Intel Arc A380

[Mussels]

It'd be a total failure from AMD or Nvidia, but for the first card from a new competitor its decent (with the beta drivers, at least)


It's not what you'd expect from intel who's had decades of GPU driver management for their IGP's, however

I don't understand why every review outfit, including the ones I trust to be objective like GN and TPU, is giving this GPU a pass. Because it's rubbish. Objectively rubbish. Intel/GUNNIR haven't even figured out fan hysteresis FFS!

Saying "RT performance is good compared to AMD" is pointless when we're talking about double or even single-digit RT framerates. And when the generation-old RTX 2060 is twice as fast.

Don't buy this rubbish. Don't pay Intel to be beta testers for them.

It's not priced or marketed to be a competitor with a 3090Ti here, if performance is within the correct segment its doing well - the RT performance is unplayable here, but the higher performance models wont be






Oh and good to see a lot of detail given on the state of the drivers and software, because holy hell that'd be a shock to someone expecting 'intel quality'

 

[Forza.Milan]

I really hope intel succeeds with their gpu planning, we need another source of gpu to tackle mining activities in the future..

 

[R0H1T]

And some rumors already suggesting that Intel may quit their dGPU gig early, brutal hit to their numbers everywhere!

Intel Reports Second-Quarter 2022 Financial Results. First Net Loss in Decades

Intel Corporation today reported second-quarter 2022 financial results. "This quarter's results were below the standards we have set for the company and our shareholders. We must and will do better. The sudden and rapid decline in economic activity was the largest driver, but the shortfall also...

www.techpowerup.com

 

[AusWolf]

And some rumors already suggesting that Intel may quit their dGPU gig early, brutal hit to their numbers everywhere!

Intel Reports Second-Quarter 2022 Financial Results. First Net Loss in Decades

Intel Corporation today reported second-quarter 2022 financial results. "This quarter's results were below the standards we have set for the company and our shareholders. We must and will do better. The sudden and rapid decline in economic activity was the largest driver, but the shortfall also...

www.techpowerup.com

I know my corporate blah-blah language skills are limited, but I don't see any mention of Intel's future dGPU plans in this article.

 

[R0H1T]

It's not from here, the rumor about exiting dGPU space, but read it on reddit & AT. With Intel you never know, outside of CPU's they've never really excelled at anything else!

And things will get much worse before they get better for them, basically if they're looking to cut even more costs (gaming?) dGPU's could be on the chopping block!

 

[AusWolf]

It's not from here, the rumor about exiting dGPU space, but read it on reddit & AT.

I don't trust reddit. If I read about it here, I'll believe it.

With Intel you never know, outside of CPU's they've never really excelled at anything else!

I don't know. Their network adapters are pretty decent.

And things will get much worse before they get better for them, basically if they're looking to cut even more costs (gaming?) dGPU's could be on the chopping block!

I think it's too soon to talk about a business segment that they haven't really started yet. First, they have to release some products to have their investments returned. If Arc ends up being a huge flop, then they might think about quitting the dGPU field. But not before - it would be foolish.

 

[R0H1T]

I don't trust reddit. If I read about it here, I'll believe it.

Hence the part about it being a rumor, I trust AT as much as TPU if not more.

Their network adapters are pretty decent.

Probably because they're much simpler?

First, they have to release some products to have their investments returned.

They have the enterprise market for that & much better margins there. The consumer market can be brutal if you've not planned it properly & don't intend to be in it for the long run! Intel probably spent 10x as much trying to fit in the mobile market & eventually left with literally billions in losses per quarter & that was back when they could afford it.

 

[AusWolf]

They have the enterprise market for that & much better margins there. The consumer market can be brutal if you've not planned it properly & don't intend to be in it for the long run! Intel probably spend 10x as much trying to fit in the mobile market & eventually left with literally billions in losses per quarter & that was back when they could afford it.

Yes, but they've tried. A-series Arc is basically done, and B-series is already in development. Making it all happen is not cheap, and leaving it all to rot without selling anything would be stupid. They might stop after B-series if the concept doesn't gain foothold, but not before. If I had an unsuccessful investment, I wouldn't shut it down straight away. I'd try to be on the market as long as I can to earn back at least some of my losses.

