Library rt.implementation.apa.schedule
Require Import rt.util.all.
Require Import rt.model.priority.
Require Import rt.model.arrival.basic.arrival_sequence rt.model.arrival.basic.task.
Require Import rt.model.schedule.global.basic.schedule.
Require Import rt.model.schedule.apa.affinity rt.model.schedule.apa.platform.
Require Import rt.model.schedule.global.transformation.construction.
Module ConcreteScheduler.
Import SporadicTaskset ArrivalSequence Schedule Platform Priority Affinity ScheduleConstruction.
(* In this section, we implement a concrete weak APA scheduler. *)
Section Implementation.
Context {Job: eqType}.
Context {sporadic_task: eqType}.
Variable job_arrival: Job → time.
Variable job_cost: Job → time.
Variable job_task: Job → sporadic_task.
(* Let num_cpus denote the number of processors, ...*)
Variable num_cpus: nat.
(* ... and let arr_seq be any arrival sequence.*)
Variable arr_seq: arrival_sequence Job.
(* Let alpha be an affinity associated with each task. *)
Variable alpha: task_affinity sporadic_task num_cpus.
(* Assume a JLDP policy is given. *)
Variable higher_eq_priority: JLDP_policy Job.
(* Next, we show how to recursively construct the schedule. *)
Section ScheduleConstruction.
(* For any time t, suppose that we have generated the schedule prefix in the
interval [0, t). Then, we must define what should be scheduled at time t. *)
Variable sched_prefix: schedule Job num_cpus.
Variable cpu: processor num_cpus.
Variable t: time.
(* For simplicity, let's use some local names. *)
Let is_pending := pending job_arrival job_cost sched_prefix.
Let actual_arrivals := jobs_arrived_up_to arr_seq.
(* Consider the list of pending jobs at time t, ... *)
Definition pending_jobs := [seq j <- actual_arrivals t | is_pending j t].
(* ...which we sort by priority. *)
Definition sorted_pending_jobs := sort (higher_eq_priority t) pending_jobs.
(* Now we implement the algorithm that generates the APA schedule. *)
(* Given a job j at time t, we first define a predicate that states
whether j should preempt a mapping (cpu, x), where x is either Some j'
that is currently mapped to cpu or None. *)
Definition should_be_scheduled (j: Job) p :=
let '(cpu, mapped_job) := p in
if mapped_job is Some j' then (* If there is a job j', check the priority and affinity. *)
(can_execute_on alpha (job_task j) cpu) &&
~~ (higher_eq_priority t j' j)
else (* Else, if cpu is idle, check only the affinity. *)
(can_execute_on alpha (job_task j) cpu).
(* Next, using the "should_be_scheduled" predicate, we define a function
that tries to schedule job j by updating a list of mappings.
It does so by replacing the first pair (cpu, x) where j can be
scheduled (if it exists). *)
Definition update_available_cpu allocation j :=
replace_first (should_be_scheduled j) (* search for processors that j can preempt *)
(set_pair_2nd (Some j)) (* replace the mapping in that processor with j *)
allocation. (* list of mappings *)
(* Consider the empty mapping. *)
Let empty_mapping : seq (processor num_cpus × option Job) :=
(zip (enum (processor num_cpus)) (nseq num_cpus None)).
(* Using the fuction "update_available_cpu", we now define an iteration
that iteratively maps each pending job to a processor.
Starting with an empty mapping,
<(cpu0, None), (cpu1, None), (cpu2, None), ...>,
it tries to schedule each job on some processor and yields an updated list:
<(cpu0, None), (cpu1, Some j5), (cpu2, Some j9), ...>. *)
Definition schedule_jobs_from_list l :=
foldl update_available_cpu empty_mapping l.
(* To conclude, we take the list of pairs and convert to a function denoting
the actual schedule. *)
Definition apa_schedule :=
pairs_to_function None (schedule_jobs_from_list sorted_pending_jobs) cpu.
End ScheduleConstruction.
(* Starting from the empty schedule, the final schedule is obtained by iteratively
picking the highest-priority job. *)
Let empty_schedule : schedule Job num_cpus := fun cpu t ⇒ None.
Definition scheduler :=
build_schedule_from_prefixes num_cpus apa_schedule empty_schedule.
