On Fri, 30 Jun 2023, Tomas Glozar wrote:
> Allows correct calculation of make job count that does not include
> isolated CPUs, on which the loads won't be scheduled.
>
> Signed-off-by: Tomas Glozar <tglozar@xxxxxxxxxx>
> ---
> rteval/modules/loads/kcompile.py | 8 ++++++--
> 1 file changed, 6 insertions(+), 2 deletions(-)
>
> diff --git a/rteval/modules/loads/kcompile.py b/rteval/modules/loads/kcompile.py
> index 35ee5cb..30c5965 100644
> --- a/rteval/modules/loads/kcompile.py
> +++ b/rteval/modules/loads/kcompile.py
> @@ -37,6 +37,7 @@ from rteval.systopology import CpuList, SysTopology
>
> expand_cpulist = CpuList.expand_cpulist
> compress_cpulist = CpuList.compress_cpulist
> +nonisolated_cpulist = CpuList.nonisolated_cpulist
>
> DEFAULT_KERNEL_PREFIX = "linux-6.1"
>
> @@ -55,17 +56,20 @@ class KBuildJob:
> if not os.path.isdir(self.objdir):
> os.mkdir(self.objdir)
>
> + # Exclude isolated CPUs if cpulist not set
> + cpus_available = len(nonisolated_cpulist(self.node.cpus.cpulist))
> +
> if os.path.exists('/usr/bin/numactl') and not cpulist:
> # Use numactl
> self.binder = f'numactl --cpunodebind {int(self.node)}'
> - self.jobs = self.calc_jobs_per_cpu() * len(self.node)
> + self.jobs = self.calc_jobs_per_cpu() * cpus_available
> elif cpulist:
> # Use taskset
> self.jobs = self.calc_jobs_per_cpu() * len(cpulist)
> self.binder = f'taskset -c {compress_cpulist(cpulist)}'
> else:
> # Without numactl calculate number of jobs from the node
> - self.jobs = self.calc_jobs_per_cpu() * len(self.node)
> + self.jobs = self.calc_jobs_per_cpu() * cpus_available
>
> self.runcmd = f"make O={self.objdir} -C {self.kdir} -j{self.jobs}"
> self.cleancmd = f"make O={self.objdir} -C {self.kdir} clean allmodconfig"
> --
> 2.41.0
>
>
Signed-off-by: John Kacur <jkacur@xxxxxxxxxx>
