[Triton] Support math and libdevice ops (#91)

This PR adds basic math ops by using `MathDialect` and `libdevice` ops by using `extern_elementwise`. This is needed to compile some tutorial code (e.g., `softmax`). This PR implements only interface till PTX (so from frontend to TritonGPU-MLIR) 
- Currently till TritonGPU. It cannot be lowered to PTX now.
- No special optimizations (e.g., constant folding etc) are applied.
  - 14.x does not define folders for many operators for math ops, but 15.x seems to increase its coverage: https://github.com/llvm/llvm-project/blob/llvmorg-15.0.0-rc3/mlir/include/mlir/Dialect/Math/IR/MathOps.td
  - No constant folding etc for `libdevice` ops.

```py
import triton
import triton.language as tl
import sys

@triton.jit
def add_kernel(
    x_ptr,
    y_ptr,
    BLOCK_SIZE: tl.constexpr,
):
    offsets = tl.arange(0, BLOCK_SIZE)
    x = tl.load(x_ptr + offsets)
    x = tl.sin(x)
    output = tl.libdevice.sin(x)
    output = tl.libdevice.fdiv_rn(output, output)
    output = tl.libdevice.fmaf_rd(output, output, output)
    tl.store(y_ptr + offsets, output)


if __name__ == "__main__" and len(sys.argv) >= 2:
    signature = "*fp32,*fp32"
    constants = {'BLOCK_SIZE': 1024}
    output = triton.compile(add_kernel, signature, device=0, constants=constants, output="ttgir")
    print(output)
```
->
```llvm
#blocked = #triton_gpu.blocked<{sizePerThread = [1], threadsPerWarp = [32], warpsPerCTA = [4], order = [0]}>
module attributes {"triton_gpu.num-warps" = 4 : i32} {
  func @add_kernel__Pfp32_Pfp32__2c1024(%arg0: !tt.ptr<f32>, %arg1: !tt.ptr<f32>) {
    %0 = tt.make_range {end = 1024 : i32, start = 0 : i32} : tensor<1024xi32, #blocked>
    %1 = tt.splat %arg0 : (!tt.ptr<f32>) -> tensor<1024x!tt.ptr<f32>, #blocked>
    %2 = tt.getelementptr %1, %0 : tensor<1024x!tt.ptr<f32>, #blocked>
    %3 = tt.load %2 {cache = 1 : i32, evict = 1 : i32, isVolatile = false} : tensor<1024xf32, #blocked>
    %4 = math.sin %3 : tensor<1024xf32, #blocked>
    %5 = tt.ext_elemwise %4 {libname = "libdevice", libpath = "/home/siwasaki/triton/python/triton/language/libdevice.10.bc", symbol = "__nv_sinf"} : tensor<1024xf32, #blocked> -> tensor<1024xf32, #blocked>
    %6 = tt.ext_elemwise %5, %5 {libname = "libdevice", libpath = "/home/siwasaki/triton/python/triton/language/libdevice.10.bc", symbol = "__nv_fdiv_rn"} : tensor<1024xf32, #blocked>, tensor<1024xf32, #blocked> -> tensor<1024xf32, #blocked>
    %7 = tt.ext_elemwise %6, %6, %6 {libname = "libdevice", libpath = "/home/siwasaki/triton/python/triton/language/libdevice.10.bc", symbol = "__nv_fmaf_rd"} : tensor<1024xf32, #blocked>, tensor<1024xf32, #blocked>, tensor<1024xf32, #blocked> -> tensor<1024xf32, #blocked>
    %8 = tt.splat %arg1 : (!tt.ptr<f32>) -> tensor<1024x!tt.ptr<f32>, #blocked>
    %9 = tt.getelementptr %8, %0 : tensor<1024x!tt.ptr<f32>, #blocked>
    tt.store %9, %7 : tensor<1024xf32, #blocked>
    return
  }
}
```

python/tests/test_math_ops.py

@@ -0,0 +1,33 @@
+
+import triton
+import triton.language as tl
+
+
+@triton.jit
+def math_kernel(x1_ptr, x2_ptr, x3_ptr, x4_ptr, n, BLOCK_SIZE: tl.constexpr):
+    offsets = tl.arange(0, BLOCK_SIZE)
+    x1 = tl.load(x1_ptr + offsets, mask=offsets < n)
+    x2 = tl.load(x2_ptr + offsets, mask=offsets < n)
+    x3 = tl.load(x3_ptr + offsets, mask=offsets < n)
+    x4 = tl.load(x4_ptr + offsets, mask=offsets < n)
+
+    y1 = tl.sin(x1)
+    y2 = tl.libdevice.sin(x2)
+    y3 = tl.libdevice.fdiv_rn(x3, x3)
+    y4 = tl.libdevice.fmaf_rd(x4, x4, x4)
+
+    tl.store(x1_ptr + offsets, y1, mask=offsets < n)
+    tl.store(x2_ptr + offsets, y2, mask=offsets < n)
+    tl.store(x3_ptr + offsets, y3, mask=offsets < n)
+    tl.store(x4_ptr + offsets, y4, mask=offsets < n)
+
+
+def test_empty_kernel_cubin_compile():
+    kernel = triton.compile(math_kernel,
+                            "*fp32,*fp32,*fp32,*fp32,i32",
+                            device=0,
+                            constants={"BLOCK_SIZE": 256},
+                            output="ttgir")  # "cubin"
+    assert kernel
+    # TODO: Check if the values are correct.
+    # TODO: Cover all the math operators