(* Then, by showing that the construction function depends only on the prefix, ... *)
Lemma scheduler_depends_only_on_prefix:
∀ sched1 sched2 cpu t,
(∀ t0 cpu0, t0 < t → sched1 cpu0 t0 = sched2 cpu0 t0) →
apa_schedule sched1 cpu t = apa_schedule sched2 cpu t.
(* ...we infer that the generated schedule is indeed based on the construction function. *)
Corollary scheduler_uses_construction_function:
∀ t cpu, scheduler cpu t = apa_schedule scheduler cpu t.
End Implementation.
(* In this section, we prove several properties about the scheduling algorithm we
implemented. For example, we show that it is APA work conserving. *)
Section Proofs.
Context {Job: eqType}.
Context {sporadic_task: eqType}.
Variable job_arrival: Job → time.
Variable job_cost: Job → time.
Variable job_task: Job → sporadic_task.
(* Assume a positive number of processors. *)
Variable num_cpus: nat.
Hypothesis H_at_least_one_cpu: num_cpus > 0.
(* Let alpha be an affinity associated with each task. *)
Variable alpha: task_affinity sporadic_task num_cpus.
(* Let arr_seq be any job arrival sequence with consistent, duplicate-free arrivals. *)
Variable arr_seq: arrival_sequence Job.
Hypothesis H_arrival_times_are_consistent: arrival_times_are_consistent job_arrival arr_seq.
Hypothesis H_arrival_sequence_is_a_set: arrival_sequence_is_a_set arr_seq.
(* Consider any JLDP policy higher_eq_priority that is transitive and total. *)
Variable higher_eq_priority: JLDP_policy Job.
Hypothesis H_priority_transitive: JLDP_is_transitive higher_eq_priority.
Hypothesis H_priority_total: ∀ t, total (higher_eq_priority t).
(* Let sched denote our concrete scheduler implementation. *)
Let sched := scheduler job_arrival job_cost job_task num_cpus arr_seq alpha higher_eq_priority.
(* Next, we provide some helper lemmas about the scheduler construction. *)
Section HelperLemmas.
(* To avoid many parameters, let's also rename the scheduling function.
We make a generic version (for any list of jobs l), ... *)
Let schedule_jobs t l := schedule_jobs_from_list job_task num_cpus alpha higher_eq_priority t l.
(* ... and a specific version (for the pending jobs in sched). *)
Let schedule_pending_jobs t :=
schedule_jobs t (sorted_pending_jobs job_arrival job_cost num_cpus arr_seq
higher_eq_priority sched t).
(* Next, we show that there are no duplicate cpus in the mapping. *)
Lemma scheduler_uniq_cpus :
∀ t l,
uniq (unzip1 (schedule_jobs t l)).
(* Next, we show that if a job j is in the mapping, then j must come from the list
of jobs l used in the construction. *)
Lemma scheduler_job_in_mapping :
∀ l j t cpu,
(cpu, Some j) \in schedule_jobs t l → j \in l.
(* Next, we prove that if a pair (cpu, j) is in the mapping, then
cpu must be part of j's affinity. *)
Lemma scheduler_mapping_respects_affinity :
∀ j t cpu,
(cpu, Some j) \in schedule_pending_jobs t →
can_execute_on alpha (job_task j) cpu.
(* Next, we show that the mapping does not schedule the same job j in two
different cpus. *)
Lemma scheduler_has_no_duplicate_jobs :
∀ j t cpu1 cpu2,
(cpu1, Some j) \in schedule_pending_jobs t →
(cpu2, Some j) \in schedule_pending_jobs t →
cpu1 = cpu2.
(* Based on the definition of the schedule, a job j is scheduled on cpu
iff (cpu, Some j) is the final mapping. *)
Lemma scheduler_scheduled_on :
∀ j cpu t,
scheduled_on sched j cpu t = ((cpu, Some j) \in schedule_pending_jobs t).
(* Now we show that for every cpu, there is always a pair in the mapping. *)
Lemma scheduler_has_cpus :
∀ cpu t l,
∃ x,
(cpu, x) \in schedule_jobs t l.
(* Next, consider a list of jobs l that is sorted by priority and does not have
duplicates.
