Is ${HOME}/NVIDIA-INSTALLER Not In L4T R19.3? Or Is It Deprecated?

Just did my flash of L4T R19.3 on Jetson. Did the apply_binaries.sh script prior to this. The README.txt still refers to ${HOME}/NVIDIA-INSTALLER:

http://developer.download.nvidia.com/embedded/jetson/TK1/README.txt

Need to know if this is an error that this one-time installer is not present, or if this is correct. I had expected it to be in the L4T so I did not back this up prior to flashing. Do we need this?

I just (re)flashed L4T R19.3. It did not include the NVIDIA_INSTALLER afterwards.

However, I was able to get up and running by installing CUDA and building the samples as documented here:

http://elinux.org/Jetson/Installing_CUDA

Now, I can run the samples and they all work.

linuxdev, NVIDIA-INSTALLER is not needed anymore.
apply_binaries.sh is enough.
I think NVIDIA-INSTALLER was for the factory build that comes with a new kit.
I guess, when you use one of the releases, it is not required anymore.

OK, but why does it not detect the 192 cuda cores?

root@tegra-ubuntu:~/cuda_samples/0_Simple/cdpSimplePrint# nvidia-detector
none
none

root@tegra-ubuntu:~/cuda_samples/0_Simple/cdpSimplePrint# ~/cuda_samples/bin/armv7l/linux/release/gnueabihf/simpleMultiGPU
Starting simpleMultiGPU
CUDA-capable device count: 1
Generating input data…

Computing with 1 GPUs…
GPU Processing time: 90.413000 (ms)

root@tegra-ubuntu:~/# nvcc -V
nvcc: NVIDIA ® Cuda compiler driver
Copyright © 2005-2013 NVIDIA Corporation
Built on Sat_Mar_15_02:05:29_PDT_2014
Cuda compilation tools, release 6.0, V6.0.1