-
message rate performance needs some tuning for the atomics pathway: notably int add/increment.
refer to the following test from Sandia SHMEM-OFI for investigation purposes:
(will need to configure…
-
Currently Servo is very slow at these (about 20x slower than FF, after bincode landed). In particular, using drawImage with a canvas calls both getImageData and putImageData. I looked into this but …
-
The nonblocking RMA routines say:
> **`shmem_put_nbi`**
> The routines return after posting the operation. The operation is considered complete after a subsequent call to `shmem_quiet`. At the com…
-
I'm having trouble starting the server when the size of shared memory is beyond a certain size (not particularly large). If I take the [shmem example](https://github.com/pgcentralfoundation/pgrx/tree/…
-
The `/src/main/shmem/shmem_cleanup.c` module could use some refactoring so that it can be used to help cleanup temp files in addition to shared memory mappings.
Currently, this module is used to fi…
-
Hi namhyung & honggyu
I'm using for performance optimizing for some app
```
This app start from systemd as below
-/usr/bin/uftrace -d /tmp/uftrace.data --no-libcall record "my app path"
```
When…
-
-
Jack said he was in problems of letting `shmem_barrier_all()` work in his `global_ptr`. In fact, I also found problems when using it for C++. The problematic program is as followed:
-
After launching a pipeline, the worker pod's memory usage slowly increases
below is corresponding cpu usage
This pipeline uses kafka source and kafka sink. During some time intervals, the sour…
-
/kind bug
**What happened?**
Lately we discovered that some of our efs csi pods crash due to oom kills.
**What you expected to happen?**
na
**How to reproduce it (as minimally and precisely…