Ops.ALU is no more, the arg is just an op (#7525)

* op arg alu [pr]

* more

* more passing

* fix more tests

* more tests passing

* fix single failing test

* so much cleaner

* noop to not have process replay trigger

* fix ptx

tinygrad/runtime/ops_python.py

@@ -7,7 +7,7 @@ import pickle, base64, itertools, time, struct
 from tinygrad.dtype import DType, dtypes, ImageDType, PtrDType, truncate
 from tinygrad.helpers import all_same, getenv, flatten
 from tinygrad.device import Compiled, Compiler, Allocator
-from tinygrad.ops import BinaryOps, TernaryOps, exec_alu, Ops, UOp
+from tinygrad.ops import BinaryOps, TernaryOps, exec_alu, Ops, UOp, GroupOp
 from tinygrad.renderer import Renderer
 from tinygrad.renderer.cstyle import CUDARenderer, MetalRenderer, AMDRenderer, IntelRenderer, ClangRenderer
 
@@ -173,10 +173,10 @@ class PythonProgram:
             def c_map(_, elem): return (elem%16, elem//16)
             ul[i] = wmma_helper(1, 1, 16, 16, 256, elem, elem, c_map)
           else: raise NotImplementedError(f"unimplemented tensor core {arg}")
-        elif uop is Ops.ALU:
-          assert all_same([len(x) for x in inp]), f"{[len(x) for x in inp]} doesn't match on {arg}"
-          assert all_same([dtype] + dtp) or arg in {BinaryOps.CMPNE, BinaryOps.CMPLT, TernaryOps.WHERE}, f"dtype mismatch on {arg}"
-          ul[i] = [exec_alu(arg, dtype, p) for p in zip(*inp)]
+        elif uop in GroupOp.ALU:
+          assert all_same([len(x) for x in inp]), f"{[len(x) for x in inp]} doesn't match on {uop}"
+          assert all_same([dtype] + dtp) or uop in {BinaryOps.CMPNE, BinaryOps.CMPLT, TernaryOps.WHERE}, f"dtype mismatch on {uop}"
+          ul[i] = [exec_alu(uop, dtype, p) for p in zip(*inp)]
         assert i in ul, (uop, dtype, idp, arg)
         i += 1
     return time.perf_counter() - st