User Profile

[coder]

I understand there are some regional variations, but don't the same conversions roughly apply to Nvidia products, wherever you are?...

[coder]

Beyond ldesnogu 's point, I wish you could add the nearest laptop & desktop CPUs to these graphs. They would provide a useful basis for comparison....

[coder]

Yeah. If Nvidia is going to the trouble and expense of sending you another Jetson board, Michael, maybe they will tell you how to configure it for max...

[coder]

Current Newegg pricing (in stock, sold by Newegg, and not open box or refurb):

Nvidia ...

[coder]

Except the 2023 card will be a RTX 5200.

I'm still bitter about the time I spec'd out an Ivy Bridge server and what got ordered was Sandy...

[coder]

I guess the title should read "C++11 Library"? Perhaps your honorary editor is too busy in another thread......

[coder]

Really? So, you're also building deep learning into commercial products?


No, it doesn't. A tensor core is so-called because it computes...

[coder]

Yes, that's the point. But it must be implemented in a very limited, power-efficient way that's customized to the specific problem of branch prediction....

[coder]

Agreed. I'm just pointing out an obstacle they'll face, when they even reach that point.


Oh, right. But it's extremely hard-wired...

[coder]

Speaking of AI, it would be interesting to see some benchmarks on this:

http://www.up-board.org/ai-edge/ai-core/

[coder]

That looks decent.

According to https://www.anandtech.com/show/12501...y-and-video-ip

it looks like a MP4 implementation should...

[coder]

See https://www.phoronix.com/forums/foru...26#post1041626


Yes, it's a given that we're talking about workloads which vectorize well....

[coder]

And certainly patented.


Wut? They did?



And they never enabled it. Assuming you're referring to primitive...

[coder]

How do you know?

The cores are virtually the same as Volta, which has been out for over a year....

[coder]

They're probably just building up a stockpile of higher-binned TU102's and maybe waiting for some market feedback to see how to price it, but it's sort...

[coder]

Not with crytpo crashing, a new wave of ASIC miners on the market, and these cards being so expensive....

[coder]

1.35x of Vega 64 is what they've said. Also, 2x performance per Watt. Calibrate your expectations to that....

[coder]

Is this rounded to the nearest $0.01? It seems odd that these 3 cards would otherwise have identical results....

[coder]

We were just listing their "firsts".

Anyway, it does have worthwhile benefits - just not at consumer price points and workload...

[coder]

Don't forget HBM. Fury had it long before Nvidia launched the P100. And Nvidia has never offered it in a consumer card, while AMD has given us two generations...