The GeForce RTX 2080 Ti Arrives For Linux Benchmarking

[phoronix]

Phoronix: The GeForce RTX 2080 Ti Arrives For Linux Benchmarking

It looks like NVIDIA has their launch-day Linux support in order for the GeForce RTX 2080 "Turing" graphics cards slated to ship later this week as arriving today at Phoronix was the RTX 2080 Ti...

http://www.phoronix.com/scan.php?page=news_item&px=GeForce-RTX-2080-Ti-Linux

[Brisse]

Apparently they are designed similiarly to something Apple would do. In short: They are not made to be taken apart.

https://youtu.be/w9FtXZGQzfM?t=6m8s

[pegasus]

If you ignore miners, the largest audience for these cards is the AI crowd. They can afford to pay for them much more than gamers, but it's still cheaper and more cost-effective for them to go for gaming cards instead of server class Teslas.

So if you can do some inference benchmarks that would be great. Tesla T4 is supposedly very good at it at 75W only and if this 2080ti shows same numbers, gamers will whine again about availability

[Michael]

If you ignore miners, the largest audience for these cards is the AI crowd. They can afford to pay for them much more than gamers, but it's still cheaper and more cost-effective for them to go for gaming cards instead of server class Teslas.

So if you can do some inference benchmarks that would be great. Tesla T4 is supposedly very good at it at 75W only and if this 2080ti shows same numbers, gamers will whine again about availability

Have any particular inference benchmarks in mind? (That are nice to build and script well.) I have had a few test profiles but they have tended to be messy with a load of dependencies, etc, so continuously looking for new tests in that area.

[tildearrow]

Michael Does this mean you are canceling your 2080 pre-order?

[miabrahams]

Very interested in the rendering benchmarks (blender, octanebench)

[miabrahams]

If you ignore miners, the largest audience for these cards is the AI crowd. They can afford to pay for them much more than gamers, but it's still cheaper and more cost-effective for them to go for gaming cards instead of server class Teslas.

So if you can do some inference benchmarks that would be great. Tesla T4 is supposedly very good at it at 75W only and if this 2080ti shows same numbers, gamers will whine again about availability

This vastly underestimates the size of the 100 billion dollar gaming market. Who knows whether things like RTX cores are even useful for AI at all? We could likely see a split with server/workstation cards dropping RTX to maximize tensor processing performance while PC cards accrue more and more gaming-specialized features.

[Xaero_Vincent]

$1200 was the entire cost of my gaming desktop I built years ago. Miners have sure driven the cost of GPUs up.

[pegasus]

Have any particular inference benchmarks in mind? (That are nice to build and script well.)

Unfortunately the only "nice to build" things in this department are provided as docker images. Building these frameworks manually is an exercise in patience. EasyBuild helps with some but not yet all of them.

[dimko]

$1200 was the entire cost of my gaming desktop I built years ago. Miners have sure driven the cost of GPUs up.

I can still compile a functioning computer for half price or so. Good one including screen with 1200$