[jax_triton] Add parameter allowing user to compile for specific comp…

…ute capability.

PiperOrigin-RevId: 612431401

jax_triton/triton_lib.py

@@ -157,10 +157,8 @@ def compile_ttir_to_ptx_inplace(
     ttir,
     cuda_backend: cb.CUDABackend,
     cuda_options: cb.CUDAOptions,
-    device: int = 0,
-    device_type: str = "cuda",
-) -> Tuple[str, str, int, int, Any]:
-  compute_capability = triton_kernel_call_lib.get_compute_capability(device)
+    compute_capability,
+) -> Tuple[str, str, int, Any]:
   if cuda_options.debug:
     print(ttir)
   if isinstance(ttir, ir.Module):
@@ -213,7 +211,7 @@ def compile_ttir_to_ptx_inplace(
     print(ptx)
   name = metadata["name"]
   cluster_dims = metadata["cluster_dims"]
-  return ptx, name, shared_mem_bytes, compute_capability, cluster_dims
+  return ptx, name, shared_mem_bytes, cluster_dims
 
 
 _COMPILED_KERNEL_CACHE = {}  # TODO(cjfj): Convert to LRU cache?
@@ -227,22 +225,23 @@ def get_or_create_triton_kernel(
     num_warps,
     num_stages,
     num_ctas,
+    compute_capability,
     enable_fp_fusion,
     metaparams,
     dump: bool,
 ) -> Tuple[triton_kernel_call_lib.TritonKernel, Any]:
-  device_type = "cuda"
   if num_warps is None:
     num_warps = 4
-  # TODO(sharadmv): handle multiple devices, right now we assume device 0
-  # which is fine when we have multiple of the same GPU but this won't work in
-  # general.
-  device = 0
-  arch = triton_kernel_call_lib.get_compute_capability(device)
-  if num_ctas > 1 and arch < 90:
-    raise ValueError("num_ctas > 1 unsupported before Hopper.")
   if num_stages is None:
     num_stages = 3
+  if compute_capability is None:
+    # TODO(sharadmv): handle multiple devices, right now we assume device 0
+    # which is fine when we have multiple of the same GPU but this won't work in
+    # general.
+    device = 0
+    compute_capability = triton_kernel_call_lib.get_compute_capability(device)
+  if num_ctas > 1 and compute_capability < 90:
+    raise ValueError("num_ctas > 1 unsupported before Hopper.")
 
   signature = dict(enumerate(arg_dtypes))
   # TODO(sharadmv,zhangqiaorjc): handle differently aligned pointers
@@ -269,15 +268,14 @@ def get_or_create_triton_kernel(
       num_warps,
       num_stages,
       num_ctas,
+      compute_capability,
       enable_fp_fusion,
   )
   kernel = _COMPILED_KERNEL_CACHE.get(cache_key)
 
   if kernel is None:
-
-    target = ("cuda", arch)
+    target = ("cuda", compute_capability)
     cuda_backend = cb.CUDABackend(target)
-
     cuda_options = cuda_backend.parse_options(
         dict(
             num_warps=num_warps,
@@ -306,13 +304,12 @@ def get_or_create_triton_kernel(
     )
 
     ttir = str(module)  # `module`` is compiled in-place, so copy TTIR here.
-    ptx, kernel_name, shared_mem_bytes, compute_capability, cluster_dims = (
+    ptx, kernel_name, shared_mem_bytes, cluster_dims = (
         compile_ttir_to_ptx_inplace(
             module,
             cuda_backend,
             cuda_options,
-            device=device,
-            device_type=device_type,
+            compute_capability,
         )
     )
 
@@ -343,6 +340,7 @@ def triton_kernel_call_lowering(
     num_warps,
     num_stages,
     num_ctas,
+    compute_capability,
     enable_fp_fusion,
     input_output_aliases,
     zeroed_outputs,
@@ -455,6 +453,7 @@ def prune_configs(configs, named_args):
         num_warps=params["num_warps"],
         num_stages=params["num_stages"],
         num_ctas=params["num_ctas"],
+        compute_capability=compute_capability,
         enable_fp_fusion=enable_fp_fusion,
         metaparams=dict(params["metaparams"]),
         dump=debug,
@@ -542,6 +541,7 @@ def triton_call(
     num_warps: Optional[int] = None,
     num_stages: Optional[int] = None,
     num_ctas: int = 1,  # TODO(giorgioa): Add support for dimensions tuple.
+    compute_capability: Optional[int] = None,
     enable_fp_fusion: bool = True,
     input_output_aliases: Optional[Dict[int, int]] = None,
     zeroed_outputs: Union[
@@ -668,6 +668,7 @@ def add(x: jnp.ndarray, y: jnp.ndarray) -> jnp.ndarray:
       num_warps=num_warps,
       num_stages=num_stages,
       num_ctas=num_ctas,
+      compute_capability=compute_capability,
       enable_fp_fusion=enable_fp_fusion,
       input_output_aliases=tuple(input_output_aliases.items()),
       zeroed_outputs=zeroed_outputs,