Julian Samaroo
Julian Samaroo
In #199 I've fixed a "bug" where non-thunk inputs are serialized during the thunk RPC call, which is silly for large inputs. That fix assumes that inputs are large enough...
If we send input data to a node (that we know is going to execute a thunk in the future) before its thunk is ready to execute, and cache it...
It can be annoying to have to pass around a `Context` to Dagger computations when you have non-default options set. We should provide a global `Context` which is used by...
And allow lazy schedulers to accept eager thunks? Without cache-sharing between schedulers, this will be rather inefficient, but could potentially be helpful for code which already has `Thunk`s describing a...
Should probably be a dagger with Julia's colors
Currently, operations on `DArray`s are staged into thunks before being passed into the scheduler, removing all information about operation structure. For some operations, such as a matmul, knowing the original...
Once #117 is merged, and #166 is realized, we'll want users to be able to use Dagger for traditional concurrent programming. In order to do so, we need to be...
The way processors are registered to Dagger is slightly ugly, and has the issue that we can't easily unregister processors later (because the callbacks in the list aren't uniquely identified)....
While round-robin work assigment is fine when first launching work without any prior knowledge, it is less efficient when individual work items are widely varying in duration, and when data...