CSRFunctions

#include <poplar/CSRFunctions.hpp>
namespace poplar

Poplar classes and functions.

Functions

void setFloatingPointBehaviour(poplar::Graph &graph, poplar::program::Sequence &prog, const FloatingPointBehaviour &behaviour, const DebugContext &debugContext = {})

Set the floating point behaviour of a tile.

Configures the floating point behaviour of a tile, affecting the treatment of exceptions and selecting stochastic rounding according to the passed behaviour structure.

Note that, in Poplar, stochastic rounding is disabled by default until either this function, setStochasticRounding() or the Engine options are used to enable it.

Parameters

  • graph – The Poplar graph.

  • prog – The program to be extended.

  • behaviour – A structure of type floatingPointBehaviour.

  • debugPrefix – The prefix prepended to debugging info.

void setFloatingPointBehaviour(poplar::Graph &graph, poplar::program::Sequence &prog, const poplar::Tensor &behaviour, const DebugContext &debugContext = {})

Set the floating point behaviour of a tile.

Configures the floating point behaviour of a tile, affecting the treatment of exceptions and selecting stochastic rounding according to the passed behaviour tensor.

The behaviour tensor must be one returned by getAndModifyFloatingPointBehaviour().

Parameters

  • graph – The Poplar graph.

  • prog – The program to be extended.

  • behaviour – A tensor containing representation of floatingPointBehaviour.

  • debugPrefix – The prefix prepended to debugging info.

void setStochasticRounding(poplar::Graph &graph, poplar::program::Sequence &prog, bool behaviour, const DebugContext &debugContext = {})

Set stochastic rounding on or off for the selected tile.

Configures the stochastic rounding operation of a tile according to the passed behaviour parameter.

Note that, in Poplar, stochastic rounding is disabled by default until either this function, setFloatingPointBehaviour() or the Engine options are used to enable it.

Parameters

  • graph – The Poplar graph.

  • prog – The program to be extended.

  • behaviour – Select stochastic rounding: true or false.

  • debugPrefix – The prefix prepended to debugging info.

poplar::Tensor getAndModifyFloatingPointBehaviour(poplar::Graph &graph, poplar::program::Sequence &prog, const FloatingPointBehaviour &clear, const FloatingPointBehaviour &set, const DebugContext &debugContext = {})

Get the current state and modify the floating-point behaviour on every tile that belongs to the target in the graph.

Returns the previous state and modifies the behaviour. Behaviour modification first clears the behaviours set in clear followed by setting the behaviours set in set.

This can be used to set the behaviour for a section of code, and then restore the original behaviour afterwards. The recommended usage of this function should be as follows to avoid unexpected numerical behaviour: 

 ...
 auto state = getAndModifyFloatingPointBehaviour(...)
 // operations that require the modified FP behaviour
 ...
 setFloatingPointBehaviour(state)

Parameters

  • graph – The Poplar graph.

  • prog – The program to be extended.

  • clear – Select the behaviours to clear. If a field in clear is set to true, then that behaviour will be cleared. If a field is false then the behaviour is unchanged.

  • set – Select behaviours to set. If a field in set is set to true, then that behaviour will be set. The values in set are applied after the values in clear. So if a field is set to true in both clear and set, then the behaviour will be set.

  • debugPrefix – The prefix prepended to debugging info.

Returns

The state before the floating-point behaviour is modified.

struct FloatingPointBehaviour
#include <CSRFunctions.hpp>

Structure to specify floating point behaviour.

Param inv

If true, a floating-point invalid operation (defined by IEEE 754) will cause an exception.

The invalid operations are:

  • Addition or subtraction where the operands are + or - infinity (inf) and the operation results in the subtraction of two infs; for example: (-inf)+(+inf) or (+inf)-(+inf).

  • Divisions: (+/-0)/(+/-0) and (+/-inf)/(+/-inf).

  • Multiplications: (+/-0)*(+/-inf) and (+/-inf)*(+/-0).

  • Remainder: x REM y where y=0 or x=(+/-inf).

  • Real operations with complex results such as the square root or logarithm of a negative number.

  • Operations with Not-a-Number as at least one operand.

  • Comparisons where one of the operands is Not-a-Number.

    See also nanoo below. It can also be set in the environment variable

    POPLAR_ENGINE_OPTIONS by setting debug.floatPointOpException to true.

Param div

If true a floating point divide by zero operation will cause an exception.

Param oflo

If true a floating point overflow will cause an exception.

Param esr

Enable stochastic rounding.

Param nanoo

Enable Not-a-Number on overflow mode. When enabled, half precision calculations that have overflowed will produce a Not-a-Number result, rather than saturating to the half precision max/min value, and the invalid operation (inv) flag will be set. It can also be enabled in the environment variable POPLAR_ENGINE_OPTIONS by setting debug.nanOverflowMode to true.

Public Functions

inline FloatingPointBehaviour(bool inv, bool div0, bool oflo, bool esr, bool nanoo)
FloatingPointBehaviour() = default
inline FloatingPointBehaviour operator!() const

Public Members

bool inv = true
bool div0 = true
bool oflo = true
bool esr = true
bool nanoo = true