-
It takes a lot of time to run it now. Thanks!
-
# Motivation
On Hopper, [efficient gemm requires warp-specialization](https://github.com/NVIDIA/cutlass/blob/main/media/docs/efficient_gemm.md#warp-specialization), which is not currently supported…
-
A few related comments:
- the `read()` algorithm verifies permissions "in parallel", which is, AFAIU, a deprecated way to describe parallelism. Presumably this should be done on the permissions tas…
-
PR https://github.com/scylladb/scylla-manager/pull/3496 brought new option for repair task to tackle rare problem with repairs getting stuck due to having too many repair jobs running on the same host…
-
To be updated.
-
We need a way to run things asynchronously within an address space.
There are lots of use cases for this, one of which is running
`DO CONCURRENT` regions on GPUs.
## Summary
Starting in Fortra…
-
**Enhancement description:**
As a user, I want to launch the same task, running n times of instances/pods.
All of them run independently.
I think that it is the k8s model ``parallel Jobs with…
-
### Description
map tasks generally cannot exceed the 5k limit. more info present in the [slack thread](https://discuss.flyte.org/t/16094488/lets-say-you-have-a-task-that-you-need-to-run-100-000-time…
-
### Feature Request / Improvement
The flink task written to iceberg can set write-parallelism, which allows it to have different parallelism degrees from the previous operator. However, when it modif…
-
### Description
Currently, the total amount of resources for the Flink Sort Job comes from the configuration file `flink-sort-plugin.properties`, so all submitted sort jobs will use the same amount…