Intel discrete GPU struggles: won't compete against NVIDIA or AMD

erek

[H]F Junkie
Joined
Dec 19, 2005
Messages
10,875
Fake news?

"A new graphics card with a 25W TDP isn't going to get anywhere near even the lowest-end Radeon or GeForce graphics card from competitors to Intel in AMD and NVIDIA, and if DG1 is this bad at this long away from release -- Koduri won't look good at all. He was the man behind the failed launch of Radeon RX Vega under AMD and Radeon Technologies Group at the time, with the Reddit post continuing: "The real question now is should we expect to see Koduri still be working for Intel after Q3'20? If not, you can expect to never see DG2 make its way to a retail product"."

Read more: https://www.tweaktown.com/news/6937...ruggles-compete-against-nvidia-amd/index.html
 
  • Like
Reactions: Mega6
like this
Fake news?

"A new graphics card with a 25W TDP isn't going to get anywhere near even the lowest-end Radeon or GeForce graphics card from competitors to Intel in AMD and NVIDIA, and if DG1 is this bad at this long away from release -- Koduri won't look good at all. He was the man behind the failed launch of Radeon RX Vega under AMD and Radeon Technologies Group at the time, with the Reddit post continuing: "The real question now is should we expect to see Koduri still be working for Intel after Q3'20? If not, you can expect to never see DG2 make its way to a retail product"."

Read more: https://www.tweaktown.com/news/6937...ruggles-compete-against-nvidia-amd/index.html

Lol 25w tdp.... my trx40 chipset pulls that.
 
Fake news?

"A new graphics card with a 25W TDP isn't going to get anywhere near even the lowest-end Radeon or GeForce graphics card from competitors to Intel in AMD and NVIDIA, and if DG1 is this bad at this long away from release -- Koduri won't look good at all. He was the man behind the failed launch of Radeon RX Vega under AMD and Radeon Technologies Group at the time, with the Reddit post continuing: "The real question now is should we expect to see Koduri still be working for Intel after Q3'20? If not, you can expect to never see DG2 make its way to a retail product"."

Read more: https://www.tweaktown.com/news/6937...ruggles-compete-against-nvidia-amd/index.html

Sounds like a load of shit. Intels first gen will likely at best reach midrange levels but nothing as pathetic as 25W TDP.
 
DG1 isn't supposed to compete with the higher end cards its more an APU esq product than anything.... That said I give it a 50/50 chance of ever coming to market in meaningful numbers
 
I'm not even sure which current Nvidia/AMD GPUs are in the 25w category for comparison...
 

Unlikely and actually, quite boring as well. I would not expect miracles from Intel, based solely on their CPU line, miracles are not forthcoming anytime soon. (Intel miracles, anyways.) Oh well, not going to buy Intel for my personal machines, nonetheless.
 
Unlikely and actually, quite boring as well. I would not expect miracles from Intel, based solely on their CPU line, miracles are not forthcoming anytime soon. (Intel miracles, anyways.) Oh well, not going to buy Intel for my personal machines, nonetheless.

You don't think the same people post here and on Reddit? I read/post on both.

Follow the source of this "rumor". It no better than if someone started a new account here and started making up BS.

I am NOT expecting miracles from Intels GPUs, just pointing out this rumor has ZERO credibility.
 
You don't think the same people post here and on Reddit? I read/post on both.

Follow the source of this "rumor". It no better than if someone started a new account here and started making up BS.

I am NOT expecting miracles from Intels GPUs, just pointing out this rumor has ZERO credibility.

Oh, I am not saying it is credible, although a 25 watt APU part might be around but, it is more likely the person who did the reddit may not even know of this sight.
 
Remember, this isn't the first time around for Intel GPUs. The truth is, they're a lot better than they were, so they're willing to give it a go again. It's just the first attempt at this "new" re-introduction. I'd cut them a lot of slack and see where this all leads.

As you know the iGPU(AGPU) of the current Intel CPU's is smoking the Vega's of the AMDs. While many might say, "who cares?"... it's interesting nonetheless.

Let's see where this goes. It's possible, that just like the last time Intel dabbled in this, that it ends up going nowhere.
 
Remember, this isn't the first time around for Intel GPUs. The truth is, they're a lot better than they were, so they're willing to give it a go again. It's just the first attempt at this "new" re-introduction. I'd cut them a lot of slack and see where this all leads.

As you know the iGPU(AGPU) of the current Intel CPU's is smoking the Vega's of the AMDs. While many might say, "who cares?"... it's interesting nonetheless.

Let's see where this goes. It's possible, that just like the last time Intel dabbled in this, that it ends up going nowhere.


cjcox https://www.vogons.org/viewtopic.php?f=5&t=64649

Intel i750 (no, not the i740) was the first back in 1988
 
Intel wont be releasing anything the enthusiast market will care about regardless of the rumors source. At best they may release a card that competes with 1650 5500 series.
 
Intel wont be releasing anything the enthusiast market will care about regardless of the rumors source. At best they may release a card that competes with 1650 5500 series.

