From aa39573ee152de0e8182719042f14c41f07d4557 Mon Sep 17 00:00:00 2001
From: Deven Desai <deven.desai.amd@gmail.com>
Date: Wed, 8 Apr 2020 22:44:41 +0000
Subject: [PATCH] Populating memory bandwidth information in the
 DeviceDescription, using information from the call to hipGetDeviceProperties

---
 tensorflow/stream_executor/rocm/rocm_gpu_executor.cc | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/tensorflow/stream_executor/rocm/rocm_gpu_executor.cc b/tensorflow/stream_executor/rocm/rocm_gpu_executor.cc
index 2a85cb820ed..dbab0304d82 100644
--- a/tensorflow/stream_executor/rocm/rocm_gpu_executor.cc
+++ b/tensorflow/stream_executor/rocm/rocm_gpu_executor.cc
@@ -856,6 +856,11 @@ GpuExecutor::CreateDeviceDescription(int device_ordinal) {
 
     float clock_rate_ghz = static_cast<float>(prop.clockRate) / 1e6;
     builder.set_clock_rate_ghz(clock_rate_ghz);
+
+    // mem_bandwidth = 2 * mem_bus_width_in_bytes * mem_clock_rate_in_hz
+    int64 memory_bandwidth = 2 * (int64(prop.memoryBusWidth) / 8) *
+                             (int64(prop.memoryClockRate) * 1000);
+    builder.set_memory_bandwidth(memory_bandwidth);
   }
 
   {