Situatie
The situation may require the diagnosis of an unexpected behavior – in addition to using the appropriate tools, it is also important to perform monitoring at an appropriate time scale in order to avoid missing short-term behaviors.
Solutie
1. We download and run the HWINFO64 application – we are interested in monitoring the system parameters (temperature, speeds, usage percentages, etc.), which is why we will use the “sensors only” check box: https://www.hwinfo.com/files/hwi_750.zip (the link is for the portable version of the application).
2. Press Run and a new window will open – we will navigate to change the settings (bottom right corner of the window), respectively we will change the sampling rate from one second (1 Hz) to 100ms (10Hz):
Warning! Increasing the update rate will also lead to increased processor usage and temperatures observed on it.
3. Having access to the “new” information that was implicitly “hidden” by the low sampling rate, we can observe new behaviors such as:
– the stability of the applied voltages (observed in tasks that require computing power compared to the values observed in “idle”/rest)
– the real percentage of time spent at a certain voltage that is more or less dangerous relative to the lifetime of the components
-observation of very short-duration tasks that can produce interruptions and an unpleasant system usage environment, both by monitoring usage percentages and temperatures
-stability of transfer speeds (for example for a DRAM-less SSD (no RAM memory for buffer) when copying a large file, an average speed of 300 MBPS can be observed, and with the new settings in HWINFO we will see that the real speed is for example 200MBPS with short jumps (hundreds of milliseconds) to values of 700-900MBPS).
Leave A Comment?