trahman-quic commented on code in PR #12415: URL: https://github.com/apache/tvm/pull/12415#discussion_r944866584
########## python/tvm/topi/hexagon/slice_ops/qadd_qsub_qmul.py: ########## @@ -0,0 +1,253 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more contributor license agreements. See the NOTICE file +# distributed with this work for additional information +# regarding copyright ownership. The ASF licenses this file +# to you under the Apache License, Version 2.0 (the +# "License"); you may not use this file except in compliance +# with the License. You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, +# software distributed under the License is distributed on an +# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY +# KIND, either express or implied. See the License for the +# specific language governing permissions and limitations +# under the License. +# pylint: disable=invalid-name + +"""Compute and schedule for add, multiply, subtract slice op + +Please note the following assumptions made by the implementation: + +1) The inputs will be multiple of crouton layout except for the axis that needs broadcasting.""" + +import tvm +from tvm import te +from tvm import tir +from tvm import topi +from ..utils import get_layout_transform_fn +import struct + + +def broadcast_axis(tensor_A, tensor_B): + A_broadcast = [] + B_broadcast = [] + + for i in range(len(tensor_A.shape)): + if tensor_A.shape[i] == tensor_B.shape[i]: + A_broadcast.append(1) + B_broadcast.append(1) + elif tensor_A.shape[i] == 1: + A_broadcast.append(0) + B_broadcast.append(1) + elif tensor_B.shape[i] == 1: + A_broadcast.append(1) + B_broadcast.append(0) + return A_broadcast, B_broadcast + + +def flt_getexp(a): + a_f = struct.pack("f", a) + a_i = struct.unpack("I", a_f) + exp_value = ((a_i[0] >> 23) & 0xFF) - 127 + return exp_value + + +# this function will return x * 2 ^ y +def flt_ldexp(x, y): + a = ((y + 127) & 0xFF) << 23 + a_i = struct.pack("I", a) + a_f = struct.unpack("f", a_i) + return (x * a_f[0]).astype("float32") + + +def saturate_16(x): + return max(-32768, min(x, 32767)) + + +def saturate(x, dtype): + if dtype == "uint8": + return te.max(0, te.min(x, 255)) + elif dtype == "int8": + return te.max(-127, te.min(x, 128)) + return x + + +def get_int_scale(scale_A, scale_B, scale_M, zero_point_A, zero_point_B, zero_point_M, op): + C_recip = 1 // scale_M + + if op == "qmul": + scale = scale_A * scale_B * C_recip + exp = flt_getexp(scale) + rsh = 14 - exp + scale_int = int(saturate_16(round(flt_ldexp(scale, rsh)))) + corr = zero_point_M << rsh + + return scale_int, rsh, corr + else: + a_scale_f = scale_A * C_recip + b_scale_f = scale_B * C_recip + + exp_a = flt_getexp(a_scale_f) + exp_b = flt_getexp(b_scale_f) + + rsh = 14 - max(exp_a, exp_b) + + scale_a = int(saturate_16(round(flt_ldexp(a_scale_f, rsh)))) + scale_b = int(saturate_16(round(flt_ldexp(b_scale_f, rsh)))) + + if op == "qadd": + corr = (zero_point_M << rsh) - (zero_point_A * scale_a + zero_point_B * scale_b) + else: + corr = (zero_point_M << rsh) - (zero_point_A * scale_a - zero_point_B * scale_b) + + return scale_a, scale_b, rsh, corr + + +def