Reordered backends in readme.md

[ROCm/rocshmem commit: cb0f6d5731]

projects/rocshmem/README.md

@@ -8,23 +8,23 @@ overlap than traditional host-driven networking.
 rocSHMEM uses a single symmetric heap (SHEAP) that is allocated on GPU memories.
 
 There are currently three backends for rocSHMEM;
-IPC, GPU-IB and Reverse Offload (RO).
+IPC, Reverse Offload (RO), and GPU-IB.
 The backends primarily differ in their implementations of intra-kernel networking.
-Currently only the IPC conduit is supported.
-The GPU-IB and RO backends are provided as-is with
+Currently, only the IPC backend is supported.
+The RO and GPU-IB backends are provided as-is with
 no guarantees of support from AMD or AMD Research.
 
 The IPC backend implements communication primitives using load/store operations issued from the GPU.
 
-The GPU InfiniBand (GPU-IB) backend implements a lightweight InfiniBand verbs interface
-on the GPU. The GPU itself is responsible for building commands and ringing
-the doorbell on the NIC to send network commands.
-
 The Reverse Offload (RO) backend has the GPU runtime forward rocSHMEM networking operations
 to the host-side runtime, which calls into a traditional MPI or OpenSHMEM
 implementation. This forwarding of requests is transparent to the
 programmer, who only sees the GPU-side interface.
 
+The GPU InfiniBand (GPU-IB) backend implements a lightweight InfiniBand verbs interface
+on the GPU. The GPU itself is responsible for building commands and ringing
+the doorbell on the NIC to send network commands.
+
 ## Requirements
 
 rocSHMEM base requirements: