[BACKEND] Compiler improvements (#557)

This PR adds several optimization capabilities in the compiler backend:
- Now using inline PTX for `tl.store`, making it possible to use things like evict_last
- For A100, mma layout can be directly converted to shared memory
- For A100, an additional "transpose" argument in `dot` allows tensors to be loaded once and used both row- and col- major.
- Fixed liveness analysis; this was broken.
- Now can load/store directly mma layout without converting. Useful for when tl.dot accumulator is initialized with DRAM data inside of an inner loop.
- `tl.dot` can now take LHS inputs in registers when it comes from a previous `tl.dot` instruction. Useful for e.g. fused attention.

lib/codegen/analysis/allocation.cc

@@ -92,8 +92,10 @@ void allocation::run(ir::module &mod) {
   }
   // Save maximum size of induced memory space
   allocated_size_ = 0;
-  for(shared_layout* x: V)
+  for(shared_layout* x: V){
     allocated_size_ = std::max<size_t>(allocated_size_, starts[x] + x->get_size());
+    // std::cout << "start: " << starts[x] << " | end: " << starts[x] + x->get_size() << std::endl;
+  }
 }
 
 }