You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
$ ls -l test.sh
-rwxr-xr-x 1 user user 32 14. Feb 11:31 test.sh*
$ cp test.sh ~/.cache/vramfs
$ ls -l ~/.cache/vramfs/
-rw-r--r-- 1 user user 32 14. Feb 12:03 test.sh
$ mkdir ~/.cache/testdir
$ cp test.sh ~/.cache/testdir
-rwxr-xr-x 1 user user 32 14. Feb 12:05 test.sh*
cp -p preserves the permissions, but why do they get lost in the first place? I wanted to do some testing on how fast a vramfs would be as a build directory compared to an SSD and a tmpfs in real RAM.
I'm using an RX480 with 8 GiB GDDR5, on an up-to-date arch linux using the opencl-amdgpu-pro-orca opencl driver.
Unrelated testing results you might find interesting, and I don't know where else to put them:
Interestingly, I'm getting really good write speed and random write IOPS on the drive, but the read speed isn't great (around 1GiB/s using 128k blocks) and especially the random read IOPS is atrocious, worse than my 6 year old sata SSD.
The vramfs is mounted to ~/.cache/vramfs
cp -p preserves the permissions, but why do they get lost in the first place? I wanted to do some testing on how fast a vramfs would be as a build directory compared to an SSD and a tmpfs in real RAM.
I'm using an RX480 with 8 GiB GDDR5, on an up-to-date arch linux using the opencl-amdgpu-pro-orca opencl driver.
Unrelated testing results you might find interesting, and I don't know where else to put them:
Interestingly, I'm getting really good write speed and random write IOPS on the drive, but the read speed isn't great (around 1GiB/s using 128k blocks) and especially the random read IOPS is atrocious, worse than my 6 year old sata SSD.
command to test:
fio --randrepeat=1 --ioengine=libaio --direct=1 --gtod_reduce=1 --name=test --filename=test --bs=4k --iodepth=64 --size=4G --readwrite=randread
Result:
4k blocksize read: IOPS=19.2k, bw=74.9MiB/s
128k blocksize read: IOPS=4354, BW=544MiB/s
4k blocksize write: IOPS=73.5k, BW=287MiB/s
128k blocksize write: IOPS=12.2k, BW=1521MiB/s
and on my SSD:
4k blocksize read: IOPS=75.4k, BW=294MiB/s
128k blocksize read: IOPS=4145, BW=518MiB/s
4k blocksize write: IOPS=6184, BW=24.2MiB/s
128k blocksize write: IOPS=609, BW=76.2MiB/s
Using the basic dd test I also get worse read than write speed.
The text was updated successfully, but these errors were encountered: