-
### What happened?
When dealing with hugepage resources, the unit of the resource is an entire page (typically either 2Mi or 1Gi sizes). The unit of the resource is not the amount of memory in bytes.…
-
I found that the latest kernel (or OVMF?) has some performance issues compared to the 6.6 kernel. For example, on a SNP guest, (1) [Intel MLC](https://www.intel.com/content/www/us/en/download/736633/i…
-
- [ICDE Workshop 21] [An Investigation of Atomic Synchronization for Sort-Based Group-By Aggregation on GPUs](https://ieeexplore.ieee.org/document/9438804)
- Experiments in II.B & Figure 3 are good…
-
JVM during its runtime, doesn't recognize the new CPUs that are added dynamically when `NUMA` is ON. New CPUs are added dynamically via CPU Hotplug or LPAR(in Linux PPC). Because of this, the new thre…
-
This is a tracking issue for handling locality aware scheduling of on-node resources (currently only GPUs I think) in Fluxion.
This is a requirement for our Production Ready System Instance milestone…
-
## What would you like to be added?
Flag to not advertise NUMA information
## What is the use case for this feature / enhancement?
Logic to generate placement hints in topology manager is …
-
I prevent core 3 (I have 4 cores numbered: 0..3) from being used by anything else on the system by adding isolcpus=3 to my kernel command line parameters.
On reboot this is OK. 'top' shows all 4 cor…
ajvok updated
3 years ago
-
BOINC client needs to be NUMA-aware to support systems with multiple CPU Groups (NUMA nodes).
- Manage pool of all client processes and explicitly assign Group Affinity via SetThreadGroupAffinity.
Se…
-
### DO NOT REMOVE OR SKIP THE ISSUE TEMPLATE
- [X] I understand that I will be **blocked** if I remove or skip any mandatory\* field
### Checklist
- [X] I'm reporting a bug unrelated to a spe…
-
Below, I have listed the five different approaches/solutions, which will improve OpenJ9's locking performance. Each approach listed below will incrementally improve OpenJ9's locking performance.
#…