-
I am running SpeedPPI on GPU nodes. But some of the jobs would run out of memory even with 250GB memory. The error says `RESOURCE_EXHAUSTED: Out of memory while trying to allocate 16508718128 bytes.` …
-
```
To use the gperftools heap-profiler, we have to use the tcmalloc memory
allocator with it. But in essence, they're independent.
I'm now trying to make the gperftools heap-profiler available with…
-
**Describe the bug**
Calling a UDF inside a CHECK constraint is causing issues.
**How to reproduce**
Run Fiddle:
https://fiddle.clickhouse.com/5da6f97a-59c9-4441-9ec0-d7d6aff8a0d0
TSAN report…
-
**Describe the bug**
On unit testing the conv ops of model_k, for the input resolution 256x256:
- Run the following command to reproduce the issue: `pytest tests/ttnn/unit_tests/operations/test_new_…
-
Having too many tablets causes all sorts of issues:
- Excessive metadata footprint / OOM
- due to `tablet_map` size in memory
- compaction group overhead, sstable metadata overhead (more relevant…
-
I have previously successfully used skia-safe in conjunction with ash and vk-mem-rs to build a Vulkan-based 3D application with an overlay rendered with Skia. This was with version 0.47.0 of skia-safe…
-
The bug in https://github.com/rust-lang/hashbrown/issues/438 has been fixed, and we need to upgrade to a new version of `hashbrown` to fix it. After upgrading, we can remove the workaround in this pr:…
-
Just as we have [`using_allocator`]( https://docs.cupy.dev/en/stable/reference/generated/cupy.cuda.using_allocator.html ) for device memory, it would be helpful to have a similar function for pinned m…
-
For some reasons, if the memory is not enough for `DoAllocMembers(count, allocator)`, it will got `nullptr` as return value. However, there seems to be no checks or assert for this sititutation. It to…
-
_Перенос предложения:_ голоса +0, -3
_Автор идеи:_ **Victor Gubin**
Мотивация. STL контейнеры позволяют определить собственный распределитель памяти, и принимают std::allocator в качестве параметр…