Currently job artifacts in CI/CD pipelines on LRZ GitLab never expire. Starting from Wed 26.1.2022 the default expiration time will be 30 days (GitLab default). Currently existing artifacts in already completed jobs will not be affected by the change. The latest artifacts for all jobs in the latest successful pipelines will be kept. More information: https://gitlab.lrz.de/help/user/admin_area/settings/continuous_integration.html#default-artifacts-expiration

Commit a115ed1f authored by Weronika's avatar Weronika
Browse files

added memory sensors to the nvml plugin

parent 132f8520
......@@ -37,6 +37,9 @@ nvmlConfigurator::nvmlConfigurator() {
_gpuFeatureMAP["GPU_POWER"] = GPU_POWER;
_gpuFeatureMAP["GPU_TEMP"] = GPU_TEMP;
_gpuFeatureMAP["GPU_FAN"] = GPU_FAN;
_gpuFeatureMAP["GPU_MEM_USED"] = GPU_MEM_USED;
_gpuFeatureMAP["GPU_MEM_FREE"] = GPU_MEM_FREE;
_gpuFeatureMAP["GPU_MEM_TOT"] = GPU_MEM_TOT;
_groupName = "group";
_baseName = "sensor";
......@@ -46,7 +49,7 @@ nvmlConfigurator::~nvmlConfigurator() {}
void nvmlConfigurator::sensorBase(nvmlSensorBase& s, CFG_VAL config) {
//ADD {
BOOST_FOREACH (boost::property_tree::iptree::value_type &val, config) {
if (boost::iequals(val.first, "feature")) {
gpuFeatureMap_t::iterator it = _gpuFeatureMAP.find(val.second.data());
......
......@@ -42,6 +42,9 @@ enum GPU_FEATURE {
GPU_POWER = 1,
GPU_TEMP = 2,
GPU_FAN = 3,
GPU_MEM_USED = 4,
GPU_MEM_TOT = 5,
GPU_MEM_FREE = 6,
};
/**
......@@ -109,6 +112,15 @@ class nvmlSensorBase : public SensorBase {
case GPU_FAN:
feature = "GPU_FAN";
break;
case GPU_MEM_USED:
feature = "GPU_MEM_USED";
break;
case GPU_MEM_TOT:
feature = "GPU_MEM_TOT";
break;
case GPU_MEM_FREE:
feature = "GPU_MEM_FREE";
break;
}
LOG_VAR(ll) << leading << " Feature type: " << feature;
}
......
......@@ -49,7 +49,7 @@ struct counters_t {
// unsigned int clockspeed_graphics;
// unsigned int clockspeed_sm;
// unsigned int clockspeed_mem;
// nvmlMemory_t memory;
nvmlMemory_t memory;
unsigned int power;
// unsigned long long ecc_counts;
// nvmlUtilization_t utilization;
......@@ -154,6 +154,18 @@ void nvmlSensorGroup::read() {
err = nvmlDeviceGetFanSpeed(env.device,&(counters.fanspeed));
reading.value = counters.fanspeed;
break;
case(GPU_MEM_USED):
err = nvmlDeviceGetMemoryInfo (env.device, &(counters.memory));
reading.value = counters.memory.used;
break;
case(GPU_MEM_TOT):
err = nvmlDeviceGetMemoryInfo (env.device, &(counters.memory));
reading.value = counters.memory.total;
break;
case(GPU_MEM_FREE):
err = nvmlDeviceGetMemoryInfo (env.device, &(counters.memory));
reading.value = counters.memory.free;
break;
}
s->storeReading(reading);
#ifdef DEBUG
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment