How are people with the gtx 970?

[watercooled]

If you do it right, even a R9 390 can sit quiet comfortably in a ITX case. Don't give in the hype

People love getting all worked up over 20W! http://www.techspot.com/review/1075-...015/page6.html

It's important to look at power consumption under the type of load you'll be doing e.g. in games. Higher power consumption under a power virus like Furmark (what a lot of sites put down as 'power consumption') is a pretty useless data point for comparing cards for gaming efficiency. Power-virus power consumption basically just reveals where Nvidia/AMD set their power limiter.

[Thorsson]

Funny that; Tom's Hardware shows after 10 mins of BF4 the 390 was using 254.5 Watts, while the 970 was using 171. That matches what friends have told me about their room temps when running the 390...

[kalniel]

Funny that; Tom's Hardware shows after 10 mins of BF4 the 390 was using 254.5 Watts, while the 970 was using 171. That matches what friends have told me about their room temps when running the 390...

If you could noticeably change the temp of room with 80W then all these 2kW fan heaters have been doing it wrong.

[1/2Genius1/2Wit]

If you could noticeably change the temp of room with 80W then all these 2kW fan heaters have been doing it wrong.

Agreed here - I've only noticed a room noticeably heated by computers when there are 5 systems in it...

Back on topic, I can thoroughly recommend the GTX 970. I currently have 2 in different systems. It's fast, can be cooled quietly and it overclocks well. It's a little bit overkill at 1080p as mentioned, but pair it up with a 144Hz display and you'll suddenly appreciate that overkill quite a lot! The only game where I can force an issue with the RAM is GTA V running at 2560x1440 downscaled.

[Thorsson]

My computer noticeably changes my room temp if I close the door. Put a 2kW fan heater in here and it would noticeably change the temperature in a minute. 80W x 1 hour is the same as 2kW for 2.4 minutes...

[watercooled]

Heat is lost through walls/windows so you can't just count joules - the timescale is massively important too.

[kalniel]

My computer noticeably changes my room temp if I close the door.

Several hundred watts might change a room temp if it's a small closed room, but 80W won't make a noticeable difference.

[Agrippa]

My entire flat is heated by a 500W heat pump. Those 500W are enough to comfortably heat 80m2 even with an outside temperature of -27C (which is the coldest I've experienced here). So yes, 80W would certainly make a noticeable difference in a small, closed room.

[Thesman106]

So i have been playing around with my r9 390 and its been amazing. My temperatures are also quite decents. At max load i get 50c with 35% fan speed with an ambient room temp of 23. To be honest you couldnt go wrong with either and just matter preference of what games you like that runs better

[kalniel]

My entire flat is heated by a 500W heat pump. Those 500W are enough to comfortably heat 80m2 even with an outside temperature of -27C (which is the coldest I've experienced here).

That's impressive - extreme insulation and more keeping up to temp than bringing up from a low start?

[Agrippa]

Perfectly ordinary insulation, or ordinary for Norway at least. It's an old house and the insulation was upgraded to modern standard some years back. No extreme measures though. It takes the temp up from around 10C (which is what I turn it to when I'm away for a while) to 20C in maybe an hour and a half, though admittedly that's assisted a little by the heat from the floor below.

[watercooled]

Heat pumps have an 'efficiency' of greater than 100%, or more accurately their COP is >1, where a purely resistive heater is 1. So depending on what sort of heat pump it is, for that 500W of electrical power you're actually getting in the area of 1500 to 3000W of thermal energy from it (or maybe more depending on the installation) .

So in reality you have far more than 500W heating your home - that's the whole point of using heat pumps rather than a resistive heater.

[CAT-THE-FIFTH]

People obsess about the graphics card,then don't appear to put anywhere as much effort into the efficiency of the rest of the parts in their system,such as the CPU,motherboard,PSU,monitor,etc and then overclock on top of that too.

[Agrippa]

Heat pumps have an 'efficiency' of greater than 100%.... So in reality you have far more than 500W heating your home - that's the whole point of using heat pumps rather than a resistive heater.

Of course, I'm well aware of that. Nevertheless I have no doubt that an additional 80W of heat output would make an appreciable difference to "a small, enclosed room". The description perfectly fits my current computer room/spare bedroom and I certainly noticed a difference in there after I built my current PC, which probably draws around 100-120W more than the old one (as a very quick, rough estimate).

[watercooled]

Sorry for going somewhat off-topic, but in reply to CAT I've found motherboard choice can make a huge difference to power consumption, and here we're talking largely about idle power consumption you'll get the entire time the system is powered on rather than a GPU's load power consumption you'll only get when actively playing games - modern processors are really very good at power gating when idle.

In fact in many if not most systems, the motherboard seems to be the biggest idle power consumer now. Smaller boards tend to fare better, and in general it seems like ASRock is one of the better manufacturers in this regard (at least from my own experience and from reviews I've read).

Of course, overclocking can destroy efficiency too - looking at Skylake power consumption it already seems like the 6700k is running at an uncomfortably high clock speed, same as the 4790k (and for what it's worth the FX-9000 series are well outside of the efficiency curve). People will claim efficiency is crucially important, then buy an inefficiently-clocked processor, and do even more damage by overclocking it. :shrug:

Monitors are another one people don't seem to consider - an awful lot of monitors ship with 100% backlight, which unless you're living on the surface of the Sun is probably massive overkill and will probably just wear out the backlights more quickly. Properly setting up a monitor and reducing the backlight to sane levels will normally cut a fair bit off their energy consumption too.