News Intel GPUs - Falcon Shores cancelled

Page 229 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

Win2012R2

Senior member
Dec 5, 2024
647
609
96
People do NOT care if Intel is losing money or not, people are mostly concerned about getting a good value product. It's as simple as that.
Would they change their mind if or when new Intel board cuts GPU money losing business and there will be no more "day 1" driver updates? I guess we'll find out soon.

Ironically the more they buy the more likely that will happen.
 

GTracing

Senior member
Aug 6, 2021
276
645
106
And the Pro Fit has spoken...


Well, even a broken clock MLID can be right from time to time, volume is clearly low because the more you buy the more Intel loses!

P.S. Looks like MLID is feeling the heat from the current popular opinion that new Arc is great, so he is adjusting his narrative a tad.

Every time I watch a MLID video I regret it. It's like he's trying to be as condescending and elitist as possible.
But I mean come, is this level on performance on screen here, a third the performance of the competition's two year old flagship that's cut down, is that enthusiast? I'm not going to answer that question for you because, honestly, if you can't answer that question honestly for your self looking at the data this week, then I think it's time to stop trying to convince you what's going on here. That a 12GB 2080 super in 2024, it shouldn't be hard to discern if that's actually enthusiast.
I guess poor people can't be hardware enthusiasts.
 

Win2012R2

Senior member
Dec 5, 2024
647
609
96
I guess poor people can't be hardware enthusiasts.
PC hardware enthusiast in old days - getting cheap stuff and overclocking it to gain a LOT more perf.
PC hardware enthusiast these days - paying top dollar (often to scalpers) for factory overclocked, supply throttled hardware, often for marginal gains.

MLID is really getting a bashing in the comments section - usually it's full of his loyal fans only, but today he is furiously copy/pasting canned reply a lot
 

Meteor Late

Senior member
Dec 15, 2023
266
292
96
Also, it's funny MLID and many comments always talking about how it will be bad soon because in a few months RTX 5000 and RX 8000 are coming, but at the same time, this doesn't matter one bit when talking about Strix Halo, always talking about "RTX 4070 level of performance" and whatnot, apparently in this case RTX 5000 are not coming in laptops next year of course...
 

Meteor Late

Senior member
Dec 15, 2023
266
292
96
Would they change their mind if or when new Intel board cuts GPU money losing business and there will be no more "day 1" driver updates? I guess we'll find out soon.

Ironically the more they buy the more likely that will happen.

That's not on consumers, that's on Intel, are we expecting every consumer to be an investor and make a full analysis of how Intel operates in terms of margins now? that's ridiculous.
 

ajsdkflsdjfio

Member
Nov 20, 2024
171
117
76
Would they change their mind if or when new Intel board cuts GPU money losing business and there will be no more "day 1" driver updates? I guess we'll find out soon.

Ironically the more they buy the more likely that will happen.
Most of your posts seem like uneducated guesses at best and intentional trolling at worst. Like you you claimed Battlemage was using 20k$ n4 wafers while the rx 7600 was using 9.5k N7 wafers....

You dumb or something lol, the battlemage cards themselves aren't a massive money sink, they just likely make little to no profit. The idea that intel is actively losing money for each battlemage card they produce is ludicrous. AMD is able to sell their cards at a margin high enough that they are able to cover all the R&D, manufacturing, marketing, etc costs while also making a +15% profit after all those costs. What does that show you about what the cards themselves cost to make. Intel using a +40 dollar die and +10 dollars for extra 4gb memory isn't enough to put their BOM over 250$ and still likely has room enough for AIB margins and associated sales costs.

Besides if Intel was so concerned about starting to turn a profit right at this instant they would have priced the card higher. I think the fact that they're willing to take a loss on margins trading for increased sales/mindshare means that they are playing the long game likely meaning they are committed to the graphics market for the foreseeable future. They aren't going to cut the graphics department simply because their second generation graphics cards aren't a massive profit machine yet, not even the board/new ceos are that short sighted.
 
Reactions: DigDog

Win2012R2

Senior member
Dec 5, 2024
647
609
96
you claimed Battlemage was using 20k$ n4 wafers while the rx 7600 was using 9.5k N7 wafers....
I provided sources to the costs, they are also easily obtained by Googling -
1) tsmc n7 wafer cost
2) tsmc n4 wafer price

Intel using a +40 dollar die

False.

In order for this to be true a 300 sq mm chip with perfect yield giving 200 per wafer would need to have wafer cost of $8000.

AMD is able to sell their cards at a margin high enough that they are able to cover all the R&D, manufacturing, marketing

AMD has got revenues from consoles that sell better than their cards, well, until Xbox sales tanked, now that's a problem.

