Try #380:

.buildkite/pipeline.yml

@@ -116,3 +116,11 @@ steps:
           queue: central
           slurm_cpus_per_task: 8
           slurm_mem_per_cpu: 6G
+
+      - label: "GPU benchmarks"
+        command: "julia --color=yes --project=gpuenv perf/benchmark.jl"
+        agents:
+          config: gpu
+          queue: central
+          slurm_ntasks: 1
+          slurm_gres: "gpu:1"

gpuenv/Manifest.toml

@@ -2,7 +2,7 @@
 
 julia_version = "1.8.5"
 manifest_format = "2.0"
-project_hash = "64f7dfb64711f57c4ddf26244a6acc5d6bb6c3aa"
+project_hash = "c2134fd3c36d974d4c4f215329ec4057c2e52fff"
 
 [[deps.AbstractFFTs]]
 deps = ["ChainRulesCore", "LinearAlgebra"]
@@ -44,6 +44,12 @@ version = "0.4.2"
 [[deps.Base64]]
 uuid = "2a0f44e3-6c83-55bd-87e4-b1978d98bd5f"
 
+[[deps.BenchmarkTools]]
+deps = ["JSON", "Logging", "Printf", "Profile", "Statistics", "UUIDs"]
+git-tree-sha1 = "d9a9701b899b30332bbcb3e1679c41cce81fb0e8"
+uuid = "6e4b80f9-dd63-53aa-95a3-0cdb28fa8baf"
+version = "1.3.2"
+
 [[deps.BitFlags]]
 git-tree-sha1 = "43b1a4a8f797c1cddadf60499a8a077d4af2cd2d"
 uuid = "d1d4a3ce-64b1-5f1a-9ba4-7e7e69966f35"
@@ -519,7 +525,7 @@ uuid = "3fa0cd96-eef1-5676-8a61-b3b8758bbffb"
 deps = ["Adapt", "CUDA", "ClimaComms", "DocStringExtensions", "GaussQuadrature", "Random", "StaticArrays"]
 path = ".."
 uuid = "a01a1ee8-cea4-48fc-987c-fc7878d79da1"
-version = "0.9.0"
+version = "0.9.1"
 
 [[deps.Random]]
 deps = ["SHA", "Serialization"]

gpuenv/Project.toml

@@ -1,6 +1,7 @@
 [deps]
 Adapt = "79e6a3ab-5dfb-504d-930d-738a2a938a0e"
 Aqua = "4c88cf16-eb10-579e-8560-4a9242c79595"
+BenchmarkTools = "6e4b80f9-dd63-53aa-95a3-0cdb28fa8baf"
 CLIMAParameters = "6eacf6c3-8458-43b9-ae03-caf5306d3d53"
 CUDA = "052768ef-5323-5732-b1bb-66c8b64840ba"
 ClimaComms = "3a4d1b5c-c61d-41fd-a00a-5873ba7a1b0d"