If you have read our RTX 5090 FE house test, done in the tradition of real tests, led by master’s hand by the master refineeur Eric, assisted by his nicolas gafet, you do well, and that is typically the kind of test to which it is necessary to return several times to assimilate the crazy amount of information unveiled. If you haven’t read it yet, you still have a chance to catch up before going to burn in hell with 50 Rocco Sifredi clones in rut. Frankly, better to read it! You could also see the GPU-Z captures of the RTX 5090 FE made with a GPU-Z anticipated beta version provided to W1zzard its creator on TPU. You have probably seen that there was no temperature for the Hotspot, the warmest point of the GPU. It is not a forgetting of development, but a desire not to make it appear on the RTX 50. Why?
Well according to W1zzard, Nvidia did not probe to watch him, and it would be the same for all RTX 50s on the market. In reality, the software thus refers an erroneous value of 255 ° C, a value postponed by default in the absence of a probe. In our many tests of custom graphics cards, which will resume more beautiful since the RTX 50 are announced and while waiting for the RDNA 4 cards, we have been able to establish a kind of “rule” between the temperature of the GPU and that of the Hotspot. There has been a differential ranging from 10 to 20 ° C maximum between these two surveys, at the expense of the hotspot. You can a priori apply this offset to find out what the temperature of the Hotspot is, but you will have no way of knowing it by conventional means.
-
Tech
