pull-kubernetes-scheduler-perf: use -benchtime=1x

Update scheduler benchmark parameters to use
-benchtime=1x instead of -benchtime=1ns to explicitly run benchmarks
one time.

Author: bart0sh

config/jobs/kubernetes/sig-scheduling/sig-scheduling-config.yaml

@@ -80,7 +80,7 @@ presubmits:
         args:
         - /bin/sh
         - -c
-        - ./hack/install-etcd.sh && PATH=${PWD}/third_party/etcd:${PATH} make test-integration WHAT=./test/integration/scheduler_perf/... KUBE_CACHE_MUTATION_DETECTOR=false KUBE_TIMEOUT=--timeout=3h55m ETCD_LOGLEVEL=warn KUBE_TEST_ARGS="-run=nothing-because-no-test-has-this-name -benchtime=1ns -bench=BenchmarkPerfScheduling -data-items-dir=${ARTIFACTS}" SHORT='--short=false' KUBE_PRUNE_JUNIT_TESTS=false
+        - ./hack/install-etcd.sh && PATH=${PWD}/third_party/etcd:${PATH} make test-integration WHAT=./test/integration/scheduler_perf/... KUBE_CACHE_MUTATION_DETECTOR=false KUBE_TIMEOUT=--timeout=3h55m ETCD_LOGLEVEL=warn KUBE_TEST_ARGS="-run=nothing-because-no-test-has-this-name -benchtime=1x -bench=BenchmarkPerfScheduling -data-items-dir=${ARTIFACTS}" SHORT='--short=false' KUBE_PRUNE_JUNIT_TESTS=false
         # We need to constraint compute resources so all the tests
         # finish approximately at the same time. More compute power
         # can increase scheduling throughput and make consequent results