We prove that for any job j in l, if j is not scheduled at time t,
then every cpu in j's affinity has some job mapped at time t. *)
Lemma scheduler_mapping_is_work_conserving :
∀ j cpu t l,
j \in l →
sorted (higher_eq_priority t) l →
uniq l →
(∀ cpu, (cpu, Some j) \notin schedule_jobs t l) →
can_execute_on alpha (job_task j) cpu →
∃ j_other,
(cpu, Some j_other) \in schedule_jobs t l.
(* Next, we prove that the mapping respects priority. *)
Lemma scheduler_priority :
∀ j j_hp cpu t,
arrives_in arr_seq j →
backlogged job_arrival job_cost sched j t →
can_execute_on alpha (job_task j) cpu →
scheduled_on sched j_hp cpu t →
higher_eq_priority t j_hp j.
End HelperLemmas.
(* Now, we prove the important properties about the implementation. *)
(* First, we show that scheduled jobs come from the arrival sequence. *)
Lemma scheduler_jobs_come_from_arrival_sequence:
jobs_come_from_arrival_sequence sched arr_seq.
(* Jobs do not execute before they arrive, ...*)
Theorem scheduler_jobs_must_arrive_to_execute:
jobs_must_arrive_to_execute job_arrival sched.
(* ..., jobs are sequential, ... *)
Theorem scheduler_sequential_jobs: sequential_jobs sched.
(* ... and jobs do not execute after completion. *)
Theorem scheduler_completed_jobs_dont_execute:
completed_jobs_dont_execute job_cost sched.
(* In addition, the scheduler is APA work conserving, ... *)
Theorem scheduler_apa_work_conserving:
apa_work_conserving job_arrival job_cost job_task arr_seq sched alpha.
(* ..., respects affinities, ... *)
Theorem scheduler_respects_affinity:
respects_affinity job_task sched alpha.
(* ... and respects the JLDP policy under weak APA scheduling. *)
Theorem scheduler_respects_policy :
respects_JLDP_policy_under_weak_APA job_arrival job_cost job_task arr_seq
sched alpha higher_eq_priority.
End Proofs.
End ConcreteScheduler.
Require Import rt.model.priority.
Require Import rt.model.arrival.basic.arrival_sequence rt.model.arrival.basic.task.
Require Import rt.model.schedule.global.basic.schedule.
Require Import rt.model.schedule.apa.affinity rt.model.schedule.apa.platform.
Require Import rt.model.schedule.global.transformation.construction.
Module ConcreteScheduler.
Import SporadicTaskset ArrivalSequence Schedule Platform Priority Affinity ScheduleConstruction.
(* In this section, we implement a concrete weak APA scheduler. *)
Section Implementation.
Context {Job: eqType}.
Context {sporadic_task: eqType}.
Variable job_arrival: Job → time.
Variable job_cost: Job → time.
Variable job_task: Job → sporadic_task.
(* Let num_cpus denote the number of processors, ...*)
Variable num_cpus: nat.
(* ... and let arr_seq be any arrival sequence.*)
Variable arr_seq: arrival_sequence Job.
(* Let alpha be an affinity associated with each task. *)
Variable alpha: task_affinity sporadic_task num_cpus.
(* Assume a JLDP policy is given. *)
Variable higher_eq_priority: JLDP_policy Job.
(* Next, we show how to recursively construct the schedule. *)
Section ScheduleConstruction.
(* For any time t, suppose that we have generated the schedule prefix in the
interval [0, t). Then, we must define what should be scheduled at time t. *)
Variable sched_prefix: schedule Job num_cpus.
Variable cpu: processor num_cpus.
Variable t: time.
(* For simplicity, let's use some local names. *)
Let is_pending := pending job_arrival job_cost sched_prefix.
Let actual_arrivals := jobs_arrived_up_to arr_seq.
(* Consider the list of pending jobs at time t, ... *)
Definition pending_jobs := [seq j <- actual_arrivals t | is_pending j t].
(* ...which we sort by priority. *)
Definition sorted_pending_jobs := sort (higher_eq_priority t) pending_jobs.
