Re: Fio on Block size 128M

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On 7/4/23 11:25?PM, Nick dan wrote:
> Hi
> 
> I have a NVMe disk, SATA SSD and SATA HDD.
> I have mkfs.ext4 on all the disks. And mounted the file system on /mnt.
> I want the block size to be set as 128MB and file size 10GB
> 
> What is the correct way to benchmark ? As it is taking a lot of time
> and the values of Throughput are not correct. What should be taken
> into consideration? Can you send the fio  command?

What are you trying to do, specifically? A block size of 128MB should of
course work, but it's somewhat nonsensical as no applications would do
that, and on the device side, it'd be chopped into much smaller pieces
anyway. That means one fio read or write of 128MB would turn into
potentially many hundreds of device requests.

For any fio question or bug report, please include what you are seeing
and what you would expect. Just saying "it's taking a lot of time and
throughput is not correct" is not super useful, be more detailed and
specific. And as also stated in the REPORTING-BUGS file, include what
version you are running, the job file being used, etc. That would make
for a complete bug report.

-- 
Jens Axboe




[Index of Archives]     [Linux Kernel]     [Linux SCSI]     [Linux IDE]     [Linux USB Devel]     [Video for Linux]     [Linux Audio Users]     [Yosemite News]     [Linux SCSI]

  Powered by Linux