Home
last modified time | relevance | path

Searched refs:approximation (Results 1 – 25 of 36) sorted by relevance

12

/llvm-project-15.0.7/libc/utils/mathtools/
H A Dexpm1f.sollya11 # Polynomial approximation for e^(x) - 1 on [-Ln2, -1/8].
22 # Polynomial approximation for e^(x) - 1 on [-1/8, 1/8].
32 # Polynomial approximation for e^(x) - 1 on [1/8, Ln2].
/llvm-project-15.0.7/libclc/generic/lib/math/
H A Dasin.cl36 // where R(x^2) is a rational minimax approximation to
40 // together with the above rational approximation, and
61 // Use a rational approximation for [0.0, 0.5]
102 // where R(x^2) is a rational minimax approximation to
106 // together with the above rational approximation, and
124 // Use a rational approximation for [0.0, 0.5]
H A Dasinpi.cl36 // where R(x^2) is a rational minimax approximation to
40 // together with the above rational approximation, and
64 // Use a rational approximation for [0.0, 0.5]
105 // where R(x^2) is a rational minimax approximation to
109 // together with the above rational approximation, and
127 // Use a rational approximation for [0.0, 0.5]
H A Dacos.cl36 // where R(x^2) is a rational minimax approximation to
40 // together with the above rational approximation, and
63 // Use a rational approximation for [0.0, 0.5]
105 // where R(x^2) is a rational minimax approximation to
109 // together with the above rational approximation, and
128 // Use a rational approximation for [0.0, 0.5]
H A Dacospi.cl37 // where R(x^2) is a rational minimax approximation to
41 // together with the above rational approximation, and
64 // Use a rational approximation for [0.0, 0.5]
105 // where R(x^2) is a rational minimax approximation to
109 // together with the above rational approximation, and
129 // Use a rational approximation for [0.0, 0.5]
H A Derf.cl41 // Coefficients for approximation to erf on [00.84375]
57 // Coefficients for approximation to erf in [0.843751.25]
73 // Coefficients for approximation to erfc in [1.251/0.35]
92 // Coefficients for approximation to erfc in [1/.3528]
228 * That is, we use rational approximation to approximate
263 * We use rational approximation to approximate
377 // Compute rational approximation
H A Derfc.cl41 // Coefficients for approximation to erf on [00.84375]
57 // Coefficients for approximation to erf in [0.843751.25]
73 // Coefficients for approximation to erfc in [1.251/0.35]
92 // Coefficients for approximation to erfc in [1/.3528]
229 * That is, we use rational approximation to approximate
264 * We use rational approximation to approximate
H A Dsincos_helpers.cl40 // We use a minimax approximation of (f(w) - 1) / w
62 // We use a minimax approximation of (f(w) - 1 + w/2) / (w*w)
95 // Core Remez [1,2] approximation to tan(x) on the interval [0,pi/4].
514 // We use a minimax approximation of (f(w) - 1) / w
518 // is an approximation to cos(x)*sin(xx) valid because
524 // We use a minimax approximation of (f(w) - 1 + w/2) / (w*w)
528 // is an approximation to sin(x)*sin(xx) valid because
H A Datan.cl75 // Core approximation: Remez(2,2) on [-7/16,7/16]
151 // Core approximation: Remez(4,4) on [-7/16,7/16]
H A Datanpi.cl75 // Core approximation: Remez(2,2) on [-7/16,7/16]
149 // Core approximation: Remez(4,4) on [-7/16,7/16]
H A Dlgamma_r.cl256 // 2. Polynomial approximation of lgamma around its
263 // 2. Rational approximation in the primary interval [2,3]
264 // We use the following approximation:
282 // Let z = 1/x, then we approximation
H A Dacosh.cl34 // For 2.0 <= x <= 1/sqrt(epsilon) the approximation is
H A Dlog1p.cl107 // for the kernel approximation.
H A Datan2pi.cl141 // Polynomial approximation to atan(r)
H A Datan2.cl158 // Polynomial approximation to atan(r)
/llvm-project-15.0.7/mlir/test/mlir-cpu-runner/X86Vector/
H A Dmath-polynomial-approx-avx2.mlir1 // RUN: mlir-opt %s -test-math-polynomial-approximation="enable-avx2" \
28 // Check that the vectorized approximation is reasonably accurate.
/llvm-project-15.0.7/llvm/test/MC/Disassembler/Hexagon/
H A Dxtype_fp.txt108 # Floating point reciprocal square root approximation
140 # Floating point reciprocal approximation
/llvm-project-15.0.7/polly/lib/External/isl/doc/
H A Dimplementation.tex684 the current approximation of the affine hull.
874 we can, in the general case, only compute an approximation
1058 We will use the following approximation $Q_i$ for $P_i'$:
1126 not compute the most accurate affine approximation of
1154 of Farkas' lemma is needed to obtain the approximation of
1208 The approximation for $k\,\Delta$,
1217 Finally, the computed approximation for $R^+$,
1330 the following approximation for $R^+$:
1343 and we obtain the approximation
1419 is clearly not exact and the approximation of the power $K$
[all …]
/llvm-project-15.0.7/compiler-rt/lib/builtins/
H A Dfp_div_impl.inc105 // polynomial approximation: x0 = 3/4 + 1/sqrt(2) - b/2.
106 // The max error for this approximation is achieved at endpoints, so
109 // The initial approximation is between x0(1.0) = 0.9571... and x0(2.0) = 0.4571...
160 // * x0 being an inherently imprecise first approximation of 1/b_hw
206 // approximation after last iteration.
/llvm-project-15.0.7/mlir/lib/Dialect/Math/Transforms/
H A DPolynomialApproximation.cpp754 Value approximation = builder.create<arith::SelectOp>( in matchAndRewrite() local
756 rewriter.replaceOp(op, approximation); in matchAndRewrite()
1061 Value approximation = builder.create<arith::SelectOp>( in matchAndRewrite() local
1064 rewriter.replaceOp(op, approximation); in matchAndRewrite()
1198 Value approximation = select(negativeRange, mul(cstNegativeOne, v6), v6); in matchAndRewrite() local
1200 rewriter.replaceOp(op, approximation); in matchAndRewrite()
/llvm-project-15.0.7/mlir/test/Dialect/Math/
H A Dpolynomial-approximation.mlir1 // RUN: mlir-opt %s -test-math-polynomial-approximation | FileCheck %s
2 // RUN: mlir-opt %s -test-math-polynomial-approximation=enable-avx2 \
540 // ATan approximation:
/llvm-project-15.0.7/llvm/docs/
H A DBranchWeightMetadata.rst205 sampling-based profiles, this operand is an approximation of how many times
/llvm-project-15.0.7/llvm/utils/vim/
H A Dvimrc50 " details. These aren't absolute rules; they're just an approximation of
/llvm-project-15.0.7/clang/test/Analysis/
H A DDeallocUseAfterFreeErrors.m266 // approximation to trigger the false positive.
/llvm-project-15.0.7/llvm/test/CodeGen/PowerPC/
H A Dfmf-propagation.ll430 ; The call is now fully 'fast'. This implies that approximation is allowed.
488 ; The call is now fully 'fast'. This implies that approximation is allowed.

12