 

[R0H1T]

Of course they wouldn't suddenly stop tomorrow or even next quarter, for that to happen things would have to get much much worse. But Intel's in a much worse shape than many others realize ~

Intel Arc Alchemist refresh reportedly in the works as next-gen Arc Battlemage supposedly gets delayed

Arc Alchemist is Intel’s first foray into dedicated gaming hardware meant to compete directly with Nvidia and AMD. Rumor has it that Intel might be planning a refresh of the Arc Alchemist GPU lineup for notebooks. The refresh, if true, will be surprising since notebooks with Arc Alchemist dGPUs...

www.notebookcheck.net

From AT ~

Q: Reserves for Sapphire and timelines - can give more specific on SPR, not just early units but volume? How does this impact/push out EMR and GNR?

A: We're already ramping a number of SKUs of SPR starting last Q. The particular issue, wasn't affecting those SKUs, so we did another tapeout for the volume SKUs, ramping in 2H. Feel comfortable. EMR goes into the SPR platform, product is healthy for 2023.

Question - Intel Q2 Results - Terrible

https://d1io3yog0oux5.cloudfront.net/_0f37de80c52d83ac6338b4805dea84aa/intel/db/887/8856/earnings_release/Q2+22_EarningsRelease+%281%29.pdf Revenue down 22%. Edit: Oh and it gets better... Intel posted a loss!

forums.anandtech.com

Their latest & greatest, in servers, is delayed again!

Just to give you a hint look at their margins ~

Now imagine AMD just around the Dozer era, except 10-15x as big!

 

[AusWolf]

Of course they wouldn't suddenly stop tomorrow or even next quarter, for that to happen things would have to get much much worse. But Intel's in a much worse shape than many others realize ~

Intel Arc Alchemist refresh reportedly in the works as next-gen Arc Battlemage supposedly gets delayed

Arc Alchemist is Intel’s first foray into dedicated gaming hardware meant to compete directly with Nvidia and AMD. Rumor has it that Intel might be planning a refresh of the Arc Alchemist GPU lineup for notebooks. The refresh, if true, will be surprising since notebooks with Arc Alchemist dGPUs...

www.notebookcheck.net

From AT ~

Question - Intel Q2 Results - Terrible

https://d1io3yog0oux5.cloudfront.net/_0f37de80c52d83ac6338b4805dea84aa/intel/db/887/8856/earnings_release/Q2+22_EarningsRelease+%281%29.pdf Revenue down 22%. Edit: Oh and it gets better... Intel posted a loss!

forums.anandtech.com

Their latest & greatest, in servers, is delayed again!

Sure, but this is still not an indication of their future dGPU plans.

 

[R0H1T]

Like I said it's rumor so take it for what it's worth, or not.

 

[Mussels]

It's not from here, the rumor about exiting dGPU space, but read it on reddit & AT. With Intel you never know, outside of CPU's they've never really excelled at anything else!

And things will get much worse before they get better for them, basically if they're looking to cut even more costs (gaming?) dGPU's could be on the chopping block!

I mean, until this year almost every single computer company used intel hardware - my AMD motherboards have intel ethernet

SATA ports, Ethernet chipsets, USB chipsets, wireless chips
Intel networking hardware is overall considered the best there is - high performance, no driver issues (and on the wifi side, incredibly cheap)

Intel make a looooot more than just CPUs

 

[Valantar]

I mean, until this year almost every single computer company used intel hardware - my AMD motherboards have intel ethernet

SATA ports, Ethernet chipsets, USB chipsets, wireless chips
Intel networking hardware is overall considered the best there is - high performance, no driver issues (and on the wifi side, incredibly cheap)

Intel make a looooot more than just CPUs

Not to mention FPGAs, datacenter/server networking and fabric, server RAID controllers/HBAs, 5G-related ASICs, SSDs, non-volatile memory solutions ... they do a ton of stuff. With varying success? Sure. But a ton of stuff still.

 

[R0H1T]

Intel make a looooot more than just CPUs

And they fail at most of them, remember (5G) modems? Most of what you quoted is part of the (mother) board or does Intel sell SATA ports now? And as I pointed out earlier they're not that complex.

FPGAs

They acquired Altera, just like AMD did with Xilinx. Outside their core area of x86 CPU, besides "networking" they have very little to show!

SSDs

Sold that a while back!

