MicroZed Performance question

Hi,
I wanted to complete some rudimentary performance tests on the MicroZed board and I am getting some disappointing results. I followed the tutorials and created an SDK. Then I ran a simple test reading from an array in RAM.

for(i=0;i<1000000;i++)
tResult = ulBit[0];

It takes 6mS to complete. This is on a single core with both I-cache and D-cache enabled. I wondered if this is the expected performance or does this indicate a implementation issue?
TIA.

0 Former Member over 11 years ago

The default compiler settings are set for debug with no optimization enabled. You might try changing the build configuration from 'debug' to 'release' or even set the compiler optimization settings directly to see if that improves your results.

-Gary
Cancel
Vote Up 0 Vote Down

Sign in to reply

Verify Answer

Cancel
0 Former Member over 11 years ago

Hi,
Optimization was turned on. Given that RAM transfer rates of 1066MT/s are supported this theoretically is 32GB/s the performance seem to be far less that I would have expected. Can anyone confirm that the results are normal/abnormal.
Thanks.
Cancel
Vote Up 0 Vote Down

Sign in to reply

Verify Answer

Cancel