[RUNTIME] Added auto-alignment mechanism (#71)

This PR adds an automatic memory alignment mechanism in the Triton runtime. Specifically, the JIT compiler detects the alignment (in bytes) of each pointer argument as well as the largest power of two divisor (between 1 and 16) of each integer argument. Proper .aligned and .multipleof attributes are then added to the Triton-IR on-the-fly for all auto-tunable kernels. There is a cache that remembers all the kernels compiled for each possible configuration.

This PR also includes substantial cleaning of the Python API. This adds 2-3us overhead, mostly due to accessing integer #defines from the auto-tuned compilation options. The previous solution was slightly faster but hacky and potentially unsafe, so this is preferred for now.

python/triton/ops/matmul.c

@@ -1,16 +1,12 @@
 #define STM 8
 #define STN 8
 
-__global__ void matmul(TYPE *A __noalias __readonly __aligned(16),
-                       TYPE *B __noalias __readonly __aligned(16),
-                       TYPE *C __noalias __aligned(16),
+__global__ void matmul(TYPE *A __noalias __readonly,
+                       TYPE *B __noalias __readonly,
+                       TYPE *C __noalias,
                        float alpha,
-                       int M,
-                       int N,
-                       int K __multipleof(16),
-                       int lda __multipleof(LDA_POW2_DIV),
-                       int ldb __multipleof(LDB_POW2_DIV),
-                       int ldc __multipleof(LDC_POW2_DIV),
+                       int M, int N, int K,
+                       int lda, int ldb, int ldc,
                        int *locks) {
   // prologue
   int pid = get_program_id(0);