(* Now we implement the algorithm that generates the APA schedule. *)
(* Given a job j at time t, we first define a predicate that states
whether j should preempt a mapping (cpu, x), where x is either Some j'
that is currently mapped to cpu or None. *)
Definition should_be_scheduled (j: Job) p :=
let '(cpu, mapped_job) := p in
if mapped_job is Some j' then (* If there is a job j', check the priority and affinity. *)
(can_execute_on alpha (job_task j) cpu) &&
~~ (higher_eq_priority t j' j)
else (* Else, if cpu is idle, check only the affinity. *)
(can_execute_on alpha (job_task j) cpu).
(* Next, using the "should_be_scheduled" predicate, we define a function
that tries to schedule job j by updating a list of mappings.
It does so by replacing the first pair (cpu, x) where j can be
scheduled (if it exists). *)
Definition update_available_cpu allocation j :=
replace_first (should_be_scheduled j) (* search for processors that j can preempt *)
(set_pair_2nd (Some j)) (* replace the mapping in that processor with j *)
allocation. (* list of mappings *)
(* Consider the empty mapping. *)
Let empty_mapping : seq (processor num_cpus × option Job) :=
(zip (enum (processor num_cpus)) (nseq num_cpus None)).
(* Using the fuction "update_available_cpu", we now define an iteration
that iteratively maps each pending job to a processor.
Starting with an empty mapping,
<(cpu0, None), (cpu1, None), (cpu2, None), ...>,
it tries to schedule each job on some processor and yields an updated list:
<(cpu0, None), (cpu1, Some j5), (cpu2, Some j9), ...>. *)
Definition schedule_jobs_from_list l :=
foldl update_available_cpu empty_mapping l.
(* To conclude, we take the list of pairs and convert to a function denoting
the actual schedule. *)
Definition apa_schedule :=
pairs_to_function None (schedule_jobs_from_list sorted_pending_jobs) cpu.
End ScheduleConstruction.
(* Starting from the empty schedule, the final schedule is obtained by iteratively
picking the highest-priority job. *)
Let empty_schedule : schedule Job num_cpus := fun cpu t ⇒ None.
Definition scheduler :=
build_schedule_from_prefixes num_cpus apa_schedule empty_schedule.
(* Then, by showing that the construction function depends only on the prefix, ... *)
Lemma scheduler_depends_only_on_prefix:
∀ sched1 sched2 cpu t,
(∀ t0 cpu0, t0 < t → sched1 cpu0 t0 = sched2 cpu0 t0) →
apa_schedule sched1 cpu t = apa_schedule sched2 cpu t.
(* ...we infer that the generated schedule is indeed based on the construction function. *)
Corollary scheduler_uses_construction_function:
∀ t cpu, scheduler cpu t = apa_schedule scheduler cpu t.
End Implementation.
(* In this section, we prove several properties about the scheduling algorithm we
implemented. For example, we show that it is APA work conserving. *)
Section Proofs.
Context {Job: eqType}.
Context {sporadic_task: eqType}.
Variable job_arrival: Job → time.
Variable job_cost: Job → time.
Variable job_task: Job → sporadic_task.
(* Assume a positive number of processors. *)
Variable num_cpus: nat.
Hypothesis H_at_least_one_cpu: num_cpus > 0.
(* Let alpha be an affinity associated with each task. *)
Variable alpha: task_affinity sporadic_task num_cpus.
(* Let arr_seq be any job arrival sequence with consistent, duplicate-free arrivals. *)
Variable arr_seq: arrival_sequence Job.
Hypothesis H_arrival_times_are_consistent: arrival_times_are_consistent job_arrival arr_seq.
Hypothesis H_arrival_sequence_is_a_set: arrival_sequence_is_a_set arr_seq.
(* Consider any JLDP policy higher_eq_priority that is transitive and total. *)
Variable higher_eq_priority: JLDP_policy Job.
Hypothesis H_priority_transitive: JLDP_is_transitive higher_eq_priority.
Hypothesis H_priority_total: ∀ t, total (higher_eq_priority t).
(* Let sched denote our concrete scheduler implementation. *)
Let sched := scheduler job_arrival job_cost job_task num_cpus arr_seq alpha higher_eq_priority.
(* Next, we provide some helper lemmas about the scheduler construction. *)
Section HelperLemmas.
