-
- [ ] I have searched the [issues](https://github.com/seata/seata/issues) of this repository and believe that this is not a duplicate.
### Ⅰ. Issue Description
global_table 2分10s延迟删除太慢了,在高并发下,…
-
-
[ncpol2sdpa](https://github.com/peterwittek/ncpol2sdpa/blob/bce75d524d0b9d0093f32e3a0a5611f8589351a7/ncpol2sdpa/sdp_relaxation.py#L1387) has the ability of adding arbitrary (linear) constraints betwee…
-
Hi:
As far as I know, there are two ways to allocate resource:
1. Coarse granularity: Partition machine into fixed-size slots, and every slot can run one task, such as Hadoop.
2. Fine-grained resour…
cxxly updated
7 years ago
-
Once we've merged https://github.com/HaxeFoundation/haxe/pull/8730 we'll be able to poll the connected socket to see if there's a pending request. This should allow us to implement a pattern like so:
…
Simn updated
7 months ago
-
In the parallelization of multiple tasks, the memory take of the spark driver is increased by the memory requirement of each task. I think the problem is in `apply_async` function which keeps the ref…
gench updated
3 years ago
-
For example, if the looping part of a recurring set of tasks dies, the whole recurrence will die. We should develop a system to allow analysts to state that, for example, a given task should be restar…
-
if there are many connections to the port, the request is processed one by one, or there is a thread pool to process it?
-
I'm trying to specify resources for builtin dask functions such a `dd.read_csv`, with an end goal of running certain functions on "CPU workers" and other functions on "GPU workers". Here's a minimal e…
-
hi, I managed to get this to run on rhel 7.5 but the data going in just doesn't show up, I see 100's pending tasks stacking up, so I know the data is definitely getting there. Is there something else…