+ :i <^RIHt^RIHt^RIt^RIHtRRltR#))cuda)driverN) numpy_supportc Ta a \VR^4pV'gzVPwr4VPPV,VPP3p\P P PWC3VVPVR7p\P!VP4o \P!4Pp\\P!^\P !V^4^, 44p\Wg, 4pW^,3o \P"V V 3Rl4p \VP^,V, ^,4\VP^,V, ^,43p W3p WW3,!W4V#)aCompute the transpose of 'a' and store it into 'b', if given, and return it. If 'b' is not given, allocate a new array and return that. This implements the algorithm documented in http://devblogs.nvidia.com/parallelforall/efficient-matrix-transpose-cuda-cc/ :param a: an `np.ndarray` or a `DeviceNDArrayBase` subclass. If already on the device its stream will be used to perform the transpose (and to copy `b` to the device if necessary). stream)dtyperc<\PPS S R7p\PPp\PP p\P P\PP,p\P P \PP ,pWc,pWT,pWd,VP^,8d:WS,VP^,8dWV,WS,3,W$V3&\P!4WP^,8d*WqP^,8dW#V3,WV3&R#R#R#))shaperN) rsharedarray threadIdxxyblockIdxblockDimr syncthreads) inputoutputtiletxtybxbyrrdt tile_shapes && Z/var/www/html/photoedit/myenv/lib/python3.14/site-packages/numba/cuda/kernels/transpose.pykerneltranspose..kernel)s{{  z < ^^   ^^   ]]__t}} . ]]__t}} . G G 7U[[^ #%++a.(@ b"'!12DRL  ||A 1||A#6Br=s, ): r