GeForce GTX 1080 Ti Announced: 3584 CUDA Cores, 11 GB vRAM, 11 Gbps

[Michael_S]

Way too god damn expensive. Insanity.

It's for the people that want to run the cutting edge of gaming, or CUDA / OpenCL or similar. And to be fair, it's way more cost and resource efficient for someone to buy this than, say, getting a mid-life crisis Mustang or BMW.

I just got a $200 AMD RX-480 GPU a few weeks ago and you wouldn't believe how good Nethack looks now.

[schmidtbag]

nvidia cards seem to be consistently more energy efficient, compared to similar amd cards.

Consistently? Not at all - have you forgotten the 400 series, or any GPUs re-branded from that?
The correct phrasing is "Nvidia cards are usually more energy efficient compared to similar AMD cards"

[bug77]

Consistently? Not at all - have you forgotten the 400 series, or any GPUs re-branded from that?
The correct phrasing is "Nvidia cards are usually more energy efficient compared to similar AMD cards"

It's pretty consistent if you can only name one exception off the top of your head

[theriddick]

AMD was only able to tell us the name of their new card, the RX VEGA. Seems they are shy to show off its power just yet, perhaps they will wait, we might see a 1080ti price drop by the time VEGA comes out. I do feel sorry for AMD, they spent a fortune on their new tech and VEGA and will have no choice but to sell it at budget discount prices by the time its released because competition is just that far ahead of them atm (GPU wise anyway).

[schmidtbag]

It's pretty consistent if you can only name one exception off the top of your head

Again, consistent isn't the right word. Consistency means no broken patterns over time. Nvidia isn't consistently more efficient, but they are most of the time. Anyway, the 400 series isn't the only exception, it's just a relatively recent one.

[drohm]

It's for the people that want to run the cutting edge of gaming, or CUDA / OpenCL or similar. And to be fair, it's way more cost and resource efficient for someone to buy this than, say, getting a mid-life crisis Mustang or BMW.

I just got a $200 AMD RX-480 GPU a few weeks ago and you wouldn't believe how good Nethack looks now.

Trust me, I get all that, I'm for the most part one of those people. $700 is ridiculous though, no question they are gouging. Most mobo/cpu combos aren't that expensive, think about it. Vega can't come soon enough. I plan on building a new rig this summer, all AMD, can't wait.

[schmidtbag]

Trust me, I get all that, I'm for the most part one of those people. $700 is ridiculous though, no question they are gouging. Most mobo/cpu combos aren't that expensive, think about it. Vega can't come soon enough. I plan on building a new rig this summer, all AMD, can't wait.

Haha, my entire gaming PC will cost roughly $700. I don't see a reason to spend more. To me, 2K isn't enough of an upgrade from the 1080p I use now. 4K displays are too expensive. VR is still too "beta" to consider upgrading my OSVR. Linux overall isn't very optimized for 4K displays. My internet connection isn't fast enough to stream 4K videos, and not even the GTX 1080 can reliably play modern games at 4K.

I figure by the time games don't play smoothly on my GPU anymore, there will be a reasonably priced GPU out there that can play games in 4K.

[L_A_G]

It's for the people that want to run the cutting edge of gaming, or CUDA / OpenCL or similar. And to be fair, it's way more cost and resource efficient for someone to buy this than, say, getting a mid-life crisis Mustang or BMW.

No, it's the Titan cards that are the cards for people who do some heavy compute on the GPU. Nvidia has a long history of intentionally crippling FP64 performance on their consumer cards so that people doing GPU compute, who actually do FP64 computations on their GPUs, won't be able to buy essentially the same cards for less. People who buy GPUs to play games don't really care as games generally use FP32 and in some cases FP16 when they don't need the resolution.

I should know as my master's thesis project was writing an OpenCL compute application for a researcher who had an original GTX Titan in his machine for compute jobs.

[schmidtbag]

No, it's the Titan cards that are the cards for people who do some heavy compute on the GPU.
...
I should know as my master's thesis project was writing an OpenCL compute application for a researcher who had an original GTX Titan in his machine for compute jobs.

Actually, it's the Quadro and Tesla cards that are meant for heavy computing. Titans are capable of it, but they're marketed as gaming cards. When writing your own software, a Titan may be a better choice, but if you're using any real-world professional applications, it's likely a poor choice. Nvidia doesn't have the same driver optimizations for Titans as they do for Quadros.

The 1080Ti definitely is not intended for OpenCL/CUDA (though again, they're capable of it).

[Michael_S]

Haha, my entire gaming PC will cost roughly $700. I don't see a reason to spend more. To me, 2K isn't enough of an upgrade from the 1080p I use now. 4K displays are too expensive. VR is still too "beta" to consider upgrading my OSVR. Linux overall isn't very optimized for 4K displays. My internet connection isn't fast enough to stream 4K videos, and not even the GTX 1080 can reliably play modern games at 4K.

I figure by the time games don't play smoothly on my GPU anymore, there will be a reasonably priced GPU out there that can play games in 4K.

If you're going to spend $700 on a video card, $500 for a 4k display won't slow you down. If you're spending $700 on your gaming PC, it will. (And there's no shame in that. Aside from my RX-480, the rest of my PC is six years old.)