After faithful service for 2.5 years, my old 7600GST has given up its ghost and died.
Considering the insane amount of abuse Crysis can give I can understand why.
So, I went ahead and got myself a replacement: an 8600GT (berate me later on the choice, I was short on cash and needed a replacement ASAP for school+work).
So I'm using it as usual, and decided to poke around on how "hot" it runs idling around with the basic non-3d stuff I do.
Its idling at 64 degrees Celsius.
Any techie person will tell you this is quite warm for a GPU core to run idling around, doing low-load work like desktop stuff and surfing.
Then again my ambient room temps are in the range of 30-37 degrees Celsius, so I can understand. Blame the iron works next door and the fact that I live in a tropical country. Yes, horrible.
I've given Nvidia a call and they say that normally the 8600GT should idle around 50, and load at 80. Rated maxing out at 127 before damage appears.
I'm not even going close to 127, but I'm still worried. Reason:
When I booted up Crysis on medium settings (autodetected settings as Crysis set, with a tweak here and there to match my aspect ratio for my widescreen), I'm getting load temperatures in excess of 85+ degrees Celsius (my alarm is set for 84 C).
I immediately closed Crysis and check my temperature logs.
I peaked at 88 C.
This is cause for concern. I've never had a videocard that ran this hot. In comparison my old 7600GST (factory overclocked) idled at 55, load at 75, and ran beautifully (albeit a bit slow with current-generation games), and even with Nvidia's support email saying otherwise, this is quite hot.
Installing Rivatuner and cranking the otherwise low-running fan (which oddly idles at 20% even under some load, WTF) up to 100% does help a bit, but still, I peak up to 90 if I dont listen to the audio warning.
I'm giving half my mind into returning the card to the shop for a no-questions-asked replacement (as I'm still within their month return-policy). Who knows, maybe I should return the card anyway. Who knows, could be just a bad card/badly seated HSF. I'd fix it myself provided that 1. I'm not hemostatic (aka, static magnet. I kill PC components by just touching them) and 2. I had some spare heat paste to put onto the GPU...