Intel Sells SSD Business and Fab to Sk Hynix, New 'Solidigm' Subsidiary Launched

Intel and SK hynix seal the first phase of the deal.

www.tomshardware.com

 

[Valantar]

They acquired Altera, just like AMD did with Xilinx. Outside their core area of x86 CPU, besides "networking" they have very little to show!

...and? That makes Altera a part of Intel, and what-used-to-be-Altera's business Intel's business. Also, saying "besides "networking"" seems to indicate a lack of understanding of just how large, profitable and fundamental to the industry that sector is. Fabric and networking is the backbone of modern server, HPC and cloud computing, and will only be growing more important - and Intel is among the longest-standing, highest performing, most respected vendors in that area. There's a reason Nvidia spent billions acquiring Mellanox - fabric is crucial.

Sold that a while back!

The main NAND unit, sure, but they still make Optane SSDs for enterprise users. As I said, not all of these ventures are very successful, but many of them are.

 

[R0H1T]

The main NAND unit, sure, but they still make Optane SSDs for enterprise users.

Did you miss the front page news?

Also, saying "besides "networking"" seems to indicate a lack of understanding of just how large, profitable and fundamental to the industry that sector is.

Your post would've made some sense had they excelled in 5G, or even 4G, but that never came to pass! In a world where QC, Broadcom, Marvell, MediaTek exist Intel doing solid networking is not earth shattering news. They've been at it for what 30~40 years now IIRC?

...and? That makes Altera a part of Intel, and what-used-to-be-Altera's business Intel's business.

And that makes what Altera developed eons back somehow Intel's innovation?

There's a reason Nvidia spent billions acquiring Mellanox - fabric is crucial.

It is & how does that make my statement less relevant? Should I post the list of their failures? I'm sure I can get more than what you posted back then.

There's a reason there's just 2 major x86 chipmakers in the world or 2 (major) dGPU makers & no it's nothing to do with just their patents! It's effin hard that's why.

 

[Valantar]

Did you miss the front page news?

Given that it was posted a couple of hours ago, yes.

Your post would've made some sense had they excelled in 5G, or even 4G, but that never came to pass! In a world where QC, Broadcom, Marvell, MediaTek exist Intel doing solid networking is not earth shattering news. They've been at it for what 30~40 years now IIRC?

... and does "having been at it for 30~40 years" somehow mean that they're not successful or a major player in this segment? I don't follow your logic here, sorry. I never claimed that it was news, neither earth shattering nor otherwise. I'm simply pointing out that your view of what Intel does is myopic.

And that makes what Altera developed eons back somehow Intel's innovation?

Have I claimed that it is? Again: Altera's current business is Intel's business, and it is relatively successful. As such, it is a field in which Intel is doing well. Is that success largely due to the work of people at what was at the time another company? Sure. But so what? Companies merge and split constantly, and Intel bought Altera in 2015, meaning that FPGAs developed fully under Intel have been on the market for a bit now.

It is & how does that make my statement less relevant? Should I post the list of their failures? I'm sure I can get more than what you posted back then.

Well, given that your statement was

outside of CPU's they've never really excelled at anything else!

then yes, it does indeed make it less relevant, as it is a major, crucial field in which Intel has been a leading actor for decades by your own admission. That would constitute "excelling" at that, no?

There's a reason there's just 2 major x86 chipmakers in the world or 2 (major) dGPU makers & no it's nothing to do with just their patents! It's effin hard that's why.

It's hard, expensive, and even more expensive if you don't have the patents. And, crucially, literally not legally possible without access to licencing those patents and core technologies. VIA could have done it, but never had the funding or market presence to get that funding (though their Chinese branches/subsidiaries/spinoffs seem to be doing quite a lot better now, with the cash injections that has brought with it).

 

[R0H1T]

That would constitute "excelling" at that, no?

No it would have to be at or near class leading, which admittedly due to lack of competition in the GPU space ~ they are, although Apple's beating them hard with the Mx gen SoCs.

It's hard, expensive, and even more expensive if you don't have the patents. And, crucially, literally not legally possible without access to licencing those patents and core technologies.

That doesn't apply to Intel in the dGPU space, they had the patent sharing (licensing?) agreement with Nvidia & then AMD. If they couldn't do it a major reason was that they were simply not good enough!

Strictly speaking outside of probably ethernet Intel's not been leading anywhere in the networking space AFAIK & even when it comes to ethernet they're not winning any benchmarks, so I'm not sure how you'd show they're excelling there? I'll give Intel bonus points for reliability though.

