LOOOL, people think this is a consumer product.

Its for people who make games or VFX. Better yet, its for companies that make companies that make games+VFX

Delicious Quadro is delicious. If only it were that useful for AE then things would be even more awesome.

If I was a millionare I'd do 4 of these in quad SLI.

I'd use OCZ ZD-XL SQL Accelerator as my boot drive and dual Intel Xeon E5-2690 as my CPU.

Then I'd play Plants VS Zombies.

AMD's CPU is a consumer market product.   This GPU is a business market product.

Totally separate markets.  nVidia's workstation GPU department really doesn't give a damn what AMD does in the consumer CPU deparment and vice versa.

Perhapse not with Nvida or AMD, but it sure has with ARM and Intel's chips.  The next couple of years should show be interesting.

Didn't Nvidia show that in a year or two they will have a 40x reduction in power soon?  Here is an other Intel chart. They have a few Haswells that hit the 6.5 watt (total.)

ARM has only because they were late to 28nm. They should have had that in 2012.

Intel's Ivy Bridge was 2012, Haswell is not a perf/watt improvement, and Broadwell/14nm isn't coming to desktops.

Nvidia are known for outright lying on charts.

No, the Haswells will not hit 6.5W. That's "scenario design power", i.e. make believe "typical" use numbers. Even TDP is lying. Intel don't even publish true max power numbers any more.

Also, note PERF/watt. Simply lowering power by lowering performance does not count. Intel could make a 0.1W chip if they wanted, it just wouldn't boot Windows.

Sigh, not sure why you don't want to see improvements but here is one:  http://finance.yahoo.com/news/intel-launch-low-power-version-172828668.html

http://news.softpedia.com/news/Teaser-Intel-Low-Power-Haswell-CPU-for-Tablets-370530.shtml

"Intel calculates TDP (thermal design power) through a set of benchmarks that include some of the most troublesome desktop and notebook chips. The SDP thermal rating uses a lesser set of benchmarks though, more appropriate for thin tablets, to determine average power.

This, AnandTech explains, allows Intel to set power ratings depending on the tasks that chips are aimed for.

Thus, a new Y-series Haswell CPU that Intel is preparing could have two ratings.

If one were to, say, run Furmark on the chip, and the PC doesn't thermally limit the processor, the TDP will be of 11.5W.

When using the chip like in a tablet, however, the TDP is of 4.5 or 6W, depending on SKU. Lower than the marks specified on the slide above.

Intel will release the Y-series unit and (probably) advertise it under the 4.5W thermal envelope, in the hopes of scoring more tablet design wins.

No details exist on the chips, however, not even an actual name. Still, this all goes to show that Intel's efforts to make its chips more power-efficient are paying off."

Claiming to launch low power XYZ doesn't mean it uses less power.

Your description of SDP and TDP matches what I said: cherry-picked benchmarks that don't give you max power.

You haven't given me evidence of increased performance-per-watt yet.

MM, Ok, well no point in trying to convince you then.  I'll just leave these for others reading it.

"Intel has already launched a line of its Atom mobile chips that are tweaked to work as low-power server chips." (from the article in the previous post)

*This was project, in reality Haswell hit much more efficient project, see previous article where it was 6.5W which is better than the 13W they expected. 

Both 6.5W and 13W bear no resemblance to actual power consumption. No matter how many times you bold "LOW POWER".

If you'd like I can get an actual chip engineer to come and explain this to you.

In the server configuration, neither part is actually competitive on performance/watt versus ARM. So it does not represent an advance. And as I said they crippled performance to get it down that low in power.

And anyway this thread is talking about DESKTOP performance, right? That's what we care about? And that's what I'm claiming no advances since 2012 for.