That's not on consumers

In the end it's always on the buyers - caveat emptor!
 

ajsdkflsdjfio

Member
Nov 20, 2024
171
117
76
I provided sources to the costs, they are also easily obtained by Googling -
1) tsmc n7 wafer cost
2) tsmc n4 wafer price
Your source said 18k for n4, battlemage uses N5 which is quoted as 15-16k. If n4 is more expensive than N5, N6 is also likely more expensive than N7. Guess what uses N6 not N7... the rx 7600. N7 is listed at 10k, even being genererous N6 is probably around 11k at least.
So instead of 20k vs 9.5k like you said, it's 15-16k vs 11k. Regardless of your grossly misquoted prices, how many times do I gotta tell you that battlemage is n5 not n4, and rx 7600 is n6 not n7. N5 =/= N4 and N7 =/= N6.
In order for this to be true a 300 sq mm chip with perfect yield giving 200 per wafer would need to have wafer cost of $8000.
I said +40 dollars meaning +40 dollars from AMD rx 7600 die(twice as expensive). and +10 for 4gb memory. All other costs being similar that means that the b580 only costs around 50 dollars more to produce than the rx 7600, well within margin to break even. AMD prices their cards to have a margin so they can pay for R&D and other costs. Intel pricing their cards low to gain mindshare and cuz even pricing their cards higher likely not enough to pay for R&D anyways, they are invested long term.
AMD has got revenues from consoles that sell better than their cards, well, until Xbox sales tanked, now that's a problem.
Their consoles sell better than their cards but at the same time have lower margin, which is the entire point of a console, low margin but high volume of the same one product. AMD makes a healthy margin off their cards, it is not crazy to think that battlemage is not losing intel money for each card they sell.
 
Last edited:
Reactions: gaav87

Win2012R2

Senior member
Dec 5, 2024
647
609
96
Your source said 18k for n4, battlemage uses N5 which is quoted as 15-16k.

"Negotiations with AI and HPC customers, such as Nvidia, suggest these clients can tolerate approximately 10% price hikes for 4nm-class wafers from around $18,000 per wafer to around $20,000 per wafer. As a result, the 4nm and 5nm nodes, primarily used by companies like AMD and Nvidia, are expected to see an 11% blended average selling price (ASP) hike." Source:
If N5 is indeed 16k, then price goes up by 11% in the new year and becomes $18k - happy now? Does it change things materially? No.

I said +40 dollars meaning +40 dollars from AMD rx 7600 die(twice as expensive)

False for obvious reasons - it simply can't be true: rx 7600 is 204 sq mm die (techpowerup data) and it's on N6 which is priced at $9500 as Google search above shows, Intel uses 50% more silicon for DOUBLE the price, yields will also be worse - they are 3-3.5 times more expensive to make than rx 7600, that can not be +40 bucks unless AMD pays 15 bucks for their dies.

Intel pricing their cards low to gain mindshare

Well assumption does not contract that they lose money on each sale, does it? Price is the only thing going for it, who'd recommend it if it was priced at $400 like it should have been for 300 sq mm N5-class silicon. PS5 Pro is on N4, I would find it very surprising if this chip isn't using it too. N4 to N5 is what N6 to N7.
 
Last edited:

ajsdkflsdjfio

Member
Nov 20, 2024
171
117
76
"Negotiations with AI and HPC customers, such as Nvidia, suggest these clients can tolerate approximately 10% price hikes for 4nm-class wafers from around $18,000 per wafer to around $20,000 per wafer. As a result, the 4nm and 5nm nodes, primarily used by companies like AMD and Nvidia, are expected to see an 11% blended average selling price (ASP) hike." Source:
If N5 is indeed 16k, then price goes up by 11% in the new year and becomes $18k - happy now? Does it change things materially? No.
What? How do you draw that N5 prices will for sure go up by 11% based on that quote from tom's hardware about N4? Why do you apply this stretch of reasoning to increasing N5 prices and not N6?
False for obvious reasons - it simply can't be true: rx 7600 is 204 sq mm die (techpowerup data) and it's on N6 which is priced at $9500 as Google search above shows, Intel uses 50% more silicon for DOUBLE the price, yields will also be worse - they are 3-3.5 times more expensive to make than rx 7600, that can not be +40 bucks unless AMD pays 15 bucks for their dies.
Intel uses +33% (not 50%) more silicon with 1.4x the wafer price MAX since N5 is quoted as being between 15-16k and N6 is likely at least 11K. Look at Joe's above posts, his own calculations also yielded around a 2x die cost. Are you intentionally making huge miscalculations in your data or are you just stupid?? How is 272 vs 204mm a 50% increase, a quick calculator result can tell you its 33% not 50%. Also even with your 50% die size and 2x wafer cost BS, how do you get 3.5 times more expensive for the die. Did you just add the .5 for fun?



