add manually triggered benchmark action (#17088)

This is a first step towards reenabling UR performance testing CI. This
introduces the reusable yml workflow and a way to trigger it manually.

Here's an example how it looks:
pbalcer/llvm#2 (comment)

Author: pbalcer

scripts/benchmarks/benches/oneapi.py

@@ -11,7 +11,7 @@
 
 class OneAPI:
     # random unique number for benchmark oneAPI installation
-    ONEAPI_BENCHMARK_INSTANCE_ID = 98765
+    ONEAPI_BENCHMARK_INSTANCE_ID = 987654
 
     def __init__(self):
         self.oneapi_dir = os.path.join(options.workdir, "oneapi")

scripts/benchmarks/main.py

@@ -262,9 +262,7 @@ def main(directory, additional_env_vars, save_name, compare_names, filter):
             compare_names.append(saved_name)
 
     if options.output_html:
-        html_content = generate_html(
-            history.runs, "oneapi-src/unified-runtime", compare_names
-        )
+        html_content = generate_html(history.runs, "intel/llvm", compare_names)
 
         with open("benchmark_results.html", "w") as file:
             file.write(html_content)

scripts/benchmarks/options.py

@@ -37,7 +37,7 @@ class Options:
     build_compute_runtime: bool = False
     extra_ld_libraries: list[str] = field(default_factory=list)
     extra_env_vars: dict = field(default_factory=dict)
-    compute_runtime_tag: str = "24.52.32224.10"
+    compute_runtime_tag: str = "25.05.32567.12"
     build_igc: bool = False
     current_run_name: str = "This PR"
 

scripts/benchmarks/requirements.txt

@@ -0,0 +1,4 @@
+matplotlib==3.9.2
+mpld3==0.5.10
+dataclasses-json==0.6.7
+PyYAML==6.0.1