cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 
nastasiyacsf
Visitor
Visitor
193 Views
Registered: ‎07-10-2018

Petalinux 2019.1 BogoMIPS

Jump to solution

Good day! I'm trying to optimize a driver for Zynq-7020. Tell me, does petalinux really use both kernels? Compare Buildroot and Petalinux

Buildroot writes on kernel boot:

 

Brought up 2 CPUs
SMP: Total of 2 processors activated (2664.32 BogoMIPS)

 

Petalinux writes:

 

Brought up 2 CPUs
SMP: Total of 2 processors activated (1332.16 BogoMIPS)

 

 Why write speed is two times less? And how to increase it?

Tags (1)
0 Kudos
1 Solution

Accepted Solutions
rfs613
Scholar
Scholar
168 Views
Registered: ‎05-28-2013

You are likely comparing two different kernel versions. The way BogoMIPS is calculated has changed over time. For a while the value was even removed, but this broke some userspace code, so it was added back (see "Bogomips" section of this LWN article).

Note that BogoMIPS does not measure write speed. Rather, it is the speed of "doing nothing", basically measuring how fast the CPU can run an empty loop. This is not, and has never been, a particularly good metric for performance, in particular on modern multicore systems.

For optimization, you should define your criteria: throughput, delay, CPU usage, memory usage, etc. These can be measured in various ways, but in the end, it is always a balancing act.

 

View solution in original post

0 Kudos
1 Reply
rfs613
Scholar
Scholar
169 Views
Registered: ‎05-28-2013

You are likely comparing two different kernel versions. The way BogoMIPS is calculated has changed over time. For a while the value was even removed, but this broke some userspace code, so it was added back (see "Bogomips" section of this LWN article).

Note that BogoMIPS does not measure write speed. Rather, it is the speed of "doing nothing", basically measuring how fast the CPU can run an empty loop. This is not, and has never been, a particularly good metric for performance, in particular on modern multicore systems.

For optimization, you should define your criteria: throughput, delay, CPU usage, memory usage, etc. These can be measured in various ways, but in the end, it is always a balancing act.

 

View solution in original post

0 Kudos