ztrans_proposal.mdwn

   1 # Ztrans - transcendental operations
   2
   3 See:
   4
   5 * <http://bugs.libre-riscv.org/show_bug.cgi?id=127>
   6 * <https://www.khronos.org/registry/spir-v/specs/unified1/OpenCL.ExtendedInstructionSet.100.html>
   7
   8 Extension subsets:
   9
  10 * **Ztrans**: standard transcendentals (best suited to 3D)
  11 * **ZtransExt**: extra functions (useful, not generally needed for 3D)
  12 * **ZtransAdv**: much more complex to implement in hardware
  13
  14 [[!toc levels=2]]
  15
  16 # List of 2-arg opcodes
  17
  18 [[!table  data="""
  19 opcode    | Description           | pseudo-code                | Extension |
  20 FATAN2    | atan2 arc tangent     | rd = atan2(rs2, rs1)       | Ztrans    |
  21 FATAN2PI  | atan arc tangent / pi | rd = atan2(rs2, rs1) / pi  | ZtransExt |
  22 FPOW      | x power of y          | rd = pow(rs1, rs2)         | ZtransAdv |
  23 FROOT     | x power 1/y           | rd = pow(rs1, 1/rs2)       | ZtransAdv |
  24 """]]
  25
  26 # List of 1-arg opcodes
  27
  28 [[!table  data="""
  29 opcode   | Description              | pseudo-code             | Extension |
  30 FCBRT    | Cube Root                | rd = pow(rs1, 3)        | Ztrans    |
  31 FEXP2    | power-of-2               | rd = pow(2, rs1)        | Ztrans    |
  32 FLOG2    | log2                     | rd = log2(rs1)          | Ztrans    |
  33 FEXPM1   | exponent minus 1         | rd = pow(e, rs1) - 1.0  | Ztrans    |
  34 FLOG1P   | log plus 1               | rd = log(e, 1 + rs1)    | Ztrans    |
  35 FEXP     | exponent                 | rd = pow(e, rs1)        | ZtransExt |
  36 FLOG     | natural log (base e)     | rd = log(e, rs1)        | ZtransExt |
  37 FEXP10   | power-of-10              | rd = pow(10, rs1)       | ZtransExt |
  38 FLOG10   | log base 10              | rd = log10(rs1)         | ZtransExt |
  39 FSIN     | sin (radians)            |                         | Ztrans    |
  40 FCOS     | cos (radians)            |                         | Ztrans    |
  41 FTAN     | tan (radians)            |                         | Ztrans    |
  42 FASIN    | arcsin (radians)         | rd = asin(rs1)          | Ztrans    |
  43 FACOS    | arccos (radians)         | rd = acos(rs1)          | Ztrans    |
  44 FATAN    | arctan (radians)         | rd = atan(rs1)          | Ztrans    |
  45 FSINPI   | sin times pi             | rd = sin(pi * rs1)      | ZtransExt |
  46 FCOSPI   | cos times pi             | rd = cos(pi * rs1)      | ZtransExt |
  47 FTANPI   | tan times pi             | rd = tan(pi * rs1)      | ZtransExt |
  48 FSINH    | hyperbolic sin (radians) |                         | ZtransExt |
  49 FCOSH    | hyperbolic cos (radians) |                         | ZtransExt |
  50 FTANH    | hyperbolic tan (radians) |                         | ZtransExt |
  51 FASINH   | inverse hyperbolic sin   |                         | ZtransExt |
  52 FACOSH   | inverse hyperbolic cos   |                         | ZtransExt |
  53 FATANH   | inverse hyperbolic tan   |                         | ZtransExt |
  54 """]]
  55
  56 # Pseudo-code ops
  57
  58 * FRCP rd, rs1 - pseudo-code alias for rd = 1.0 / rs1
  59 * FATAN - pseudo-code alias for rd = atan2(rs1, 1.0) - FATAN2
  60 * FATANPI - pseudo alias for rd = atan2pi(rs1, 1.0) - FATAN2PI
  61 * FSINCOS - fused macro-op between FSIN and FCOS (issued in that order).
  62 * FSINCOSPI - fused macro-op between FSINPI and FCOSPI (issued in that order).
  63