qadd_broadcast_compute( + tensor_A, + tensor_B, + output_shape, + zero_point_A, + scale_A, + zero_point_B, + scale_B, + zero_point_M, + scale_M, +): + A_broadcast, B_broadcast = broadcast_axis(tensor_A, tensor_B) + n_a, h_a, w_a, c_a = A_broadcast + n_b, h_b, w_b, c_b = B_broadcast + + scale_a, scale_b, rsh, corr = get_int_scale( + scale_A, scale_B, scale_M, zero_point_A, zero_point_B, zero_point_M, "qadd" + ) + + return te.compute( + output_shape, + lambda n, h, w, c: saturate( + ( + ( + (tensor_A[n * n_a, h * h_a, w * w_a, c * c_a] * scale_a) + + (tensor_B[n * n_b, h * h_b, w * w_b, c * c_b] * scale_b) + + corr + ) + >> rsh + ), + "uint8", + ).astype("uint8"), + ) + + """ + + return te.compute( + output_shape, + lambda n, h, w, c: (((scale_A/scale_M) * (tensor_A[n * n_a, h * h_a, w * w_a, c * c_a] - zero_point_A)) + + ((scale_B/scale_M) * (tensor_B[n * n_b, h * h_b, w * w_b, c * c_b] - zero_point_B)) + + zero_point_M).astype('uint8'), + ) + """ + + +def qsubtract_broadcast_compute( + tensor_A, + tensor_B, + output_shape, + zero_point_A, + scale_A, + zero_point_B, + scale_B, + zero_point_M, + scale_M, +): + A_broadcast, B_broadcast = broadcast_axis(tensor_A, tensor_B) + n_a, h_a, w_a, c_a = A_broadcast + n_b, h_b, w_b, c_b = B_broadcast + + scale_a, scale_b, rsh, corr = get_int_scale( + scale_A, scale_B, scale_M, zero_point_A, zero_point_B, zero_point_M, "qsub" + ) + + return te.compute( + output_shape, + lambda n, h, w, c: saturate( + ( + ( + (tensor_A[n * n_a, h * h_a, w * w_a, c * c_a] * scale_a) + - (tensor_B[n * n_b, h * h_b, w * w_b, c * c_b] * scale_b) + + corr + ) + >> rsh + ), + "uint8", + ).astype("uint8"), + ) + + +def qmultiply_broadcast_compute( + tensor_A, + tensor_B, + output_shape, + zero_point_A, + scale_A, + zero_point_B, + scale_B, + zero_point_M, + scale_M, +): + A_broadcast, B_broadcast = broadcast_axis(tensor_A, tensor_B) + n_a, h_a, w_a, c_a = A_broadcast + n_b, h_b, w_b, c_b = B_broadcast + + scale_int, rsh, corr = get_int_scale( + scale_A, scale_B, scale_M, zero_point_A, zero_point_B, zero_point_M, "qmul" + ) + + return te.compute( + output_shape, + lambda n, h, w, c: saturate( + ( + ( + scale_int + * (tensor_A[n * n_a, h * h_a, w * w_a, c * c_a] - zero_point_A) + * (tensor_B[n * n_b, h * h_b, w * w_b, c * c_b] - zero_point_B) + + corr + ) + >> rsh + ), + "uint8", + ).astype("uint8"), + ) + + +def tir_schedule( + out_m, + tensor_A, + tensor_B, + output_layout: str, + tensor_A_layout: str, + tensor_B_layout: str, + op_name: str, +): + """Schedule for input and output layout nhwc-8h8w32c-2d""" + func = te.create_prim_func([tensor_A, tensor_B, out_m]) + + print(func) + + s = tir.Schedule(func) + + block = s.get_block("compute") + + if tensor_A_layout == "nhwc-8h8w32c-2d": + tensor_A_transformed_layout = get_layout_transform_fn(tensor_A_layout) + s.transform_layout(block, buffer=("read", 0), index_map=tensor_A_transformed_layout) Review Comment: I believe the error is caught here. buffer=("read", 0) should return tensor_A and that has the layout "nhwc-8h8w32c-2d" so we should be able to apply transform_layout() for this buffer. But buffer=("read", 0) is returning tensor_B (shape - [1, 8, 8, 1]) in this case and since it is not a multiple of crouton, the code fails -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