Also based on your own source (first google search results) N7 is 10k not 9500. And what should be obvious to you is that if N7 is already 10k, then N6 is obviously more expensive. Why are you using a random 9.5k number when its 10k for N7, and why are you using N7 prices for N6 and N4 prices for N5. The fuck?? Ironically you'd be perfect for Intel's marketing and sales department since you already know how to max abuse and misrepresent data to further your point.

 
Last edited:

Win2012R2

Senior member
Dec 5, 2024
647
609
96
if N7 is already 10k, then N6 is obviously more expensive.
No, that's not obvious at all if you actually follow how semi manufacturing nodes get progressed and for what reasons.
Intel uses +33% (not 50%) more silicon
I rounded it a bit up (200 vs 300), relax, now factor in yield that you conveniently ignore, you might get to that 50% a lot closer than you think.
 
Feb 4, 2009
35,664
17,154
136
Irregardless intel is an experienced company, they understand margins and chip costs. This has already been budgeted and I am glad we have a third competitor that is trying to do something different in this field.
 

ajsdkflsdjfio

Member
Nov 20, 2024
171
117
76
No, that's not obvious at all if you actually follow how semi manufacturing nodes get progressed and for what reasons.
Yes it is obvious. N4 is 2-3k more expensive while providing around 6 percent transistor density and 11% performance. N6 provides 18% transistor density and double digit performance numbers and somehow you think it isn't more expensive than N7?
I rounded it a bit up (200 vs 300), relax, now factor in yield that you conveniently ignore, you might get to that 50% a lot closer than you think.
I don't purposefully ignore it, you never addressed it in your calculations which I was responding to. Even factoring in yield, it's unlikely to make more than a single digit difference as yields on these more mature nodes are all relatively good. Also it's not just about chip size and defect rate, it's also about parametric yield. Intel's chip is much less dense compared to other N5 chips meaning it likely yields better.
 

ajsdkflsdjfio

Member
Nov 20, 2024
171
117
76
Yes, more than a year ago when Pat was running the show. Nothing changed since then...
Like I said, they could have priced it at 300$ if they wished to making up for the +40 dollars die +10 ram. Regardless of the budget of the BOM, Pat is no longer at Intel and they could have chosen to price it 300, they don't need a year in advance to change the price. They priced it at 250$ because their margins can likely take it and they aren't focused on maximizing profit right now anyways. If they were losing money on each battlemage card at 250$ I guarantee they would price it higher because it would still sell even at 300 dollars.
 

Win2012R2

Senior member
Dec 5, 2024
647
609
96
N6 provides 18% transistor density and double digit performance numbers and somehow you think it isn't more expensive than N7?
Not now it isn't, what was N7 became N7/6 - same rules, but it's up to chip designers to shrink design to take full advantage of N6 which will give benefit of reduction in area size/perf. If it was more expensive than Sony would have never switched to it in PlayStation 5.

they aren't focused on maximizing profit

It's their fidiciary duty to do so, they could go to jail if they don't.
 

ajsdkflsdjfio

Member
Nov 20, 2024
171
117
76
Not now it isn't, what was N7 became N7/6 - same rules, but it's up to chip designed to shrink design to take full advantage of N6 which will give benefit of reduction in area size/perf. If it was more expensive than Sony would have never switched to it in PlayStation 5.
No they aren't the same. They are separate products that customers can choose between. Just look at N3, N3E, N3P, N3x, N3B. They all still exist separately and have different costs. Why if TSMC is maintaining so many separate N3 N4 etc derivatives at once, you think that N7 is discontinued because N6 exists...

Many of TSMC nodes are design rule compatible, for example their upcoming N2 node will be fully design compatible with A16, TSMC still keeps both as options for customers regardless because they have different costs and different advantages.
It's their fidiciary duty to do so, they could go to jail if they don't.
What are you even talking about. It's not companies fidiciary duty to maximize short term profits, there is more to a company's strategy than just milking as much money as they can right now. Did you even read what I said btw? I said they aren't focused on maximizing profits RIGHT NOW, why did you conveniently leave that out of the quote??? They are clearly trading short term profits on this generation for mind/market share and to generally get their cards out there so they can have a testing bed to improve drivers.
 
Last edited:
Reactions: Grazick

Joe NYC

Platinum Member
Jun 26, 2021
2,790
4,103
106
I still think it’s most likely that Intel is building this business for their fabs to have a good customer in the future vs making money now. A calculated investment. Probably

Yup. That's the only way it makes sense to me. What I don't get is why not have these on "Intel 3". Then, Intel can have these as sort of a "variable" load on the fab, build a few here and there when the fabs are not fully utilized.

