From 22f8bcd0d6d1f503083c0ba680ed0b2759def201 Mon Sep 17 00:00:00 2001 From: Bobbins228 Date: Wed, 14 Feb 2024 10:31:18 +0000 Subject: [PATCH 1/2] Added mcad bool for Job submission with no cluster --- src/codeflare_sdk/job/jobs.py | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-) diff --git a/src/codeflare_sdk/job/jobs.py b/src/codeflare_sdk/job/jobs.py index 655107df..174e0257 100644 --- a/src/codeflare_sdk/job/jobs.py +++ b/src/codeflare_sdk/job/jobs.py @@ -69,6 +69,7 @@ def __init__( scheduler_args: Optional[Dict[str, str]] = None, image: Optional[str] = None, workspace: Optional[str] = f"file://{Path.cwd()}", + mcad: Optional[str] = False, ): if bool(script) == bool(m): # logical XOR raise ValueError( @@ -93,6 +94,7 @@ def __init__( ) self.image = image self.workspace = workspace + self.mcad = mcad def _dry_run(self, cluster: "Cluster"): j = f"{cluster.config.num_workers}x{max(cluster.config.num_gpus, 1)}" # # of proc. = # of gpus @@ -136,6 +138,9 @@ def _dry_run_no_cluster(self): if self.scheduler_args is not None: if self.scheduler_args.get("namespace") is None: self.scheduler_args["namespace"] = get_current_namespace() + scheduler = "kueue_job" + if self.mcad == True: + scheduler = "kubernetes_mcad" runner = get_runner() return ( runner.dryrun( @@ -172,7 +177,7 @@ def _dry_run_no_cluster(self): if self.image is not None else self._missing_spec("image"), ), - scheduler="kubernetes_mcad", + scheduler=scheduler, cfg=self.scheduler_args, workspace="", ), From cbba3bba337288962bd1abe4619088e3a6bb31c4 Mon Sep 17 00:00:00 2001 From: Bobbins228 Date: Mon, 25 Mar 2024 12:20:24 +0000 Subject: [PATCH 2/2] Updated scheduler name --- src/codeflare_sdk/job/jobs.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/codeflare_sdk/job/jobs.py b/src/codeflare_sdk/job/jobs.py index 174e0257..73a95e98 100644 --- a/src/codeflare_sdk/job/jobs.py +++ b/src/codeflare_sdk/job/jobs.py @@ -138,7 +138,7 @@ def _dry_run_no_cluster(self): if self.scheduler_args is not None: if self.scheduler_args.get("namespace") is None: self.scheduler_args["namespace"] = get_current_namespace() - scheduler = "kueue_job" + scheduler = "kueue" if self.mcad == True: scheduler = "kubernetes_mcad" runner = get_runner()