The primary scheduling domain represents a group of CPUs in the system
where the scheduler will initially attempt to assign tasks. Tasks will
only be dispatched to CPUs within this primary domain until they are
fully utilized, after which tasks may overflow to other available CPUs.
The primary scheduling domain can defined using the option
`--primary-domain CPUMASK` (by default all the CPUs in the system are
used as primary domain).
This change introduces two new special values for the CPUMASK argument:
- `performance`: automatically detect the fastest CPUs in the system
and use them as primary scheduling domain,
- `powersave`: automatically detect the slowest CPUs in the system and
use them as primary scheduling domain.
The current logic only supports creating two groups: fast and slow CPUs.
The fast CPU group is created by excluding CPUs with the lowest
frequency from the overall set, which means that within the fast CPU
group, CPUs may have different maximum frequencies.
When using the `performance` mode the fast CPUs will be used as primary
domain, whereas in `powersave` mode, the slow CPUs will be used instead.
This option is particularly useful in hybrid architectures (with P-cores
and E-cores), as it allows the use of bpfland to prioritize task
scheduling on either P-cores or E-cores, depending on the desired
performance profile.
Example:
- Dell Precision 5480
- CPU: 13th Gen Intel(R) Core(TM) i7-13800H
- P-cores: 0-11 / max freq: 5.2GHz
- E-cores: 12-19 / max freq: 4.0GHz
$ scx_bpfland --primary-domain performance
0[||||||||| 24.5%] 10[|||||||| 22.8%]
1[|||||| 14.9%] 11[||||||||||||| 36.9%]
2[|||||| 16.2%] 12[ 0.0%]
3[||||||||| 25.3%] 13[ 0.0%]
4[||||||||||| 33.3%] 14[ 0.0%]
5[|||| 9.9%] 15[ 0.0%]
6[||||||||||| 31.5%] 16[ 0.0%]
7[||||||| 17.4%] 17[ 0.0%]
8[|||||||| 23.4%] 18[ 0.0%]
9[||||||||| 26.1%] 19[ 0.0%]
Avg power consumption: 3.29W
$ scx_bpfland --primary-domain powersave
0[| 2.5%] 10[ 0.0%]
1[ 0.0%] 11[ 0.0%]
2[ 0.0%] 12[|||| 8.0%]
3[ 0.0%] 13[||||||||||||||||||||| 64.2%]
4[ 0.0%] 14[|||||||||| 29.6%]
5[ 0.0%] 15[||||||||||||||||| 52.5%]
6[ 0.0%] 16[||||||||| 24.7%]
7[ 0.0%] 17[|||||||||| 30.4%]
8[ 0.0%] 18[||||||| 22.4%]
9[ 0.0%] 19[||||| 12.4%]
Avg power consumption: 2.17W
(Info collected from htop and turbostat)
Signed-off-by: Andrea Righi <andrea.righi@linux.dev>
While the system is not saturated the scheduler will use the following
strategy to select the next CPU for a task:
- pick the same CPU if it's a full-idle SMT core
- pick any full-idle SMT core in the primary scheduling group that
shares the same L2 cache
- pick any full-idle SMT core in the primary scheduling grouop that
shares the same L3 cache
- pick the same CPU (ignoring SMT)
- pick any idle CPU in the primary scheduling group that shares the
same L2 cache
- pick any idle CPU in the primary scheduling group that shares the
same L3 cache
- pick any idle CPU in the system
While the system is completely saturated (no idle CPUs available), tasks
will be dispatched on the first CPU that becomes available.
Signed-off-by: Andrea Righi <andrea.righi@linux.dev>
Allow to specify a primary scheduling domain via the new command line
option `--primary-domain CPUMASK`, where CPUMASK can be a hex number of
arbitrary length, representing the CPUs assigned to the domain.
If this option is not specified the scheduler will use all the available
CPUs in the system as primary domain (no behavior change).
Otherwise, if a primary scheduling domain is defined, the scheduler will
try to dispatch tasks only to the CPUs assigned to the primary domain,
until these CPUs are saturated, at which point tasks may overflow to
other available CPUs.
This feature can be used to prioritize certain cores over others and it
can be really effective in systems with heterogeneous cores (e.g.,
hybrid systems with P-cores and E-cores).