That's actually a good first step for discrete, to be honest.

They need to prove that their technology is solid when it has enough grunt to run more than Indy/MOBA-level games at playable framerates. Drivers, stability, compatibility, all the stuff that one expects of AMD and Nvidia GPUs, before they can start convincing buyers on more powerful and more expensive parts.
 
Intel wont be releasing anything the enthusiast market will care about regardless of the rumors source. At best they may release a card that competes with 1650 5500 series.

I'll bet in the first year, of discrete cards the will get closer to a 1660 series.

But drivers would be my bigger concern.
 
Intel wont be releasing anything the enthusiast market will care about regardless of the rumors source. At best they may release a card that competes with 1650 5500 series.

But you know, even reaching those levels would be huge for Intel, right?
 
  • Like
Reactions: Auer
like this
True or not, I don't expect this card to make a legit splash in the gaming market either way.
 
I'm not even sure which current Nvidia/AMD GPUs are in the 25w category for comparison...

The GT1030 comes in 20 and 30W versions depending if you want DDR4 or GDDR5, as does its mobile equivalent the MX250. It's half the size of a GTX 1050.

The Raedon 520/530 are probably in the same ballpark but no specific power level is given.

That power level is too low to make a particularly useful card; assuming the rumor is correct about Intel targeting it, the product isn't intended to be much more than an engineering/manufacturing pathfinder. A commercially viable entry level part would need to be scaled up 2-3x.
 
But you know, even reaching those levels would be huge for Intel, right?

I dont think so, they have been making integrated graphics for years. They should have a clue to how to scale up to a discrete level graphics card. Problem is they lack the mentality for a gaming point of view, they would rather have the server side business and will never be a priority to them. $1,000 AI card is more valuable then the gaming business at $150 per card and we know Intel is having volume issues as it is. I may end up wrong but that would fit the typical business pattern of Intel.
 
As you know the iGPU(AGPU) of the current Intel CPU's is smoking the Vega's of the AMDs.

If I remember correctly, and I might not, the highest (G7) iGPU smokes the Vega 3, but is roughly on par with the higher-end one (Vega 8/11?). We'll see what happens when the Zen 2 APUs come out.
 
I dont think so, they have been making integrated graphics for years. They should have a clue to how to scale up to a discrete level graphics card. Problem is they lack the mentality for a gaming point of view, they would rather have the server side business and will never be a priority to them. $1,000 AI card is more valuable then the gaming business at $150 per card and we know Intel is having volume issues as it is. I may end up wrong but that would fit the typical business pattern of Intel.

They've been making non-performant integrated graphics for years.

The post I replied to suggested performance on par with PCIe bus-powered discrete offerings. I just suggested that even attaining that level would be an achievement for Intel.
 
They've been making non-performant integrated graphics for years.

To add, they've been at it for decades, really. They were despised up until the last decade or so when they started taking more than 2D desktop acceleration seriously. They're not a complete unknown, and the lingering questions are largely academic: it's not trying to figure out what does work, because almost everything works and works well, but discerning what hasn't been tested due to the limited nature of IGPs / APUs.

The post I replied to suggested performance on par with PCIe bus-powered discrete offerings. I just suggested that even attaining that level would be an achievement for Intel.

Agreed, though on the one hand, had Intel not romeo-foxed their foundry upgrade progression, they could max out their reticules and build as big as they want -- on the other hand, that's a poor way to gain market share even if they produce an efficient, well-priced, class- or market-leading product as there just aren't enough buyers at that level.

Targeting < 75w (PCIe slot limit) makes quite a bit of sense because they can easily push out their install base, build some marketshare and mindshare, and start getting some feedback on what's working and what isn't when people use their technology in ways that weren't previously practical.
 
Problem is they lack the mentality for a gaming point of view, they would rather have the server side business and will never be a priority to them. $1,000 AI card is more valuable then the gaming business at $150 per card and we know Intel is having volume issues as it is. I may end up wrong but that would fit the typical business pattern of Intel.
Agreed. They hired Raja for a reason. It is obvious that his GPU expertise is more oriented towards compute.
 
...and start getting some feedback on what's working and what isn't when people use their technology in ways that weren't previously practical.
It's probably because my brain's fuzzy from waking up, but the first thing that came to mind reading this is, "A new host of attack vectors leading to a market for GPU cooking shows?".
 
I would say the probability of Intel launching a 25w discrete GPU in the current market as a finished product is basically zero. My guess is their primary target will initially be the data centre space before branching out into gaming, and in neither context will a 25w GPU be worth anyone’s time. I’m extremely skeptical of the veracity of this rumour.
 
I dont think so, they have been making integrated graphics for years. They should have a clue to how to scale up to a discrete level graphics card. Problem is they lack the mentality for a gaming point of view, they would rather have the server side business and will never be a priority to them. $1,000 AI card is more valuable then the gaming business at $150 per card and we know Intel is having volume issues as it is. I may end up wrong but that would fit the typical business pattern of Intel.

My feelings as well.
 
Back
Top