gem5/src/mem/ruby/common/Histogram.cc

/*
 * Copyright (c) 1999-2008 Mark D. Hill and David A. Wood
 * All rights reserved.
 *
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions are
 * met: redistributions of source code must retain the above copyright
 * notice, this list of conditions and the following disclaimer;
 * redistributions in binary form must reproduce the above copyright
 * notice, this list of conditions and the following disclaimer in the
 * documentation and/or other materials provided with the distribution;
 * neither the name of the copyright holders nor the names of its
 * contributors may be used to endorse or promote products derived from
 * this software without specific prior written permission.
 *
 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
 * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
 * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
 * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
 * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
 * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
 * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
 * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
 * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
 * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 */

#include <cmath>
#include <iomanip>

#include "base/intmath.hh"
#include "mem/ruby/common/Histogram.hh"

using namespace std;

Histogram::Histogram(int binsize, uint32_t bins)
{
    m_binsize = binsize;
    clear(bins);
}

Histogram::~Histogram()
{
}

void
Histogram::clear(int binsize, uint32_t bins)
{
    m_binsize = binsize;
    clear(bins);
}

void
Histogram::clear(uint32_t bins)
{
    m_largest_bin = 0;
    m_max = 0;
    m_data.resize(bins);
    for (uint32_t i = 0; i < bins; i++) {
        m_data[i] = 0;
    }

    m_count = 0;
    m_max = 0;
    m_sumSamples = 0;
    m_sumSquaredSamples = 0;
}

void
Histogram::doubleBinSize()
{
    assert(m_binsize != -1);
    uint32_t t_bins = m_data.size();

    for (uint32_t i = 0; i < t_bins/2; i++) {
        m_data[i] = m_data[i*2] + m_data[i*2 + 1];
    }
    for (uint32_t i = t_bins/2; i < t_bins; i++) {
        m_data[i] = 0;
    }

    m_binsize *= 2;
}

void
Histogram::add(int64 value)
{
    assert(value >= 0);
    m_max = max(m_max, value);
    m_count++;

    m_sumSamples += value;
    m_sumSquaredSamples += (value*value);

    uint32_t index;

    if (m_binsize == -1) {
        // This is a log base 2 histogram
        if (value == 0) {
            index = 0;
        } else {
            index = floorLog2(value) + 1;
            if (index >= m_data.size()) {
                index = m_data.size() - 1;
            }
        }
    } else {
        // This is a linear histogram
        uint32_t t_bins = m_data.size();

        while (m_max >= (t_bins * m_binsize)) doubleBinSize();
        index = value/m_binsize;
    }

    assert(index < m_data.size());
    m_data[index]++;
    m_largest_bin = max(m_largest_bin, index);
}

void
Histogram::add(Histogram& hist)
{
    uint32_t t_bins = m_data.size();

    if (hist.getBins() != t_bins) {
        if (m_count == 0) {
            m_data.resize(hist.getBins());
        } else {
            fatal("Histograms with different number of bins "
                  "cannot be combined!");
        }
    }

    m_max = max(m_max, hist.getMax());
    m_count += hist.size();
    m_sumSamples += hist.getTotal();
    m_sumSquaredSamples += hist.getSquaredTotal();

    // Both histograms are log base 2.
    if (hist.getBinSize() == -1 && m_binsize == -1) {
        for (int j = 0; j < hist.getData(0); j++) {
            add(0);
        }

        for (uint32_t i = 1; i < t_bins; i++) {
            for (int j = 0; j < hist.getData(i); j++) {
                add(1<<(i-1));  // account for the + 1 index
            }
        }
    } else if (hist.getBinSize() >= 1 && m_binsize >= 1) {
        // Both the histogram are linear.
        // We are assuming that the two histograms have the same
        // minimum value that they can store.

        while (m_binsize > hist.getBinSize()) hist.doubleBinSize();
        while (hist.getBinSize() > m_binsize) doubleBinSize();

        assert(m_binsize == hist.getBinSize());

        for (uint32_t i = 0; i < t_bins; i++) {
            m_data[i] += hist.getData(i);

            if (m_data[i] > 0) m_largest_bin = i;
        }
    } else {
        fatal("Don't know how to combine log and linear histograms!");
    }
}

// Computation of standard deviation of samples a1, a2, ... aN
// variance = [SUM {ai^2} - (SUM {ai})^2/N]/(N-1)
// std deviation equals square root of variance
double
Histogram::getStandardDeviation() const
{
    if (m_count <= 1)
        return 0.0;

    double variance =
        (double)(m_sumSquaredSamples - m_sumSamples * m_sumSamples / m_count)
        / (m_count - 1);
    return sqrt(variance);
}

void
Histogram::print(ostream& out) const
{
    printWithMultiplier(out, 1.0);
}

void
Histogram::printPercent(ostream& out) const
{
    if (m_count == 0) {
        printWithMultiplier(out, 0.0);
    } else {
        printWithMultiplier(out, 100.0 / double(m_count));
    }
}

void
Histogram::printWithMultiplier(ostream& out, double multiplier) const
{
    if (m_binsize == -1) {
        out << "[binsize: log2 ";
    } else {
        out << "[binsize: " << m_binsize << " ";
    }
    out << "max: " << m_max << " ";
    out << "count: " << m_count << " ";
    //  out << "total: " <<  m_sumSamples << " ";
    if (m_count == 0) {
        out << "average: NaN |";
        out << "standard deviation: NaN |";
    } else {
        out << "average: " << setw(5) << ((double) m_sumSamples)/m_count
            << " | ";
        out << "standard deviation: " << getStandardDeviation() << " |";
    }

    for (uint32_t i = 0; i <= m_largest_bin; i++) {
        if (multiplier == 1.0) {
            out << " " << m_data[i];
        } else {
            out << " " << double(m_data[i]) * multiplier;
        }
    }
    out << " ]";
}

bool
node_less_then_eq(const Histogram* n1, const Histogram* n2)
{
    return (n1->size() > n2->size());
}