[RFC PATCH 0/9 v4] A new CPU load metric for power-efficient scheduler: CPU ConCurrency

[Yuyang Du]

The current schedulerâs load balancing is completely work-conserving. In some
workload, generally low CPU utilization but immersed with CPU bursts of
transient tasks, migrating task to engage all available CPUs for
work-conserving can lead to significant overhead: cache locality loss,
idle/active HW state transitional latency and power, shallower idle state,
etc, which are both power and performance inefficient especially for todayâs
low power processors in mobile. 

This RFC introduces a sense of idleness-conserving into work-conserving (by
all means, we really donât want to be overwhelming in only one way). But to
what extent the idleness-conserving should be, bearing in mind that we donât
want to sacrifice performance? We first need a load/idleness indicator to that
end.

Thanks to CFSâs âmodel an ideal, precise multi-tasking CPUâ, tasks can be seen
as concurrently running (the tasks in the runqueue). So it is natural to use
task concurrency as load indicator. Having said that, we do two things:

1) Divide continuous time into periods of time, and average task concurrency
in period, for tolerating the transient bursts:
a = sum(concurrency * time) / period
2) Exponentially decay past periods, and synthesize them all, for hysteresis
to load drops or resilience to load rises (let f be decaying factor, and a_x
the xth period average since period 0):
s = a_n + f^1 * a_n-1 + f^2 * a_n-2 +, ..., + f^(n-1) * a_1 + f^n * a_0

We name this load indicator as CPU ConCurrency (CC): task concurrency
determines how many CPUs are needed to be running concurrently.

Another two ways of how to interpret CC:

1) the current work-conserving load balance also uses CC, but instantaneous
CC.

2) CC vs. CPU utilization. CC is runqueue-length-weighted CPU utilization. If
we change: "a = sum(concurrency * time) / period" to "a' = sum(1 * time) /
period". Then a' is just about the CPU utilization. And the way we weight
runqueue-length is the simplest one (excluding the exponential decays, and you
may have other ways).

To track CC, we intercept the scheduler in 1) enqueue, 2) dequeue, 3)
scheduler tick, and 4) enter/exit idle.

After CC, in the consolidation part, we do 1) attach the CPU topology to be
adaptive beyond our experimental platforms, and 2) intercept the current load
balance for load and load balancing containment.

Currently, CC is per CPU. To consolidate, the formula is based on a heuristic.
Suppose we have 2 CPUs, their task concurrency over time is ('-' means no
task, 'x' having tasks):

1)
CPU0: ---xxxx---------- (CC[0])
CPU1: ---------xxxx---- (CC[1])

2)
CPU0: ---xxxx---------- (CC[0])
CPU1: ---xxxx---------- (CC[1])

If we consolidate CPU0 and CPU1, the consolidated CC will be: CC' = CC[0] +
CC[1] for case 1 and CC'' = (CC[0] + CC[1]) * 2 for case 2. For the cases in
between case 1 and 2 in terms of how xxx overlaps, the CC should be between
CC' and CC''. So, we uniformly use this condition for consolidation (suppose
we consolidate m CPUs to n CPUs, m > n):

(CC[0] + CC[1] + ... + CC[m-2] + CC[m-1]) * (n + log(m-n)) >=<? (1 * n) * n *
consolidating_coefficient

The consolidating_coefficient could be like 100% or more or less.

By CC, we implemented a Workload Consolidation (WC) patch on two Intel mobile
platforms (a quad-core composed of two dual-core modules): contain load and
load balancing in the first dual-core when aggregated CC low, and if not in
the full quad-core. Results show that we got power savings and no substantial
performance regression (even gains for some). The workloads we used to
evaluate the Workload Consolidation include 1) 50+ perf/ux benchmarks (almost
all of the magazine ones), and 2) ~10 power workloads, of course, they are the
easiest ones, such as browsing, audio, video, recording, imaging, etc.

v4:
- Reuse per task load average to calculate CC
- Enable SD_WORKLOAD_CONSOLIDATION in sched_domain initialization
- Reuse active_load_balance_cpu_stop

v3:
- Removed rq->avg first, and base our patch on it
- Removed all CONFIG_CPU_CONCURRENCY and CONFIG_WORKLOAD_CONSOLIDATION
- CPU CC will be updated mandatory
- CPU WC can be enabled/disabled by flags per domain level on the fly 
- CPU CC and WC is completely fair scheduler thing, don't touch RT anymore
 
v2:
- Data type defined in formation

Patchset against linux-next v3.16-rc2.

Yuyang Du (9):
  sched: Remove update_rq_runnable_avg
  sched: Precise accumulated time and acount runnable number in
    update_entity_runnable_avg
  How CPU ConCurrency (CC) accrues with runqueue change and time
  Define SD_WORKLOAD_CONSOLIDATION and attach to sched_domain
  Workload Consolidation: Consolidating workload to a subset of CPUs if
    possible
  Implement Workload Consolidation in wakeup/fork/exec
  Implement Workload Consolidation in idle_balance
  Implement Workload Consolidation in nohz_idle_balance
  Implement Workload Consolidation in periodic load balance

 include/linux/sched.h        |    9 +-
 include/linux/sched/sysctl.h |    4 +
 kernel/sched/core.c          |   46 +++-
 kernel/sched/debug.c         |    8 -
 kernel/sched/fair.c          |  576 +++++++++++++++++++++++++++++++++++++++---
 kernel/sched/sched.h         |   17 +-
 kernel/sysctl.c              |    9 +
 7 files changed, 612 insertions(+), 57 deletions(-)

-- 
1.7.9.5

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/