Qualcomm sues Apple for making its modems look inferior to Intel’s

The chipset company countersues after it was hit with numerous lawsuits for overcharging manufacturers. Qualcomm filed a lawsuit against Apple in the US....

www.gsmarena.com

 

[Valantar]

No it would have to be at or near class leading, which admittedly due to lack of competition in the GPU space ~ they are, although Apple's beating them hard with the Mx gen SoCs.

... so Intel is "at or near class leading" in the GPU space, but by some measure not in the networking space? Sorry, but what planet are you on? Intel is far closer to the front of the pack in networking than they are in GPUs. It's also a more diverse field and thus more difficult to judge, but Intel is way, way up there in terms of networking both for consumers, enterprise, and high end datacenter stuff. As for GPUs, they are ... doing okay for a first effort, I guess?

That doesn't apply to Intel in the dGPU space, they had the patent sharing (licensing?) agreement with Nvidia & then AMD. If they couldn't do it a major reason was that they were simply not good enough!

Wait, I was responding to your argument about X86 CPUs here, this wasn't about GPUs, was it?

Strictly speaking outside of probably ethernet Intel's not been leading anywhere in the networking space AFAIK & even when it comes to ethernet they're not winning any benchmarks, so I'm not sure how you'd show they're excelling there? I'll give Intel bonus points for reliability though.

Qualcomm sues Apple for making its modems look inferior to Intel’s

The chipset company countersues after it was hit with numerous lawsuits for overcharging manufacturers. Qualcomm filed a lawsuit against Apple in the US....

www.gsmarena.com

From the way you're speaking of this it seems your limited understanding of what goes into "networking" is rather limited, as you're speaking either of client controllers (ethernet/Wifi) or mobile data (which isn't generally considered networking in that sense, though of course there's an overlap in functionality). You're entirely glossing over the enterprise networking market, datacenters, HPC, server fabric solutions, +++. These are huge, high performance, high cost and high margin areas. Intel is by no means an out-and-out market leader in this space, but they've delivered among the largest, best supported and best performing solutions in many areas of these complex markets for years and years. Claiming they "haven't really excelled" in this space is pure nonsense.

 

[R0H1T]

so Intel is "at or near class leading" in the GPU space, but by some measure not in the networking space?

I mentioned at least 4 other competitors in the "networking" space, besides ethernet ~ which probably you can't prove they're leaders in based on benchmarks otherwise you would've done it! Where have they excelled? They're kinda "leading" in the GPU space due to lack of competition, much like x86 CPU's.

Wait, I was responding to your argument about X86 CPUs here, this wasn't about GPUs, was it?

What? You need to get your story straight, I already covered this here ~

2 (major) dGPU makers & no it's nothing to do with just their patents! It's effin hard that's why.

I clearly mentioned both CPU & GPU, the reason Intel couldn't do it because it's difficult (to compete) & complex!

Wanna go back to your favorite point about networking now?

You're entirely glossing over the enterprise networking market, datacenters, HPC, server fabric solutions, +++. These are huge, high performance, high cost and high margin areas. Intel is by no means an out-and-out market leader in this space, but they've delivered among the largest, best supported and best performing solutions in many areas of these complex markets for years and years. Claiming they "haven't really excelled" in this space is pure nonsense.

Let me give you a hint ~ it's because of that virtual server (& until recently x86) monopoly!

Apple Removes Remaining Intel Components from M2 MacBooks

Apple has removed the final remaining Intel components from its latest M2 MacBooks with the Intel JHL8040R USB4 timer chips being replaced with a pair of custom U09PY3 chips. This change was discovered by iFixIt during a recent teardown and documented by Twitter user SkyJuice with the exact...

www.techpowerup.com

You seem to excel in pedantry, how about you back your arguments with more than just "the world knows"? I gave you real world example of how Intel's failed at 4G modem, they utterly failed in even getting a 5G one to work!

 

[noel_fs]

The card isn't garbage, but it is terrible value.

I think every article/review about low-end cards needs a disclaimer that these cards have many valid uses beyond gaming, especially those wanting to add new capabilities to older machines which still have are working fine.

If Intel priced this card below $100 and didn't market it as a gaming card, the reception would be fine.

i absolutely agree

but since thats not the case, we gotta call it for what it is.

 

[efikkan]

i absolutely agree

but since thats not the case, we gotta call it for what it is.

