deprecate MTIA_WORKLOADD from kineto (pytorch#1093)

Summary:

X-link: pytorch/pytorch#154609

Remove MTIA_WORKLOADD activity from PyTorch

Differential Revision: D75535500

Author: fenypatel99

libkineto/include/ActivityType.h

@@ -41,7 +41,6 @@ enum class ActivityType {
   HPU_OP, // HPU host side runtime event
   XPU_RUNTIME, // host side xpu runtime events
   COLLECTIVE_COMM, // collective communication
-  MTIA_WORKLOADD, // MTIA workloadd events
 
   // PRIVATEUSE1 Activity types are used for custom backends.
   // The corresponding device type is `DeviceType::PrivateUse1` in PyTorch.

libkineto/src/ActivityType.cpp

@@ -39,7 +39,6 @@ static constexpr std::array<ActivityTypeName, activityTypeCount + 1> map{
      {"hpu_op", ActivityType::HPU_OP},
      {"xpu_runtime", ActivityType::XPU_RUNTIME},
      {"collective_comm", ActivityType::COLLECTIVE_COMM},
-     {"mtia_workloadd", ActivityType::MTIA_WORKLOADD},
      {"privateuse1_runtime", ActivityType::PRIVATEUSE1_RUNTIME},
      {"privateuse1_driver", ActivityType::PRIVATEUSE1_DRIVER},
      {"ENUM_COUNT", ActivityType::ENUM_COUNT}}};