(* To avoid many parameters, let's also rename the scheduling function.
We make a generic version (for any list of jobs l), ... *)
Let schedule_jobs t l := schedule_jobs_from_list job_task num_cpus alpha higher_eq_priority t l.
(* ... and a specific version (for the pending jobs in sched). *)
Let schedule_pending_jobs t :=
schedule_jobs t (sorted_pending_jobs job_arrival job_cost num_cpus arr_seq
higher_eq_priority sched t).
(* Next, we show that there are no duplicate cpus in the mapping. *)
Lemma scheduler_uniq_cpus :
∀ t l,
uniq (unzip1 (schedule_jobs t l)).
(* Next, we show that if a job j is in the mapping, then j must come from the list
of jobs l used in the construction. *)
Lemma scheduler_job_in_mapping :
∀ l j t cpu,
(cpu, Some j) \in schedule_jobs t l → j \in l.
(* Next, we prove that if a pair (cpu, j) is in the mapping, then
cpu must be part of j's affinity. *)
Lemma scheduler_mapping_respects_affinity :
∀ j t cpu,
(cpu, Some j) \in schedule_pending_jobs t →
can_execute_on alpha (job_task j) cpu.
(* Next, we show that the mapping does not schedule the same job j in two
different cpus. *)
Lemma scheduler_has_no_duplicate_jobs :
∀ j t cpu1 cpu2,
(cpu1, Some j) \in schedule_pending_jobs t →
(cpu2, Some j) \in schedule_pending_jobs t →
cpu1 = cpu2.
(* Based on the definition of the schedule, a job j is scheduled on cpu
iff (cpu, Some j) is the final mapping. *)
Lemma scheduler_scheduled_on :
∀ j cpu t,
scheduled_on sched j cpu t = ((cpu, Some j) \in schedule_pending_jobs t).
(* Now we show that for every cpu, there is always a pair in the mapping. *)
Lemma scheduler_has_cpus :
∀ cpu t l,
∃ x,
(cpu, x) \in schedule_jobs t l.
(* Next, consider a list of jobs l that is sorted by priority and does not have
duplicates.
We prove that for any job j in l, if j is not scheduled at time t,
then every cpu in j's affinity has some job mapped at time t. *)
Lemma scheduler_mapping_is_work_conserving :
∀ j cpu t l,
j \in l →
sorted (higher_eq_priority t) l →
uniq l →
(∀ cpu, (cpu, Some j) \notin schedule_jobs t l) →
can_execute_on alpha (job_task j) cpu →
∃ j_other,
(cpu, Some j_other) \in schedule_jobs t l.
(* Next, we prove that the mapping respects priority. *)
Lemma scheduler_priority :
∀ j j_hp cpu t,
arrives_in arr_seq j →
backlogged job_arrival job_cost sched j t →
can_execute_on alpha (job_task j) cpu →
scheduled_on sched j_hp cpu t →
higher_eq_priority t j_hp j.
End HelperLemmas.
(* Now, we prove the important properties about the implementation. *)
(* First, we show that scheduled jobs come from the arrival sequence. *)
Lemma scheduler_jobs_come_from_arrival_sequence:
jobs_come_from_arrival_sequence sched arr_seq.
(* Jobs do not execute before they arrive, ...*)
Theorem scheduler_jobs_must_arrive_to_execute:
jobs_must_arrive_to_execute job_arrival sched.
(* ..., jobs are sequential, ... *)
Theorem scheduler_sequential_jobs: sequential_jobs sched.
(* ... and jobs do not execute after completion. *)
Theorem scheduler_completed_jobs_dont_execute:
completed_jobs_dont_execute job_cost sched.
(* In addition, the scheduler is APA work conserving, ... *)
Theorem scheduler_apa_work_conserving:
apa_work_conserving job_arrival job_cost job_task arr_seq sched alpha.
(* ..., respects affinities, ... *)
Theorem scheduler_respects_affinity:
respects_affinity job_task sched alpha.
(* ... and respects the JLDP policy under weak APA scheduling. *)
Theorem scheduler_respects_policy :
respects_JLDP_policy_under_weak_APA job_arrival job_cost job_task arr_seq
sched alpha higher_eq_priority.
End Proofs.
End ConcreteScheduler.