AI/rocm-systems
2
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

SWDEV-193956 - [hipclang-vdi-rocm][perf]

Browse Source 
~45% to 50% of Performance drop on rocBLAS_int8 test

Use the last command in the queue for a wait.
Add extra print information about processed commands.
Add an option to disable file location printing.

Change-Id: I4187883e1a90e571fde3128af98368108fda8785
This commit is contained in:
German Andryeyev
2020-02-20 16:24:10 -05:00
parent fea14a7201
commit a66d09f5a3
3 changed files with 36 additions and 11 deletions
Download Patch File Download Diff File Expand all files Collapse all files
rocclr/platform/command.cpp
+13 -2
View File
@@ -85,6 +85,9 @@ uint64_t Event::recordProfilingInfo(int32_t status, uint64_t timeStamp) {
return timeStamp;
}
// Global epoch time since the first processed command
uint64_t epoch = 0;
bool Event::setStatus(int32_t status, uint64_t timeStamp) {
assert(status <= CL_QUEUED && "invalid status");
@@ -96,6 +99,9 @@ bool Event::setStatus(int32_t status, uint64_t timeStamp) {
if (profilingInfo().enabled_) {
timeStamp = recordProfilingInfo(status, timeStamp);
if (epoch == 0) {
epoch = profilingInfo().queued_;
}
}
if (!make_atomic(status_).compareAndSet(currentStatus, status)) {
@@ -112,8 +118,6 @@ bool Event::setStatus(int32_t status, uint64_t timeStamp) {
}
if (status <= CL_COMPLETE) {
ClPrint(LOG_DEBUG, LOG_CMD, "command %p complete", &command());
// Before we notify the waiters that this event reached the CL_COMPLETE
// status, we release all the resources associated with this instance.
releaseResources();
@@ -123,6 +127,13 @@ bool Event::setStatus(int32_t status, uint64_t timeStamp) {
if (referenceCount() > 1) {
signal();
}
ClPrint(LOG_DEBUG, LOG_CMD, "command %p complete (Wall: %ld, CPU: %ld, GPU: %ld us)",
&command(),
((profilingInfo().end_ - epoch) / 1000),
((profilingInfo().submitted_ - profilingInfo().queued_) / 1000),
((profilingInfo().end_ - profilingInfo().start_) / 1000));
release();
}
rocclr/platform/commandqueue.cpp
+17 -8
View File
@@ -86,15 +86,24 @@ bool HostQueue::terminate() {
}
void HostQueue::finish() {
// Send a finish to make sure we finished all commands
Command* command = new Marker(*this, false);
if (command == NULL) {
return;
Command* command = nullptr;
if (IS_HIP) {
command = getLastQueuedCommand(false);
if (nullptr != command) {
command->awaitCompletion();
}
}
if (nullptr == command) {
// Send a finish to make sure we finished all commands
command = new Marker(*this, false);
if (command == NULL) {
return;
}
ClPrint(LOG_DEBUG, LOG_CMD, "marker is queued");
command->enqueue();
command->awaitCompletion();
command->release();
}
ClPrint(LOG_DEBUG, LOG_CMD, "marker is queued");
command->enqueue();
command->awaitCompletion();
command->release();
ClPrint(LOG_DEBUG, LOG_CMD, "All commands finished");
}
rocclr/utils/debug.hpp
+6 -1
View File
@@ -47,6 +47,7 @@ enum LogMask {
LOG_AQL2 = 0x00002000, //!< Show raw bytes of AQL packet
LOG_CODE = 0x00004000, //!< Show code creation debug
LOG_CMD2 = 0x00008000, //!< More detailed command info, including barrier commands
LOG_LOCATION = 0x00010000, //!< Log message location
LOG_ALWAYS = 0xFFFFFFFF, //!< Log always even mask flag is zero
};
@@ -178,7 +179,11 @@ inline void warning(const char* msg) { amd::report_warning(msg); }
do { \
if (LOG_LEVEL >= level) { \
if (GPU_LOG_MASK & mask || mask == amd::LOG_ALWAYS) { \
amd::log_printf(level, __FILE__, __LINE__, format, ##__VA_ARGS__); \
if (GPU_LOG_MASK & amd::LOG_LOCATION) { \
amd::log_printf(level, __FILE__, __LINE__, format, ##__VA_ARGS__); \
} else { \
amd::log_printf(level, "", 0, format, ##__VA_ARGS__); \
} \
} \
} \
} while (false)