Jetson TX2 NX and Xavier NX accelerated performance compare

TX2 NX max accelerated 1.33 tflops,Xavier NX max accelerated 21 tops
how to compare their float accelerated performance?
our program run 50ms and 200ms on these two hardware platforms,Is it consistent with these hardware performance?

Not an answer, but it is useful to know that NX is just a form factor…the module-on-a-DIMM-connector form factor. The TX2 is quite old in terms of the actual SoC chip, the Xavier is an entirely newer generation in just about every way. This is to be expected that the Xavier is faster than the TX2. It isn’t entirely clear about the “50ms and 200ms” for which is which, and whether “faster is better” or “longer is better” or which of the TX2 or Xavier is 50ms versus 200ms.

sorry,I didn’t make it clear
the same float calculation program ,TX2 is 200ms, Xavier is 50ms
Does it match the hardware parameters?

200ms and 50ms for what? Total time to run the same operation? If so, then Xavier beats TX2, and this is how it should be. Xavier is newer/faster than TX2 by quite a bit. Xavier hardware is “top of the line” so far as Jetson model speeds are concerned, and TX2 is just “medium” speed.

Note: Ignore any “NX” in naming, this is a form factor, not a specification for the CPU/GPU.

1.33 tflops->21 tops,
How to compare the performance of two hardware platforms?
Does it mean 4 times faster execution for float operation?

Hi,

You can compare the performance in the below page:

Thanks.

1 Like