# # Copyright (C) 2016 Intel Corporation # # Permission is hereby granted, free of charge, to any person obtaining a # copy of this software and associated documentation files (the "Software"), # to deal in the Software without restriction, including without limitation # the rights to use, copy, modify, merge, publish, distribute, sublicense, # and/or sell copies of the Software, and to permit persons to whom the # Software is furnished to do so, subject to the following conditions: # # The above copyright notice and this permission notice (including the next # paragraph) shall be included in all copies or substantial portions of the # Software. # # THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR # IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, # FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL # THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER # LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING # FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS # IN THE SOFTWARE. import nir_algebraic # Prior to Kaby Lake, The SIN and COS instructions on Intel hardware can # produce values slightly outside of the [-1.0, 1.0] range for a small set of # values. Obviously, this can break everyone's expectations about trig # functions. This appears to be fixed in Kaby Lake. # # According to an internal presentation, the COS instruction can produce # a value up to 1.000027 for inputs in the range (0.08296, 0.09888). One # suggested workaround is to multiply by 0.99997, scaling down the # amplitude slightly. Apparently this also minimizes the error function, # reducing the maximum error from 0.00006 to about 0.00003. trig_workarounds = [ (('fsin', 'x'), ('fmul', ('fsin', 'x'), 0.99997)), (('fcos', 'x'), ('fmul', ('fcos', 'x'), 0.99997)), ] print '#include "brw_nir.h"' print nir_algebraic.AlgebraicPass("brw_nir_apply_trig_workarounds", trig_workarounds).render()