Precisely how is a product with the wrong price "garbage"?
I believe any reasonable mind would distinguish between an overpriced product and a garbage product.
And prices can change, you know

 

[Valantar]

I mentioned at least 4 other competitors in the "networking" space, besides ethernet ~ which probably you can't prove they're leaders in based on benchmarks otherwise you would've done it! Where have they excelled? They're kinda "leading" in the GPU space due to lack of competition, much like x86 CPU's.

But ... does the existence of competitors somehow constitute Intel not excelling in networking? I don't follow your logic here, sorry. To excel at something doesn't mean you have to be the best at all times, nor that you have to be the indisputed leader at something.

What? You need to get your story straight, I already covered this here ~

I clearly mentioned both CPU & GPU, the reason Intel couldn't do it because it's difficult (to compete) & complex!

Yet they are competing - with a sub-par solution, sure, but for a first effort GPU it's really not bad - and, crucially, you said "it has nothing to do with their patents" which is just plain untrue. It's not like having access to the patented tech makes you immediately produce a competent product, obviously, but without those you're screwed before you even start.

I know you mentioned both CPUs and GPUs, but I responded specifically to the CPU part, using VIA as an example. I could have been more explicit about that selection, sure, but it should be pretty clear from what I wrote that I was talking specifically about CPUs. And, crucially, patent licencing in GPUs is much, much more open than CPUs since X86 is locked into a mutually supporting duopoly, while there are many actors holding important GPU patents.

Wanna go back to your favorite point about networking now?

Let me give you a hint ~ it's because of that virtual server (& until recently x86) monopoly!

Apple Removes Remaining Intel Components from M2 MacBooks

Apple has removed the final remaining Intel components from its latest M2 MacBooks with the Intel JHL8040R USB4 timer chips being replaced with a pair of custom U09PY3 chips. This change was discovered by iFixIt during a recent teardown and documented by Twitter user SkyJuice with the exact...

www.techpowerup.com

... so? Someone picking another option, or Intel failing, is not proof of them "never really excelling" at something. A current failure does not negate previous success; a failure in one field does not negate a success in another. You're making major logical leaps here with nothing much supporting them, all to back up an exceptionally overbroad statement rather than admitting that it was bombastic and overbroad to begin with.

You seem to excel in pedantry, how about you back your arguments with more than just "the world knows"? I gave you real world example of how Intel's failed at 4G modem, they utterly failed in even getting a 5G one to work!

So, let me get this straight: you make an exceptionally overbroad statement of "outside of x86 CPUs, Intel has never really excelled at anything else", which I object to, bringing up multiple examples of product segments in which they have been or are currently very successful, and that amounts to pedantry to you? I don't think you're quite grasping just how broad and bombastic your first claim was - I'm just trying to add some desperately needed nuance here. You seem to express yourself in bold capital letter-type statements, which means there is a lot of detail lacking in both arguments and purportedly factual statements. Are x86 CPUs the main reason for Intel's success, and the core of their business throughout the past three decades or so? Yes, absolutely. Does that amount to them not excelling at anything else? Not even slightly.

 

[Selaya]

i mean ... arent/werent intel ssds like, par excellence (ignoring the p660 and other qlc memes) - the problem is just, the nand/ssd market is an open one where intel doesn't hold the duopoly, so they were forced to compete on price (-performance), and weren't able to just like hey, this is what we've got, and this is how much we're charging for it - take it or leave it as they are w/ their cpus, basically so they eventually just gave up on that market

 

[80-watt Hamster]

i mean ... arent/werent intel ssds like, par excellence (ignoring the p660 and other qlc memes) - the problem is just, the nand/ssd market is an open one where intel doesn't hold the duopoly, so they were forced to compete on price (-performance), and weren't able to just like hey, this is what we've got, and this is how much we're charging for it - take it or leave it as they are w/ their cpus, basically so they eventually just gave up on that market

This seems to be an emerging theme (or more likely one I'm just now noticing) in highly competitive markets. "Gaming" memory brands have been popping up like weeds, so Crucial axes Ballistix. Chinese headphones are taking over the world, so Sennheiser exits the consumer market entirely (licensing the Sennheiser name to someone else, IIRC). Kingston sells HyperX to HP. Honeywell sold/spun off all of its consumer product lines decades ago, as did GE. There are probably innumerable other examples.