== Example (hybrid architecture) ==
Hardware:
- Dell Precision 5480 with 13th Gen Intel(R) Core(TM) i7-13800H
- 6 P-cores 0..5 with 2 CPUs each (CPU from 0..11)
- 8 E-cores 6..13 with 1 CPU each (CPU from 12..19)
== Test ==
WebGL application (https://webglsamples.org/aquarium/aquarium.html):
this allows to generate a steady workload in the system without
over-saturating the CPUs.
Use different scheduler configurations:
- EEVDF (default)
- scx_bpfland using P-cores only (--primary-domain 0x00fff)
- scx_bpfland using E-cores only (--primary-domain 0xff000)
Measure performance (fps) and power consumption (W).
== Result ==
+-----+-----+------+-----+----------+
| min | max | avg | | |
| fps | fps | fps | stdev | power |
+-----------------+-----+-----+------+-------+--------+
| EEVDF | 28 | 34 | 31.0 | 1.73 | 3.5W |
| bpfland-p-cores | 33 | 34 | 33.5 | 0.29 | 3.5W |
| bpfland-e-cores | 25 | 26 | 25.5 | 0.29 | 2.2W |
+-----------------+-----+-----+------+-------+--------+
Using a primary scheduling domain of only P-cores with scx_bpfland
allows to achieve a more stable and predictable level of performance,
with an average of 33.5 fps and an error of ±0.5 fps.
In contrast, using EEVDF results in an average frame rate of 31.0 fps
with an error of ±3.0 fps, indicating slightly less consistency, due to
the fact that tasks are evenly distributed across all the cores in the
system (both slow and fast cores).
On the other hand, using a scheduling domain solely of E-cores with
scx_bpfland results in a lower average frame rate (25.5 fps), though it
maintains a stable performance (error of ±0.5 fps), but the power
consumption is also reduced, averaging 2.2W, compared to 3.5W with
either of the other configurations.
== Conclusion ==
In summary, with this change users have the flexibility to prioritize
scheduling on performance cores for better performance and consistency,
or prioritize energy efficient cores for reduced power consumption, on
hybrid architectures.
Moreover, this feature can also be used to minimize the number of cores
used by the scheduler, until they reach full capacity. This capability
can be useful for reducing power consumption even in homogeneous systems
or for conducting scheduling experiments with smaller sets of cores,
provided the system is not overcommitted.
Signed-off-by: Andrea Righi <andrea.righi@linux.dev>
Abbreviate the statistics reported to stdout and remove the slice_ms
metric: this metric can be easily derived from slice_ns, slice_ns_min
and nr_wait, which is already reported to stdout.
Signed-off-by: Andrea Righi <andrea.righi@linux.dev>
Use the same idle selection logic used in scx_bpfland also in
scx_rustland_core.
Also drop fifo_mode and always use the BPF idle selection logic by
default as long as the system is not saturated, unless full_user is
specified.
This approach allows user-space schedulers aiming for maximum
performance to leverage the BPF idle selection logic (bypassing
user-space), while those seeking full control can enable full_user to
bypass the BPF CPU idle selection logic and choose the target CPU for
each task from user-space.
Signed-off-by: Andrea Righi <andrea.righi@linux.dev>
Immediately re-align p->scx.dsq_vtime to the global vruntime (+/- slice
lag) as soon as we are evaluating the task's vruntime.
This allows rapidly chase the minimum global vruntime, ensuring to not
over prioritize tasks tasks with a predominantly sleeping behavior
pattern.
Signed-off-by: Andrea Righi <righi.andrea@gmail.com>
Using negative values with --slice-us-lag can be useful to make
performance more consistent and prioritize newly created tasks over the
running tasks.
Therefore, allow to specify negative values from the command line and
also update the documentation of this option.
Signed-off-by: Andrea Righi <righi.andrea@gmail.com>
Periodically report to stdout samples of the effective time slice
applied to tasks.
While one could determine this metric by examining the max slice_ns and
nr_waiting metrics, directly reporting it to stdout allows users to
quickly identify what is happening and it provides a clearer overview of
the scheduling behavior.
Signed-off-by: Andrea Righi <righi.andrea@gmail.com>
Dispatching per-CPU kthreads directly is disabled by default, reporting
this metric can generate some confusion (since it is always 0), and even
if local kthread dispatches are enabled, they should be still considered
as regular direct dispatches (there is no difference in practice).