Increasing the utilization of the fab would increase its profitability and lower the cost to fab the parts...

"Intel 3" should not be too inferior to NVidia N5. It should be in the same ballpark.
 

Hitman928

Diamond Member
Apr 15, 2012
6,524
11,796
136
No they aren't the same. They are separate products that customers can choose between. Just look at N3, N3E, N3P, N3x, N3B. They all still exist separately and have different costs. Why if TSMC is maintaining so many separate N3 N4 etc derivatives at once, you think that N7 is discontinued because N6 exists...

Many of TSMC nodes are design rule compatible, for example their upcoming N2 node will be fully design compatible with A16, TSMC still keeps both as options for customers regardless because they have different costs and different advantages.

N3 and N3B are the same process and it is going away. TSMC will continue to run it for Intel/Apple in smaller numbers as they need to support already existing products but no one's going to use it for new products. N7 is pretty much gone as well, TSMC has pushed all new designs to use N6 because it is cheaper due to it being a EUV process that needs fewer masks and passes and there's really no reason to continue to run N7 when N6 is a win-win for customers and foundry.
 
Reactions: Tlh97 and Win2012R2

Win2012R2

Senior member
Dec 5, 2024
647
609
96
They are clearly trading short term profits on this generation for mind/market share and to generally get their cards out there so they can have a testing bed to improve drivers.

Let's bookmark this page and revisit it soon when the beancounting BoD cans the ArchMage like a tuna

What I don't get is why not have these on "Intel 3"

Must be cost (of opportunity as well as ~300 sq mm of Xeon is worth a lot more than 250 bucks retail), plus that node probably would not live long enough anyway.
 
Last edited:

Hitman928

Diamond Member
Apr 15, 2012
6,524
11,796
136
In regards to BM, I am intrigued for sure by the B580, it has great perf/$ in general with a good amount of VRAM. It wouldn't be an upgrade for my computer, but I'm considering it for the kids' PC. My concern is both with having to deal with any driver issues for games they want to play (lesser concern) and Intel not giving good driver support in the future like has happened with the Alchemist line. I don't want to buy it just to have it drop in competitive performance when Intel is preparing for Celestial because their resources are now all switched to supporting the incoming generation. If Alchemist still had good driver support for new games, I wouldn't be as concerned. Something to think about but I might just wait for Celestial to see if Intel will continue to support their N-1 dGPUs going forward.
 
Reactions: Racan
sale-70-410-exam    | Exam-200-125-pdf    | we-sale-70-410-exam    | hot-sale-70-410-exam    | Latest-exam-700-603-Dumps    | Dumps-98-363-exams-date    | Certs-200-125-date    | Dumps-300-075-exams-date    | hot-sale-book-C8010-726-book    | Hot-Sale-200-310-Exam    | Exam-Description-200-310-dumps?    | hot-sale-book-200-125-book    | Latest-Updated-300-209-Exam    | Dumps-210-260-exams-date    | Download-200-125-Exam-PDF    | Exam-Description-300-101-dumps    | Certs-300-101-date    | Hot-Sale-300-075-Exam    | Latest-exam-200-125-Dumps    | Exam-Description-200-125-dumps    | Latest-Updated-300-075-Exam    | hot-sale-book-210-260-book    | Dumps-200-901-exams-date    | Certs-200-901-date    | Latest-exam-1Z0-062-Dumps    | Hot-Sale-1Z0-062-Exam    | Certs-CSSLP-date    | 100%-Pass-70-383-Exams    | Latest-JN0-360-real-exam-questions    | 100%-Pass-4A0-100-Real-Exam-Questions    | Dumps-300-135-exams-date    | Passed-200-105-Tech-Exams    | Latest-Updated-200-310-Exam    | Download-300-070-Exam-PDF    | Hot-Sale-JN0-360-Exam    | 100%-Pass-JN0-360-Exams    | 100%-Pass-JN0-360-Real-Exam-Questions    | Dumps-JN0-360-exams-date    | Exam-Description-1Z0-876-dumps    | Latest-exam-1Z0-876-Dumps    | Dumps-HPE0-Y53-exams-date    | 2017-Latest-HPE0-Y53-Exam    | 100%-Pass-HPE0-Y53-Real-Exam-Questions    | Pass-4A0-100-Exam    | Latest-4A0-100-Questions    | Dumps-98-365-exams-date    | 2017-Latest-98-365-Exam    | 100%-Pass-VCS-254-Exams    | 2017-Latest-VCS-273-Exam    | Dumps-200-355-exams-date    | 2017-Latest-300-320-Exam    | Pass-300-101-Exam    | 100%-Pass-300-115-Exams    |
http://www.portvapes.co.uk/    | http://www.portvapes.co.uk/    |