Tegrastats on TX2

Robotics & Edge Computing Jetson & Embedded Systems Jetson TX2
1

Hi
This is my output from TX2 Tegrastats:
RAM 1893/7854MB (lfb 12x4MB) CPU [0%@1113,0%@1346,0%@2036,0%@2061,0%@2035,0%@2036] EMC_FREQ 12%@1866 GR3D_FREQ 0%@1300 APE 150 MTS
RAM 1893/7854MB (lfb 12x4MB) CPU [0%@1113,0%@1346,0%@2036,0%@2061,0%@2035,0%@2036] EMC_FREQ 23%@1866 GR3D_FREQ 0%@1300 APE 150 MTS
RAM 1893/7854MB (lfb 12x4MB) CPU [0%@1113,0%@1346,0%@2036,0%@2061,0%@2035,0%@2036] EMC_FREQ 25%@1866 GR3D_FREQ 0%@1300 APE 150 MTS

I want to calculate memory bandwidth during run time with this info. Do you think using this formula is right?
(emc_freq%1866) MHz * (128/8) bytes bus width * (2) DDR4 multiplier

For example:
the first line output will be: (0.12*1866)MHz * (128/8)bytes * 2 = 716544 MB/s

If not, can you tell me where I am going wrong?

2

Cookie007,

The maximum bw for our DDR4 is 59.7 GB/s.

I’ll try to find out some tools for you.

3

Cookie007,

Sorry that I just noticed that those tools are for internal use only. I cannot release it to you.

Any special usecase that needs to check the bandwidth?

4

I want to stream 5 cameras on TX2 and save it to a file; but it seems to keep crashing. I wanted to check how much memory bandwidth my code utilizes for saving the file from stream and see if there is anything to optimize it.

5

Hi Cookie007,
Do you use USB cameras? How do you connect your cameras to TX2?

6

I use the Leopard Imaging MIPI cameras for TX2 :https://leopardimaging.com/product-category/nvidia-jetson-camera/nvidia_tx1tx2_camera_kit/

7

Hi Cookie007,
Your issue is not same as this thread. Please start a new thread.