GeForce RTX 5000, HWMonitor Restores Hotspot Temperature Monitoring. The Method Remains Unclear
The 1.65 version of HWMonitor introduces a feature that is likely to interest users of NVIDIA GeForce RTX 5000 graphics cards based on the Blackwell architecture. The update indeed reintroduces hotspot temperature monitoring, a functionality that NVIDIA had made inaccessible to third-party monitoring applications, despite the hardware sensor being present. This was also confirmed by the developer of GPU-Z, who stated that the data was no longer accessible via the public API provided by NVIDIA.
However, things might change now, and that is a good thing.
The importance of knowing the hotspot temperature goes beyond mere technical curiosity. This parameter represents the hottest point detected inside the GPU and can highlight contact issues between the chip and the heatsink that do not become apparent when solely observing the average temperature of the graphics processor.
An excessive difference between the two values can be a symptom of an uneven distribution of thermal material or suboptimal installation of the cooling system, conditions that can lead to the activation of thermal throttling and a reduction in performance during operation.
With HWMonitor 1.65, users can now view this information in real-time directly from the application. According to the initial images shared by users, the software displays two different Hotspot values: one related to the GPU and a second referred to the power delivery circuit (VRM). The availability of both readings allows for a more comprehensive view of the thermal behavior of the graphics card and enables more precise identification of any anomalies in the cooling system.
Among the first tests conducted is that of overclocker Madness, who tested the new feature using a modified GeForce RTX 5090 Frostbite with shunt mod and liquid cooling. During a brief benchmark with 3DMark 11, the system reportedly exceeded 900 watts of total power draw from the outlet. HWMonitor recorded a Hotspot temperature of approximately 68 °C, representing a difference of about 17 °C compared to the average GPU temperature, a value considered compatible with the type of cooling applied and the load subjected.
CPUID has not explained which method is used to access the data of the GeForce RTX 5000, so it is not possible to establish whether NVIDIA has restored access via the public API or if HWMonitor exploits a different undocumented monitoring path. It remains to be seen whether other widely used tools by enthusiasts, such as HWiNFO, GPU-Z, and AIDA64, will adopt a similar approach to reintroduce hotspot temperature monitoring on Blackwell cards.