Therefore, merge direct kthread dispatches into direct dispatches and
drop the separate nr_kthread_dispatches metric.
Signed-off-by: Andrea Righi <righi.andrea@gmail.com>
Scale the task's time slice based on the average amount of tasks that
are currently waiting to be dispatched.
Use a moving average for the amount of waiting tasks to smooth out
potential spikes caused by temporary bursts of tasks piling in the wait
queues.
This was initially modeled in scx_rustland and it seems to work pretty
well also in scx_bpfland now.
Signed-off-by: Andrea Righi <righi.andrea@gmail.com>
Instead of using a static value to classify tasks based on their average
amount of voluntary context switches, try to periodically evaluate an
optimal threshold, based on a global average of voluntary context
switches among of all the running tasks.
Tasks with an average amount of voluntary context switches greater than
the global average will be classified as interactive.
The global average is evaluated as an exponentially weighted moving
average (EWMA), as:
avg(t) = avg(t - 1) * 0.75 - task_avg(t) * 0.25
This approach is more efficient than iterating through all tasks and it
helps to prevent rapid fluctuations that may be caused by bursts of
voluntary context switch events.
The dynamic nvcsw threshold enables a more precise adjustment of
the classification criteria to swiftly respond to global system changes:
tasks can be quickly classified as interactive, but if the system
experiences too many interactive events, the criteria for maintaining
interactive status become stricter. This creates a natural selection
process where only the most deserving tasks remain interactive.
Additionally, introduce the new option `--nvcsw-max-thresh N`, which
allows to extend or restrict the fluctuation range of the global average
threshold for voluntary context switches.
Tested-by: Piotr Gorski <piotrgorski@cachyos.org>
Signed-off-by: Andrea Righi <righi.andrea@gmail.com>
Update libbpf-rs & libbpf-cargo to 0.24. Among other things, generated
skeletons now contain directly accessible map and program objects, no
longer necessitating the use of accessor methods. As a result, the risk
for mutability conflicts is reduced greatly.
Signed-off-by: Daniel Müller <deso@posteo.net>
We always use nr_cpu_ids to represent the maximum CPU id returned by
scx_bpf_nr_cpu_ids().
Replace cpu_max with nr_cpu_ids to be more consistent with the rest of
the code.
Signed-off-by: Andrea Righi <righi.andrea@gmail.com>
We can rely on scx_bpf_nr_cpu_ids() to create all the possible per-CPU
DSQs, eliminating the need for the hard-coded limit MAX_CPUS.
In this way scx_bpfland can support the same amount of CPUs that the
kernel can handle.
Signed-off-by: Andrea Righi <righi.andrea@gmail.com>
Instead of constantly checking the need to drain tasks from the DSQs of
the offline CPUs, provide an atomic flag to notify when there are tasks
to be drained from the offline CPUs.
Signed-off-by: Andrea Righi <righi.andrea@gmail.com>
Refine the safeguard mechanism to avoid generating too many interactive
tasks in the system, which could nullify the effect of the
interactive/regular task classification.
The safeguard mechanism operates by pausing the promotion of new tasks
to interactive status during the task wake-up process, whenever the
number of interactive tasks in the priority queue exceeds a specific
limit (set to 4x the number of online CPUs).
Halting the promotion of additional interactive tasks allows to
prioritize those already classified as interactive, thereby preventing
potential "bursts" of excessive interactive tasks in the system.
This refines the mitigation already provided by commit 640bd562
("scx_bpfland: prevent tasks from abusing interactive priority boost").
Fixes: 640bd562 ("scx_bpfland: prevent tasks from abusing interactive priority boost")
Signed-off-by: Andrea Righi <righi.andrea@gmail.com>
Always assign the maximum time slice if there are idle CPUs in the
system.
Otherwise, double the task's unused time slice to reward tasks that use
less CPU time and at the same time refill the time slice of the tasks
every time they're dispatched.
Signed-off-by: Andrea Righi <righi.andrea@gmail.com>
sched_ext is about to be merged upstream. There are some compatibility
breaking changes and we're making the current sched_ext/for-6.11
1edab907b57d ("sched_ext/scx_qmap: Pick idle CPU for direct dispatch on
!wakeup enqueues") the baseline.
