-
Once the last cache has been implemented we will want to run a series of load generator tests to see how it performs compared to SQL queries that would be used in its absence.
The only setup requir…
-
### 🐛 Describe the bug
When I use a tensor with complex128 type as input for torch.acos, the outputs from cpu and gpu mismatch (with the output from gpu matching the numpy implementation).
Code: […
-
## 关于CPU cache
### 简要
一般来说x86 结构cpu cache line 的大小是64 byte, arm cacheline 是32 byte。
CPU 的 Cache 从上至下可以分为 L1(L1d 数据缓存, L1i指令缓存), L2, L3(多个 core 共享一个) 三层cache. 每层的访问速度由上至下依次递减, 可以使用`lscpu`命令查看.
```
…
-
This happened to me twice in the last 2 weeks and we've received a report from someone else too that it happened to them. I have absolutely no idea how to reproduce it.
Very rarely our process gets…
-
Good morning .
Can you create an ipk for a router with mipsel_24kc , please. ?
`root@Huastlink:/tmp# cat /proc/cpuinfo
system type : MediaTek MT7621 ver:1 eco:3
machine …
-
### Is your feature request related to a problem? Please describe
_No response_
### Describe the solution you'd like
If the number of rows submitted in a single batch is small, a large number…
whh32 updated
4 months ago
-
### 🐛 Describe the bug
When importing all metrics, the import will fail if `skimage` is not installed. This also happens when importing an unrelated metric such as `F1Score`.
Reproducer:
```
imp…
-
I build odyssey with [C Prometheus client library](https://github.com/digitalocean/prometheus-client-c)
Enable metrics in config with parameters:
```
promhttp_server_port 3422
log_general_stats_pr…
-
### Describe the bug
The pixel pipeline processing with **opencl disabled** takes considerably longer with kernel 6.10.6 compared to 6.6.47.
With kernel 6.6.47:
```
4,2765 [dev_process_e…
-
CPUs typically have cache lines. ~~For performance reasons, it is normally a good idea to not flush these cache lines too often. Consequently, it would be a good idea to be able to "pin" a process (an…