Dunedan/0ad
1
0
Fork 0
forked from 0ad/0ad
Code Pull Requests Wiki Activity

Reduces number of allocations during GPU profiler processing for Intel queries.

Browse Source 
This was SVN commit r26011.
This commit is contained in:
Vladislav Belov 2021-11-25 17:33:17 +00:00
parent 265a2246f0
commit 6efa293fd1
1 changed files with 11 additions and 12 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

23
source/ps/Profiler2GPU.cpp
View File

@ -25,12 +25,12 @@
#include "Profiler2GPU.h"
#include "lib/ogl.h"
#include "lib/allocators/shared_ptr.h"
#include "ps/ConfigDB.h"
#include "ps/Profiler2.h"
#include <deque>
#include <stack>
#include <vector>
#if !CONFIG2_GLES
@ -605,6 +605,7 @@ private:
void ProcessFrames()
{
std::vector<char> buffer;
while (!m_Frames.empty())
{
SFrame& frame = m_Frames.front();
@ -613,8 +614,8 @@ private:
// trying to read the results from any
for (size_t j = 0; j < m_QueryTypes.size(); ++j)
{
size_t size = m_QueryTypes[j].counterBufferSize;
std::shared_ptr<char> buf(new char[size], ArrayDeleter());
const size_t size = m_QueryTypes[j].counterBufferSize;
buffer.resize(size);
for (size_t i = 0; i < frame.events.size(); ++i)
{
@ -622,7 +623,7 @@ private:
continue;
GLuint length = 0;
pglGetPerfQueryDataINTEL(frame.events[i].queries[j], INTEL_PERFQUERIES_NONBLOCK, size, buf.get(), &length);
pglGetPerfQueryDataINTEL(frame.events[i].queries[j], INTEL_PERFQUERIES_NONBLOCK, size, buffer.data(), &length);
ogl_WarnIfError();
if (length == 0)
return;
@ -648,8 +649,8 @@ private:
for (size_t j = 0; j < m_QueryTypes.size(); ++j)
{
GLuint length;
char* buf = new char[m_QueryTypes[j].counterBufferSize];
pglGetPerfQueryDataINTEL(frame.events[i].queries[j], INTEL_PERFQUERIES_BLOCK, m_QueryTypes[j].counterBufferSize, buf, &length);
buffer.resize(m_QueryTypes[j].counterBufferSize);
pglGetPerfQueryDataINTEL(frame.events[i].queries[j], INTEL_PERFQUERIES_BLOCK, m_QueryTypes[j].counterBufferSize, buffer.data(), &length);
ogl_WarnIfError();
ENSURE(length == m_QueryTypes[j].counterBufferSize);
@ -663,14 +664,14 @@ private:
{
ENSURE(counter.size == 4);
GLuint value = 0;
memcpy(&value, buf + counter.offset, counter.size);
memcpy(&value, buffer.data() + counter.offset, counter.size);
m_Storage.RecordAttributePrintf("%s: %u", counter.name.c_str(), value);
}
else if (counter.type == INTEL_PERFQUERIES_TYPE_UNSIGNED_INT64)
{
ENSURE(counter.size == 8);
GLuint64 value = 0;
memcpy(&value, buf + counter.offset, counter.size);
memcpy(&value, buffer.data() + counter.offset, counter.size);
m_Storage.RecordAttributePrintf("%s: %.0f", counter.name.c_str(), (double)value);
if (counter.name == "TotalTime")
@ -680,14 +681,14 @@ private:
{
ENSURE(counter.size == 4);
GLfloat value = 0;
memcpy(&value, buf + counter.offset, counter.size);
memcpy(&value, buffer.data() + counter.offset, counter.size);
m_Storage.RecordAttributePrintf("%s: %f", counter.name.c_str(), value);
}
else if (counter.type == INTEL_PERFQUERIES_TYPE_BOOL)
{
ENSURE(counter.size == 4);
GLuint value = 0;
memcpy(&value, buf + counter.offset, counter.size);
memcpy(&value, buffer.data() + counter.offset, counter.size);
ENSURE(value == 0 || value == 1);
m_Storage.RecordAttributePrintf("%s: %u", counter.name.c_str(), value);
}
@ -696,8 +697,6 @@ private:
//debug_warn(L"unrecognised Intel performance counter type");
}
}
delete[] buf;
}
endTimes.push(lastTime + elapsed);