Tag everything except scx_mitosis as 1.0.0. As scx_mitosis is still in early
development and is currently temporarily disabled, only the patchlevel is
bumped.
The priority boost for interactive tasks can be exploited to render the
system nearly unresponsive by creating numerous tasks that constantly
switch between wait/wakeup states.
For example, stress tests like `hackbench -l 10000` can significantly
degrade system responsiveness.
To mitigate this, limit the number of interactive tasks added to the
priority queue to 4x the number of online CPUs.
This simple approach appears to be a quite effective at identifying
potential spam of "fake" interactive tasks, while still prioritizing
legitimate interactive tasks.
Additionally, periodically refresh the interactive status of the tasks
based on their most recent average of voluntary context switches,
preventing the interactive status from being too "sticky".
Tested-by: Piotr Gorski <lucjan.lucjanov@gmail.com>
Signed-off-by: Andrea Righi <righi.andrea@gmail.com>
Avoid dispatching per-CPU kthreads directly, since this may cause
interactivity problems or unfairness, for example if there are too many
softirqs being scheduled (e.g., in presence of high RX network traffic
or when running certain stress tests, like hackbench).
Moreover, in order to help with testing and benchmarks, introduce the
option --local-kthread, that allows to restore the old behavior if
enabled.
Tested-by: Piotr Gorski <lucjan.lucjanov@gmail.com>
Signed-off-by: Andrea Righi <righi.andrea@gmail.com>
When updating the task vruntime, ensure the time slice delta is always a
positive value. Failing to do so may cause the global vruntime to
increase excessively due to overflows.
Tested-by: Piotr Gorski <lucjan.lucjanov@gmail.com>
Signed-off-by: Andrea Righi <righi.andrea@gmail.com>
Periodically report the amount of online CPUs to stdout.
The online CPUs are initially evaluated looking at the online cpumask,
then the value is updated in the .cpu_offline() / .cpu_online()
callbacks.
Tested-by: Piotr Gorski <lucjan.lucjanov@gmail.com>
Signed-off-by: Andrea Righi <righi.andrea@gmail.com>
Keep track of the CPUs that are running interactive tasks and report
their amount to stdout.
Tested-by: Piotr Gorski <lucjan.lucjanov@gmail.com>
Signed-off-by: Andrea Righi <righi.andrea@gmail.com>
The correct default value of slice_ns 5ms, not 5s.
This change doesn't really make any difference in practice, since these
values are changed by the Rust part when the scheduler is started, but
it's good to keep this aligned to the proper values for consistency.
Tested-by: Piotr Gorski <lucjan.lucjanov@gmail.com>
Signed-off-by: Andrea Righi <righi.andrea@gmail.com>
Every time we need to dispatch a task re-evalate its time slice as:
(unused_time_slice + min_time_slice) / 2
This allows to refill the time slice for tasks that haven't used much of
their previously assigned time, improving fairness.
Signed-off-by: Andrea Righi <andrea.righi@canonical.com>
Make sure to always classify interactive tasks, even when the system is
not fully utilized. This ensures that if the system suddenly becomes
overloaded, we already know which tasks need to be dispatched to the
priority DSQ.
Signed-off-by: Andrea Righi <andrea.righi@canonical.com>
Tasks are consumed from various DSQs in the following order:
per-CPU DSQs => priority DSQ => shared DSQ
Tasks in the shared DSQ may be starved by those in the priority DSQ,
which in turn may be starved by tasks dispatched to any per-CPU DSQ.
To mitigate this, record the timestamp of the last task scheduling event
both from the priority DSQ and the shared DSQ.
If the starvation threshold is exceeded without consuming a task, the
scheduler will be forced to consume a task from the corresponding DSQ.
The starvation threshold can be adjusted using the --starvation-thresh
command line parameter (default is 5ms).
Signed-off-by: Andrea Righi <andrea.righi@canonical.com>
There is no need to RCU protect the cpumask for the offline CPUs: it is
created once when the scheduler is initialized and it's never
deallocated.
Signed-off-by: Andrea Righi <andrea.righi@canonical.com>
Reduce the default time slice down to 5ms for a faster reaction and
system responsiveness when the system is overcomissioned.
