Glad to see the Intel shills went to panic stations over this.

So many people convinced it's amazing for having a 35W TDP - but where did they get that number? All we know is it's somewhere south of 75W.

It's not an iGPU, so comparing it to the 1050 absolutely makes sense, as that's roughly the level of performance you'd expect from an entry-level GPU 3 years later. If it can't hit that, it's not particularly interesting.

Quote from: william blake on May 31, 2020, 21:05:55

QuoteIf that is the case, entry level discrete chips would continue to be the better option next year.

oh christ here we go again.. do you understand they can make entry discrete exactly sligtly faster than the fastest integrated? every time. there is no competition between these two type of products. zero. one single fact is enough-the same piece of silicon is faster if made as separate chip with its own memory. and this separate chip is much more profitable for the manufacturer.

I guess it was true in world withhout competition but now low tier AMD laptops make NVIDIA MX 250 look stupid. See dave 2d latest review on $600 laptop against Razer Stealth for example.

Its amazing how the Intel fangirls just get their knickers in a twist over this article. Its not new, every other article out there already paints a very disappointing picture of its performance, so why blame this particular messenger?

The main problem is that you fangirls don't seem to get the point. Its about power-performance ratio!

@william blake - if you can't punctuate, nobody bothers to read. But based on what little I gather caught in my peripheral vision, it is not true that an integrated GPU made into a dGPU will perform better. That is nonsense. The far greater distance to the CPU means that for the same performance, it will consume far more power than the iGPU. Which is why the AMD Zen 2 Renoir APUs have blown away EVERYTHING else in terms of performance per watt for both CPU and GPU. That is still the best performance metric. Any idiot can overclock some lousy GPU to pervert performance numbers.

@akala - Similar problem. Of course the comparison of the Intel Xe DG1 with the 1050 is valid! It is a laptop dGPU too isn't it? Therefore, it is roughly within the same ballpark TDP isn't it? That's also why it is being compared with the other cited candidates!

Low end or high end is irrelevant and NOT THE POINT! We all know the Intel Xe DG1 is an entry level, low cost and low power product. It's power consumption/TDP is anything up to 75W according to its spec sheet. Of course any comparison has to be against any & all Nvidia or AMD competitors within that range! That means anything from AMD APU Vega iGPUs 14nm to 7nm, up to any other laptop dGPU by Nvidia/AMD, including the 65W AMD cited, whether they are 3 years old or 10 years old!

What's with you Intel fangirls? Duh...!

QuoteIf that is the case, entry level discrete chips would continue to be the better option next year.

oh christ here we go again.. do you understand they can make entry discrete exactly sligtly faster than the fastest integrated? every time. there is no competition between these two type of products. zero. one single fact is enough-the same piece of silicon is faster if made as separate chip with its own memory. and this separate chip is much more profitable for the manufacturer.

To be a discrete graphic card is very good, and with a 35W TDP awesome! I would like to see it inside next intel APU with 65W TDP to compete against Vega8 of AMD that with 7nm has stolen lot of space for other things...

But this is expected, isn't it?
It's Intel GPU, the last one was barel able to get close to MX150 ... so this is a huge step.
I mean nobody expected high end ... or am I wrong?

1050 is absolute trash level.
2020 iGPU could have that performance if not shared DDR4 memory.

Nvidia themselves can barely match the gtx 1050 in a less than < 20w package.

When AMD takes 5 years to match, 860m/960m performance it's seen as amazing and a major breakthrough.

But intel matching a 3 year old gtx 1050? And actually being released on time in actual products this September unlike some other companies? Super duper worrying indeed.

IMO one should expect a 20-40% jump in performance from an engineering sample of a brand new product to stable retail drivers.

What is the point of this article? Why you compare it with GTX 1050? Was there any clue that this card is a high-end? There is no information about the Intel card but you compare it with others. Even worse, you say:

"Intel's Xe architecture barely keeps up with 3 year-old Pascal"

where did you get that information to compare architectures?

pointless article..

Why would it overtake 1050? Where exactly did Intel state that this card is a contender for 1050? As far as I know this card is based on the same solution as used in Tigerlake U 28W so it is a very very low end GPU (read GT1030 or MX line from nvidia, which this card handily beats).
So please, use your stupid brain before just copy pasting the idea of another article. If this card consumes 20W for this performance than this architecture is MUCH better than 1050 and the 128EU variant will fight with GTX1650.