Rectify Vulkan Flags and Command to obtain Vulkan Target Triple (#77)

-Add correct vulkan flag to run minilm_jit
-Add automatic specification of target triple for better speed.
-Will work when https://github.com/nod-ai/SHARK-Runtime/releases/tag/candidate-20220601.20 gets
released, for now we can replace the shark release link to this on setup_venv.sh to work.

shark/examples/shark_inference/minilm_jit.py

@@ -31,4 +31,5 @@ shark_module = SharkInference(MiniLMSequenceClassification(), (test_input,),
                               jit_trace=True)
 
 shark_module.compile()
-shark_module.forward((test_input,))
+result = shark_module.forward((test_input,))
+print("Obtained result", result)

shark/iree_utils.py

@@ -89,12 +89,23 @@ def get_iree_gpu_args():
     else:
         return ["--iree-hal-cuda-disable-loop-nounroll-wa"]
 
-def get_iree_vulkan_args():
-    return [
-        "--iree-flow-demote-i64-to-i32=false",
-        "--iree-flow-demote-f64-to-f32=true"
-    ]
+def get_vulkan_triple_flag():
+    vulkan_device_cmd = "vulkaninfo | grep deviceName | awk \'END{{print $3}}\'"
+    vulkan_device = run_cmd(vulkan_device_cmd).strip()
+    if vulkan_device == "apple":
+        return "-iree-vulkan-target-triple=m1-moltenvk-macos"
+    elif vulkan_device == "A100-SXM4-40GB":
+        return "-iree-vulkan-target-triple=ampere-rtx3080-linux"
+    else:
+        print("Optimized kernel for your target device is not added yet. Contact SHARK Admin on discord or pull up an issue.")
+        return None
 
+def get_iree_vulkan_args():
+    vulkan_flag = ["--iree-flow-demote-i64-to-i32"]
+    vulkan_triple_flag = get_vulkan_triple_flag()
+    if vulkan_triple_flag is not None:
+        vulkan_flag.append(vulkan_triple_flag)
+    return vulkan_flag
 
 def get_iree_device_args(device):
     if device == "cpu":