This also helps to provide a more predictable level of performance.
Signed-off-by: Andrea Righi <andrea.righi@canonical.com>
Always use direct CPU dispatch for kthreads, there is no need to treat
kthreads in a special way, simply reuse direct CPU dispatch to
prioritize them.
Moreover, change direct CPU dispatches to use scx_bpf_dispatch_vtime(),
since we may dispatch multiple tasks to the same per-CPU DSQ now.
Signed-off-by: Andrea Righi <andrea.righi@canonical.com>
Small refactoring of the idle CPU selection logic:
- optimize idle CPU selection for tasks that can run on a single CPU
- drop the built-in idle selection policy and completely rely on the
custom one
Signed-off-by: Andrea Righi <andrea.righi@canonical.com>
We are incorrectly using the SMT idle cpumask to find any idle CPU, fix
by using the generic idle cpumask.
Signed-off-by: Andrea Righi <andrea.righi@canonical.com>
Implement CPU hotplugging in scx_bpfland without restarting the
scheduler.
The idle selection logic has been updated to consider online CPUs.
Additionally, a cpumask for offline CPUs has been introduced. Tasks
that have been dispatched to the DSQs associated with offline CPUs are
consumed by the other CPUs that are still online.
Moreover, the dependency on the Topology crate is temporarily dropped
and instead, /sys/devices/system/cpu/smt/active is used to determine if
SMT should be taken into account during idle selection. The Topology
crate will be re-introduced later when scx_bpfland will gain more
topology-aware capabilities.
This fixes#406.
Signed-off-by: Andrea Righi <andrea.righi@canonical.com>
Introduce a tunable to set a limit of the minimum vruntime that is used
when a task is dispatched, as:
vtime_min = vtime_now - slice_lag_ns
Increasing the time slice lag can make interactive tasks even more
responsive at the cost of starving regular and newly created tasks.
Default time slice lag is 0.
Signed-off-by: Andrea Righi <andrea.righi@canonical.com>
Overview
========
This scheduler is derived from scx_rustland, but it is fully implemented
in BFP with minimal user-space Rust part to process command line
options, collect metrics and logs out scheduling statistics.
Unlike scx_rustland, all scheduling decisions are made by the BPF
component.
Motivation
==========
The primary goal of this scheduler is to act as a performance baseline
for comparison with scx_rustland, allowing for a better assessment of
the overhead caused by kernel/user-space interactions.
It can also be used to deploy prototypes initially tested in the
scx_rustland scheduler. In fact, this scheduler is expected to
outperform scx_rustland, due to the elimitation of the kernel/user-space
overhead.
Scheduling policy
=================
scx_bpfland is a vruntime-based sched_ext scheduler that prioritizes
interactive workloads. Its scheduling policy closely mirrors
scx_rustland, but it has been re-implemented in BPF with some small
adjustments.
Tasks are categorized as either interactive or regular based on their
average rate of voluntary context switches per second: tasks that exceed
a specific voluntary context switch threshold are classified as
interactive.
Interactive tasks are prioritized in a higher-priority DSQ, while
regular tasks are placed in a lower-priority DSQ. Within each queue,
tasks are sorted based on their weighted runtime, using the built-in scx
vtime ordering capabilities (scx_bpf_dispatch_vtime()).
Moreover, each task gets a time slice budget. When a task is dispatched,
it receives a time slice equivalent to the remaining unused portion of
its previously allocated time slice (with a minimum threshold applied).
This gives latency-sensitive workloads more chances to exceed their time
slice when needed to perform short bursts of CPU activity without being
interrupted (i.e., real-time audio encoding / decoding workloads).
Results
=======
According to the initial test results, using the same benchmark "playing
a videogame while recompiling the kernel", this scheduler seems to
provide a +5% improvement in the frames-per-second (fps) compared to
scx_rustland, with video games such as Cyberpunk 2077, Counter-Strike 2
and Baldur's Gate 3.
Initial test results indicate that this scheduler offers around a +5%
improvement in frames-per-second (fps) compared to scx_rustland when
using the benchmark "playing a video game while recompiling the kernel".
This improvement was observed in games such as Cyberpunk 2077,
Counter-Strike 2, and Baldur's Gate 3.
Signed-off-by: Andrea Righi <andrea.righi@canonical.com>
