You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
[1] HIP/ROCm implements unified memory using slow device accessible host memory. This means that hipSYCL's call to hipMallocManaged cannot produce efficient shared allocations.
Is this still true? There is some change in version 4.5:
You are right, things are changing in ROCm. As hipSYCL just calls hipMallocManaged(), those changes directly translate to hipSYCL as well. However, from what I have seen and heard, performance of hipMallocManaged() even with page migration is still very, very bad.
Is this still true? There is some change in version 4.5:
https://github.com/RadeonOpenCompute/ROCm#unified-memory-support-in-rocm
The text was updated successfully, but these errors were encountered: