Slashdot Mirror
The NVIDIA GeForce 7900 Series
An anonymous reader writes "HardOCP has posted their evaluation of the new GeForce 7900 technology. They fully cover widescreen gaming this time around too. 'NVIDIA has worked hard to try and produce a more powerful, albeit power-efficient GPU in the 7900 GTX and GT, and they've succeeded. They run cooler; are smaller, have less transistors, and they don't make you stuff cotton in your ears. The 7900 GTX and GT are just more efficient while being lightning fast.'"
And yet another graphics card is released. Is it worth my money to upgrade my dual 6800 XTs? Let's find out by reading the review.
Unfortunately, I can't. I'm better off going to NVidia and trusting their product sheets. Why? Because I'm not looking to play Need for Speed Most Wanted or Quake Four or Half Life Two, I'm looking to do some actual graphics processing with an SLI setup. Yes, brace yourselves, I don't actually use these beasts for gaming.
If you read the reviews, it may look like these cards have no purpose other than to play the higher end games.
It is my responsibility to make a kind of "Google Earth on Steroids" for my employer. And this requires that five (yes, five) terabytes of mapping data be available for a multi-monitor (and by "multi" I mean many) display. What's my current choke point? Simply data bandwidth into the card.
Where does this review leave me? I now know intimately how high I can get my frame rate up in a first person shooter. Huzzah!
I know there are product sheets that tell me what kind of bandwidth I have but I'm more interested in what a non-interested third party has to say about it. Where are the real benchmarking tests? What about a simple program that loads up the card with as much data as possible as quickly as possible? I'm not even sure if the choking point is on the card or at the interface level with the motherboard (PCIe 16x).
Why can I not find objective reviews that aim to look at cold hard numbers?
My work here is dung.
The biggest news (for me at least) is that the MSRP of the 7900GT is $299. Considering the 7900GT performs on par with the 7800GTX, which is about $100 more, the 7900GT is starting to look like a bargain.
If any of you bleeding-edge gamers want to sell off your "old" 7800GTX for $250 or so, drop me a line
hmmm I wonder if these cards will be HDCP compatible?
63/64 FPS Max in Q4... Did they even bother to remove the vsync?
Your eyes are full of hate. That's good. Hate keeps a man alive. It gives him strength.
While I agree with the other respondant - that simplicity is inversely correlated with featureset - I was turned off by the Xbox 360 demo I saw at the store. It went something like this:
Click the game I wanted to see...
Wait...
Get the developer logos...
Wait...
Get the instructions...
Wait...
Select character...
Wait...
Watc^H^H^H^H Skip intro movie...
Wait...
After 45 seconds of waiting for the game to load, I forgot why I was even playing.
I mean, UT2004 didn't take that long to get me into a game on a 600MHz laptop.
I'd rather have someone respond than be modded up.
Yeah, encoding MP3s at 0.9x was awesome.
You have a choice when you buy a computer, and you could buy one that was cool, but you obviously didn't, so stop whining about it.
It doesn't mean much now, it's built for the future.
I'd love to read some general purpose GPU type benchmarks for these cards. I'm really curious how they perform compared to say the original Nvidia 6800 card. It might be fun to graph the performance and see what the curve looks like compared to CPU performance graphs.
The short answer is they could.
The long answer is that it's not a linear ratio of price to sales, and there are way more factors that go into pricing than you've probably ever thought of. First of all, you want to get the most money for your product, but eveyone has a different idea of what they're willing to pay. Some would pay $1000. Some would pay $500. Some would pay $25. Obviously the best thing would be to just ask them how much they're willing to pay and then charge them that, but in practice it's not that easy. So you can build a chart, find out how many people would pay X, and find the perfect price point, but there are still more factors. You're missing potential sales from people who would pay less, and potential profits from people who would pay more. So what's the solution? Different product lines. Charge the rich guy Z, the average guy Y, and the budget guy X.
Second, when you charge more, people feel that they're getting more.If you're selling shirts for $1, and the guy across the street is selling shirts for $250, the expensive ones must be better, right? They have to be, otherwise how could he charge $250? Because people will pay it. And if they feel they're getting something for their money, then they are.. at least, according to many economists. If they enjoy the goods received at the price they paid, then it was a fair price.
I'm too sleepy to keep writing, but as long as Nvidia is selling their chips and turning a profit, they must be doing ok.
https://www.eff.org/https-everywhere