Closed zaferarican closed 3 months ago
Never mind. I figured out that it was due to the kernel 6.1.43. It worked on 5.10.
what version of rknntoolkit2 do you use? 2.0.0?
It works both on 1.6.0 and 2.0.0. For me the key point was using the 5.10 kernel for Ubuntu, not the 6.1.43 version.
core_mask is used to select which core to use. RK3588 has 3 NPU cores so you can run 3 models at the same time.
thanks for your reply.I also have a issue.When i run my code with kernel 5.10,it reminds me that my Rknn driver version is too low(0.8.2),does it matter?
thanks for your reply.I also have a issue.When i run my code with kernel 5.10,it reminds me that my Rknn driver version is too low(0.8.2),does it matter?
It doesnt matter at all.
Mine is also 0.8.2. It seems ok. I did not have a problem. I did not perform comprehensive tests though.
Is it possible to run multiple models on RK3588 by distributing them to different cores by using for example rknn_set_core_mask?
When I try to test core_mask using the rank_benchmark example, I get
Is it even possible to use multiple cores in RK3588?
I am running on Orange Pi 5 Plus with 16 GB Ram.