Simon Perkins
Simon Perkins
I noticed that the warp_id calculation within much of the code assumes a 1D thread block. Is this an assumed limitation when using trove? ```c++ int warp_id = threadIdx.x &...
Found a bug? Maybe our [Slack Community](https://slack.cloudposse.com) can help. [](https://slack.cloudposse.com) ## Describe the Bug In https://github.com/cloudposse/terraform-aws-security-group/issues/38, I was referencing the `security_group_id` output instead of the `eks_cluster_managed_security_group_id` output, mostly because...
In [depthtoflow_cuda.cu](https://github.com/lmb-freiburg/lmbspecialops/blob/d1598c8ea5595d0f8e0f920af409593872b42152/src/depthtoflow_cuda.cu#L333-L335) I see the following pattern has been used: ```cpp //device.memcpyDeviceToHost(in_cpu, in_gpu, sizeof(T)*w_size*step); // Is this async? cudaMemcpyAsync(in_cpu, in_gpu, sizeof(T)*w_size*step, cudaMemcpyDeviceToHost, device.stream() ); cudaStreamSynchronize(device.stream()); ... depthtoflow_gpu(device.stream(), ``` As I...
Update https://github.com/dask/distributed/pull/2180 to use the approach suggested by @madsbk's in https://github.com/dask/dask/pull/6059.
Based on this discussion: - https://github.com/google/tensorstore/issues/114#issuecomment-1936418508 TODO: - ~~[x] Integrate AWS CPP SDK #defines into a `config.h` file variant. https://github.com/google/tensorstore/pull/149#discussion_r1548284979~~
This issue tracks remaining issues from the TODO list here: https://github.com/google/tensorstore/pull/91#issue-1649705676 Related: - https://github.com/google/tensorstore/issues/46 - https://github.com/google/tensorstore/pull/91 # KVStore - [ ] Versioned Bucket Support. From a high-level perspective might involve...
In the averaging code, UVW coordinates are linearly interpolated which probably introduces slight error. At the moment this seems to be acceptable (https://github.com/ska-sa/xova/issues/8#issuecomment-578886660) but if it ever becomes a problem...
- [x] Tests added / passed ```bash $ py.test -v -s africanus ``` If the pep8 tests fail, the quickest way to correct this is to run `autopep8` and then...
Within the FUSED RIME, the beam is sampled twice per row, and there is a non-trivial degree of computation involved: https://github.com/ska-sa/codex-africanus/blob/174645637b39ea037a7b557e29cdebd55ba0146c/africanus/experimental/rime/fused/terms/cube_dde.py#L149-L295 The performant solution would to evaluate antenna terms per...