Fix lack of uniqueId on AMD GPU OpenCL without AMD extensions #92
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
AMD GPU systems need not have a functional clGetDeviceInfo(..., CL_DEVICE_TOPOLOGY_AMD, ...). E.g., Fedora OpenCL with AMD GPU drivers, it seems. As the code stands it doesn't create a uniqueId, and leaves it null. If there are multiple GPUs, the device info is stored to null uniqueId and overwritten - and only one of these GPUs will be enumerated and visible to the miner. Fix this by providing a fall-back, constructing the uniqueId from the platform and device indexes. As is already the case for Intel GPU and CPU devices. Additionally, cosmetically, add a clearer, more self-describing struct for the AMD topology data.
Hard coded Boost 1.66.0 no longer available, cmake fails. Use the built-in Hunter default.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
AMD GPU systems need not have a functional
clGetDeviceInfo(..., CL_DEVICE_TOPOLOGY_AMD, ...).
E.g., Fedora OpenCL with AMD GPU drivers, it seems. As the code stands it
doesn't create a uniqueId, and leaves it null. If there are multiple GPUs,
the device info is stored to null uniqueId and overwritten - and only one of
these GPUs will be enumerated and visible to the miner.
Fix this by providing a fall-back, constructing the uniqueId from the
platform and device indexes. As is already the case for Intel GPU and CPU
devices.
Additionally, cosmetically, add a clearer, more self-describing struct for
the AMD topology data.
To get Hunter to work I had to update the Boost version.
The previously specified version is no longer available for hunter to download.