Thursday, August 19th 2021

Intel Beats AMD to 6nm GPUs, Arc "Alchemist" Built on TSMC N6 Process

In its 2021 Architecture Day presentation, Intel revealed that its first performance gaming GPU, the Arc "Alchemist," is built on the TSMC N6 silicon fabrication node (6 nm). A more advanced node than the N7 (7 nm) used by AMD for its current RDNA2 GPUs, TSMC N6 leverages EUV (extreme ultraviolet) lithography, and offers 18% higher transistor density, besides power improvements. "With N6, TSMC provides an optimal balance of performance, density, and power-efficiency that are ideal for modern GPUs," said Dr Kevin Zhang, SVP of Business Development at TSMC.

With working prototypes of "Alchemist" already internally circulating as the "DG2," Intel has beaten AMD to 6 nm. Team Red is reportedly planning optical-shrinks of its RDNA2-based "Navi 22" and "Navi 23" chips to TSMC N6, and assigning them mid-range SKUs in the Radeon RX 7000 series. The company will build two higher-segment RDNA3 GPUs on the more advanced TSMC N5 (5 nm) process, which will release in 2022, and power successors to the RX 6700 series and RX 6800/6900 series.
Add your own comment

50 Comments on Intel Beats AMD to 6nm GPUs, Arc "Alchemist" Built on TSMC N6 Process

#26
AnarchoPrimitiv
In a just world, Intel wouldn't be allowed to use TSMC and would have to stick with their own fabs. Tgis however, proves beyond any doubt, that Intel entering the GPU market will do absolutely NOTHING to alleviate shortage since it will just be stretching an already thin supply thinner.
Posted on Reply
#27
HisDivineOrder
So... another company heavily reliant on the same fab that can't produce enough parts for AMD?

Great. That sounds like it's going to end well.
Posted on Reply
#28
InVasMani
Article on Intel GPU and most interesting part about is about AMD's RNDA3. Really curious if the 6nm shrink will come before the 5nm shrink for mid range and high end SKU's on RX7000. If AMD manages to get 6nm mid range RDNA3 released before the high end RNDA3 that would be really good at a time like now with the supply constraints. The market could very much stand for that to happen.
Posted on Reply
#29
Bomby569
ZoneDymoNo mentioning of beating Nvidia?
that's an editorial note, Intel didn't claim that
Posted on Reply
#30
Hachi_Roku256563
Im actually exicted for intel
as someone who likes to try all the brands probs gonna buy a intel gpu
Posted on Reply
#31
eidairaman1
The Exiled Airman
dyonoctisI do hope for Intel that they won't suffer the Raja curse (Hyping a product to the moon, only to have the end result being a tad underwhelming)
Too late
Posted on Reply
#32
watzupken
Eh, TSMC 6nm is basically a refined 7nm, so I don't really think it is a significant feat is it? And based on rumours, AMD is only using 6nm for their lower end 7000 series GPU. The node aside, I think I would be more interested to see if GPUs from Intel are more power efficient as compared to RDNA2 and Ampere.
Posted on Reply
#33
InVasMani
CGI Raja looking very fake...nice try Intel.
Posted on Reply
#34
watzupken
HisDivineOrderSo... another company heavily reliant on the same fab that can't produce enough parts for AMD?

Great. That sounds like it's going to end well.
I feel the supply for Intel chips may be better than AMD despite the fact that both are using TSMC. There is certainly a supply crunch at the fab, but I suspect one of the reason for AMD falling so behind in GPU supply is because,
1. Console SOC - I believe this is one of the sources that is keeping AMD really busy
2. Prioritisation of high margin products - GPUs are harder to manufacturer (and smaller in margin) and I believe AMD is prioritising CPUs, particularly EPYC series, to maximise the profit they can make. That is why it is so hard to find AMD GPUs, particularly at the higher end, i.e. RX 6800 series. The 6900 series is simply being priced out of the market, so it is not uncommon to see a lot of 6900 XT.
Posted on Reply
#35
bug
thesmokingman:laugh:


"They'll find a way to make it the best number"
Except these will be built by TSMC, Intel doesn't get to name their process ;)
Posted on Reply
#36
medi01
thesmokingman:laugh:


"They'll find a way to make it the best number"
Bashing Intel for using more realistic 7 for their process formerly called "10nm" is not warranted.
Intel was not in fab business and they didn't need to state BS figures for marketing purposes.

I keep repeating it, transistor size in L1 cache, Ryzen 7nm TSMC vs Intel 14nm (yes, FOURTEEN):
AMD: 22x22nm
Intel: 24x24nm

Not what a sane person would have expected from "7 vs 14".

