-
Krylov exponential scheme to solve the linear homogeneous ODE
$$
\dot{\mathbf{x}} = \mathbf{Ax}.
$$
Here are some useful references:
- [1] Niesen & Wright. A Krylov subspace algorithm for e…
-
# Sighting report
The default qos->thread of vhost-blk is app_thread. So when i deploy vhost-blk instances over all the NUMA nodes. All of the qos caculating will be sent to the app_thread. The …
-
Hey, @ankane all the vector database shilling on HN is pissing me off. Would you be open to receiving sponsorships for this project? If not, I think we may get something started for people like @jeadi…
-
I am facing this error in the scenario where I tried to run only a single phase of the test. **For example, when I tried to run only ior-easy phase(both read and write) what I did is that for other ph…
-
This has come up before, e.g., #195, but it seemed difficult to reproduce. Not sure if it got tracked down, if this is something else, or if it shouldn't be expected to work.
On my machine I can re…
olof3 updated
3 years ago
-
## Current Behavior
NVMe-oF latency increased due to recent merges on master branch.
Between de8f3a50ee and a620cd198f commits average TCP latency using SPDK Target and SPDK Initiator increased fr…
-
4K 100% read miss IOPS (with cache occupancy close to 0%) is order of magnitude lower on kernel 4.13 vs 5.3 :
CAS 19.6 + kernel 4.19 = ~50 KIOPS [expected]
CAS 19.9 + kernel 4.19 = ~27 KIOPS …
arutk updated
4 years ago
-
Hi!
I'm facing similar issue as with io500-dev repo (https://github.com/VI4IO/io-500-dev/issues/62)
I'm starting ior_hard only part and after 300s it doesn't stop.
Here is slurm out file
```
I…
-
When I set the queue depth to something above 1M, the system returns immediately finding 0 files:
[carlilek@h10u01 io-500-script]$ mpirun -np 32 -ppn 32 -f hostfile-10 -bootstrap ssh ./bin//pfind ./…
-
Hello TUNA, I'm minhoryang from South Korea.
I want to serve a mirror in my university, so I have a question for size of total mirror data.
- Total Mirror Size(镜像大小)?
- Average Storage IOPS?
Th…