08-16-2019 02:02 PM
Hi Ben,
I disabled the folder directory and its still saving images very slow.
08-16-2019 02:25 PM
@JohnnyDoe771 wrote:
Hi Ben,
I disabled the folder directory and its still saving images very slow.
No, put a diagram disable around your code that interacts with the disk and that included saving etc. I realize that is not your final desire but it will tell you if the slow down is due to the disk actions or elsewhere.
If it runs as fast as the acquisition you will know the bottleneck is the disk and a SSD may be needed. If it not the disk, you can focus on other things.
"divide and conquer"
Ben
08-16-2019 04:15 PM
By now, you've probably made several changes to your code. These comments are based on the code attached with your original post.
Bob Schor
08-18-2019 03:38 AM
On current processors or unless you have a really devilish fast SSD with the bus connection that can sustain its speed, compression is almost always cheaper performance wise than the IO bottleneck. So compressing data before writing it to disk not only saves disk space but also speeds up the IO operation. LabVIEW compresses large parts of its internal VI parts mainly for this reason as zlib compressed streams on disk!
08-19-2019 03:44 AM
Minor point on "Producer-Consumer" and DAQmx. A DAQmx acquisition task is already a Producer, with LabVIEW the Consumer, with an image buffer in between, so coding another Producer loop gains one nothing and adds complexity and (minor) overhead.
08-21-2019 10:29 AM
This is an interesting thread. I'd like to know how the final VI looks.
08-21-2019 02:06 PM
Hey Eric,
The conversation is continuing at in this link: https://forums.ni.com/t5/LabVIEW/How-to-Optimize-Camera-Speed-fps/td-p/3958048/highlight/false/page/...
There's also my most recent code uploaded there 🙂