Skip to content

Fix errors with 2+ GPUs and simplify total speed calculation #22

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
zZoMROT wants to merge 2 commits into
base: master
Choose a base branch
from
Open

Fix errors with 2+ GPUs and simplify total speed calculation #22

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
0b86f12
Fix errors with 4+ GPUs and simplify total speed calculation
zZoMROT Jul 17, 2024
a5b6d63
Remove unused code
zZoMROT Jul 17, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
18 changes: 7 additions & 11 deletions Speed.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -35,7 +35,6 @@ void Speed::update(const unsigned int numPoints, const unsigned int indexDevice)
const auto ns = std::chrono::steady_clock::now().time_since_epoch().count();
const bool bPrint = ((ns - m_lastPrint) / 1000000) > m_intervalPrintMs;

updateList(numPoints, ns, m_lSamples);
updateList(numPoints, ns, m_mDeviceSamples[indexDevice]);

if (bPrint) {
Expand All @@ -44,10 +43,6 @@ void Speed::update(const unsigned int numPoints, const unsigned int indexDevice)
}
}

double Speed::getSpeed() const {
return this->getSpeed(m_lSamples);
}

double Speed::getSpeed(const unsigned int indexDevice) const {
return m_mDeviceSamples.count(indexDevice) == 0 ? 0 : this->getSpeed(m_mDeviceSamples.at(indexDevice));
}
Expand All @@ -70,20 +65,21 @@ void Speed::updateList(const unsigned int & numPoints, const long long & ns, sam
l.push_back(samplePair(ns, numPoints));

// Pop old samples until time difference between first and last element is less than or equal to m_sampleSeconds
// We don't need to check size of m_lSamples since it's always >= 1 at this point
while (l.size() > 2 && (l.back().first - l.front().first) / 1000000 > m_intervalSampleMs) {
l.pop_front();
}
}

void Speed::print() const {
const std::string strVT100ClearLine = "\33[2K\r";
std::cout << strVT100ClearLine << "Speed: " << formatSpeed(this->getSpeed());

double totalSpeed = 0.0;
std::ostringstream oss;
// std::map is sorted by key so we'll always have the devices in numerical order
for (auto it = m_mDeviceSamples.begin(); it != m_mDeviceSamples.end(); ++it) {
std::cout << " GPU" << it->first << ": " << formatSpeed(this->getSpeed(it->second));
const double speed = this->getSpeed(it->second);
totalSpeed += speed;
oss << " GPU" << it->first << ": " << formatSpeed(speed);
}

std::cout << "\r" << std::flush;
const std::string strVT100ClearLine = "\33[2K\r";
std::cout << strVT100ClearLine << "Speed: " << formatSpeed(totalSpeed) << oss.str() << "\r" << std::flush;
}
2 changes: 0 additions & 2 deletions Speed.hpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -18,7 +18,6 @@ class Speed {
void update(const unsigned int numPoints, const unsigned int indexDevice);
void print() const;

double getSpeed() const;
double getSpeed(const unsigned int indexDevice) const;

private:
Expand All @@ -31,7 +30,6 @@ class Speed {

long long m_lastPrint;
mutable std::recursive_mutex m_mutex;
sampleList m_lSamples;
std::map<unsigned int, sampleList> m_mDeviceSamples;
};

Expand Down