[test] t3-operator x fix and output

[occa] operator & qfunction still push/pull

[occa] all but ex1

[occa] vector sanitized

Update nek example to work with the GPUs

[occa] restrict tries

[occa] t10-basis output

[occa] cleanup

[occa] h & d arrays for the vector

add some comments

simple MAGMA CUDA kernels generator

Fill up more for the magmabackend. make test pass for magma

Modify Makefile to add magma backend

[make] merge phonies

Merge branch 'master' into occa-dev

[asan] CR in ignore file

[test] fortran OKL links

[occa] cleanup

[test] All 315 tests successful: C, Fortran @ CPU, GPU, OpenCL, OpenMP

[okl] pointer arith & cleanup

[okl] local pointers arith

[occa] t10 output & OpenCL backend

[occa] WIP qfunctions t20, t30 & ex1-volume-annulus examples

[occa] wip qfunction t30, ex1-volume-annulus

[occa] qfunction t20 cpu + gpu