"
cp $PATCH_PATH/0001-futex-Add-support-for-multiple-keys-at-the-same-time.patch .;
patch -p1 < ./0001-futex-Add-support-for-multiple-keys-at-the-same-time.patch;

#https://github.com/sirlucjan/kernel-patches/tree/master/5.4/clearlinux-patches-v6
#https://github.com/sirlucjan/kernel-patches/tree/master/5.4/clearlinux-patches-v6-sep
echo "*** Copying and applying Clear Linux patches.. ✓"
cp $PATCH_PATH/clearlinux-patches-v6/0001-clearlinux-patches.patch .;
patch -p1 < ./0001-clearlinux-patches.patch;

#https://github.com/sirlucjan/kernel-patches/tree/master/5.4/bfq-patches
#https://github.com/sirlucjan/kernel-patches/tree/master/5.4/bfq-patches-sep
echo "*** Copying and applying BFQ patches.. ✓"
cp $PATCH_PATH/bfq/$KERNEL_BASE_VER/0001-blkcg-Make-bfq-disable-iocost-when-enabled.patch .;
cp $PATCH_PATH/bfq/$KERNEL_BASE_VER/0002-block-bfq-present-a-double-cgroups-interface.patch .;
cp $PATCH_PATH/bfq/$KERNEL_BASE_VER/0003-block-bfq-Skip-tracing-hooks-if-possible.patch .;
patch -p1 < ./0001-blkcg-Make-bfq-disable-iocost-when-enabled.patch;
patch -p1 < ./0002-block-bfq-present-a-double-cgroups-interface.patch;
patch -p1 < ./0003-block-bfq-Skip-tracing-hooks-if-possible.patch;

(big shout out to sirlucjan for his excellent github repository and maintaining the patches!)

*EDIT 01/15/20: I removed the BMQ patches from above. After testing, I don't think it's a "must" and I underrated how well the default CFS scheduler behaved when it was under heaving load (i.e. compiling the kernel and continuing to use the desktop). BMQ was noticeably laggier/slower and was very obvious that something was compiling in the background (make -j4 on 4c/4t Haswell).

The following kernel settings have worked best for me, and the benchmarks consistently show it as well: CFS scheduler, tickless idle, 1000 Hz Timer, and Low-Latency Preemptive Kernel.

On CFS, there is practically zero slowdown while using the desktop as the kernel is compiling (great work Ingo Molnar and co!). So, while BMQ may behave better with gaming (and jury is still out on that) at the time of this writing, it was not worth it for me to give up CFS for other use cases.

Moving forward, I rather not deviate from such a core kernel feature like the CFS scheduler to an out-of-tree scheduler. Nothing against the hard work Con Kolivas, Alfred Chen, etc are doing and I think it's necessary, but I think for my Haswell system, I'm sticking with CFS.

As an aside- here are also the patches for CONFIG_PREEMPT_RT which I recently starting experimenting with. I found the results to be good, but it ultimately led to some system stability (because of all the different pieces involved, not necessarily because of the -rt patchset). There's also a bug in VirtualBox with spinlocks that causes it not to work with Windows 10 with the -rt patchset. However, for gaming and real-time low latency use cases, you should definitely give it a shot. Xanmod kernel can provide a -rt kernel if you're looking for something without compiling your own.

#https://rt.wiki.kernel.org/index.php...EEMPT_RT_Patch
#https://mirrors.edge.kernel.org/pub/...ojects/rt/5.4/
#Be sure to turn on in the menuconfig: CONFIG_PREEMPT_RT=y if using this
echo "*** Copying and applying CONFIG_PREEMPT_RT patches.. ✓"
cp $PATCH_PATH/patch-5.4.10-rt5.patch .;
patch -p1 < ./patch-5.4.10-rt5.patch;

tl;dr I'd personally avoid the BMQ/MuQSS scheduler patches for now and stick with the (very good) default stock CFS scheduler.

*END EDIT

[Mario Junior]

>not compiling your minimal kernel with a patch for enable GCC optimizations for additional CPUs, BMQ schedule patch, fsync patch and some clear linux patchs

[yokem55]

Take a look at total frame time then. CFS low latency is just awesome except if you're some awkward Google 'developer'.

The problem the the 'total' frame time graph is that it masks the relative quantities of low-frame times and high frame times. Real-time applications show perceived 'smoothness' not from really low latency (which does help!) but from really consistent latency. Game consoles get away with relatively lower frame rates because they invest more in frame-time consistency. VoIP applications can get away with relatively high levels of latency if the jitter is really low.

That's what 95th (98th, 99th) percentile frame-time measurements can show that the total frame-time graphs just don't show.

[ms178]

From my own testing with Company of Heros 2, with a low-latency kernel, aggressive compiler flags and the BMQ scheduler, it could also improve performance (in total roughly 10 % above stock).

[tildearrow]

Typo:

lower with the Liquprix kernel...

[Volta]

You need to do latency benchmarks to see improvement over standard kernel.
Low-latency kernels were always known to offer worse performance at the price of.. lower latency!

Take a look at total frame time then. CFS low latency is just awesome except if you're some awkward Google 'developer'.

[Yalok]

You need to do latency benchmarks to see improvement over standard kernel.
Low-latency kernels were always known to offer worse performance at the price of.. lower latency!