Intel Xe DG1 SDV PCB Pictured, Looks Desolate

[bug]

Good point. Are we expecting 3rd-party designs? What is Intel's AIB strategy? I'd expect the entire thing to be a flash in a pan and to work miserably as a companion card for Gen 9, 10 and 11 CPUs, to boost iGPU performance. I can't see a single scenario where this would justify taking it to 3rd parties or, in fact, buying a retail version of this Intel card at all...

What is Intel offering here for casual? A sub-par Fortnite/Rocket League/LOL experience? Or are there other gains to be had from the use of this add-in card?

I think Intel is currently going after OEM and server/compute. The desktop market is much more saturated and probably unattractive at this point, considering Intel's lack of production capacity.
Personally, I think they'll release something for the desktop as well, once they fix their capacity problems.

 

[Aquinus]

I don't see why hatin on Intel. I really hope they release something competitive in near future as we can only benefit from one more player in the discrete GPU space.

The hate is from how literally none of the benchmarks worked from the source linked in the article, so we have absolutely no idea what performance actually is because it's so buggy. That definitely deserves some scrutiny. I expect my GPUs to work.

 

[bug]

The hate is from how literally none of the benchmarks worked from the source linked in the article, so we have absolutely no idea what performance actually is because it's so buggy. That definitely deserves some scrutiny. I expect my GPUs to work.

You're just nitpicking

Edit: There's something strange in the original article. They keep saying that after a year or so they were expecting things to work. While I didn't find a clarification for this, it leads me to believe we are looking at an older engineering sample retested with more up-to-date software. But I'm really not sure.

 

[r9]

The hate is from how literally none of the benchmarks worked from the source linked in the article, so we have absolutely no idea what performance actually is because it's so buggy. That definitely deserves some scrutiny. I expect my GPUs to work.

You expected wrong.

 

[Aquinus]

You're just nitpicking

Edit: There's something strange in the original article. The keeps saying that after a year or so they were expecting things to work. While I didn't find a clarification for this, it leads me to believe we are looking at an older engineering sample retested with more up-to-date software. But I'm really not sure.

In their defense, Intel's open source drivers in Linux have become a lot better in the last couple of years, like, a lot better. Their GPUs are still relatively slow, but they've been making good strides preparing for these GPUs, so it is a little weird that they're so blatantly broken.

 

[Mussels]

Whats the weird connector on the back of the card?

 

[DonKnotts]

Whats the weird connector on the back of the card?

There is a row of LEDs on the shroud between the fins, I assume that's the power cable for them.

 

[bug]

In their defense, Intel's open source drivers in Linux have become a lot better in the last couple of years, like, a lot better. Their GPUs are still relatively slow, but they've been making good strides preparing for these GPUs, so it is a little weird that they're so blatantly broken.

It may not be the drivers at fault here, but the possibly pre-production board.

 

[Steevo]

So.... What does it do?

C'mon.....do something.

 

[Stephen.]

Whats the weird connector on the back of the card?

It's for debugging.