I didn't dig into details, but wouldn't be surprised if Intel 10nm actually beats TSMC 7nm.
TheoneandonlyMrKWhat the actual Fffff, Intel haven't released an actual GPU.
You are new to Raja The Smoke And Mirrors (also known as Raja The Great Embarassment) aren't you? :D
AnarchoPrimitivn a just world, Intel wouldn't be allowed to use TSMC and would have to stick with their own fabs.
But why the heck not?

I don't think that's how things work.
Customers book capacity in advance.
Customers use the said capacity.

All parties are happy.


Now, call me a pessimist, but I cannot imagine Intel making any sort of significant volume booking at TSMC at this point in time. And if so, their impact on the availability will be negligible.
Posted on Reply
#37
TheoneandonlyMrK
medi01Bashing Intel for using more realistic 7 for their process formerly called "10nm" is not warranted.
Intel was not in fab business and they didn't need to state BS figures for marketing purposes.

I keep repeating it, transistor size in L1 cache, Ryzen 7nm TSMC vs Intel 14nm (yes, FOURTEEN):
AMD: 22x22nm
Intel: 24x24nm

Not what a sane person would have expected from "7 vs 14".

I didn't dig into details, but wouldn't be surprised if Intel 10nm actually beats TSMC 7nm.



You are new to Raja The Smoke And Mirrors (also known as Raja The Great Embarassment) aren't you? :D
No and I'm not new to Intel's four year advertising campaign for Xe, who is at this point.
Posted on Reply
#38
Jism
watzupkenEh, TSMC 6nm is basically a refined 7nm, so I don't really think it is a significant feat is it? And based on rumours, AMD is only using 6nm for their lower end 7000 series GPU. The node aside, I think I would be more interested to see if GPUs from Intel are more power efficient as compared to RDNA2 and Ampere.
even 7nm cant be compared to intels 10nm or even 14nm really.

It's just advances in nodes really, with either lower power consumption at same clocks compared to 7nm or higher performance at the same power enveloppe to also 7nm.

All this fuss; just release the product(s).
Posted on Reply
#39
Bomby569
JismAll this fuss; just release the product(s).
I think releasing the hardware is easier then having it work with proper software. AMD never got their drivers right, i fear Intel will have a long road ahead in making this work.
Posted on Reply
#40
Richards
I
medi01Bashing Intel for using more realistic 7 for their process formerly called "10nm" is not warranted.
Intel was not in fab business and they didn't need to state BS figures for marketing purposes.

I keep repeating it, transistor size in L1 cache, Ryzen 7nm TSMC vs Intel 14nm (yes, FOURTEEN):
AMD: 22x22nm
Intel: 24x24nm

Not what a sane person would have expected from "7 vs 14".

I didn't dig into details, but wouldn't be surprised if Intel 10nm actually beats TSMC 7nm.



You are new to Raja The Smoke And Mirrors (also known as Raja The Great Embarassment) aren't you? :D


But why the heck not?

I don't think that's how things work.
Customers book capacity in advance.
Customers use the said capacity.

All parties are happy.


Now, call me a pessimist, but I cannot imagine Intel making any sort of significant volume booking at TSMC at this point in time. And if so, their impact on the availability will be negligible.
Intel has bigger density sram. Thats why intel's 14nm is still close to tsmc 7nm in performance
Posted on Reply
#41
medi01
Bomby569never got their drivers righ

Jesus Christ, unbelievable that someone pops up with this kind of low effort garbage in 2021...
Posted on Reply
#42
ZoneDymo
medi01
Jesus Christ, unbelievable that someone pops up with this kind of low effort garbage in 2021...
ignorant fanboysm does not conform to a year or period sadly.
Posted on Reply
#43
Bomby569
medi01
Jesus Christ, unbelievable that someone pops up with this kind of low effort garbage in 2021...
being an ex-owner of the disastrous rx5700xt i know what i'm talking about.
Posted on Reply
#44
TheoneandonlyMrK
Bomby569being an ex-owner of the disastrous rx5700xt i know what i'm talking about.
How's about you put your big boy pants on , sit chill and have a think.

Is it possible your experience isn't everyone's.

Yes, yes it is.

Now your initial point was valid, Intel have some work ahead they need to get the driver right.
But bringing AMD driver's to the table is just toolish , pointless, trolly and distracting, but do ramble on off topic about your butt hurt.
Posted on Reply
#45
TheinsanegamerN
TheoneandonlyMrKHow's about you put your big boy pants on , sit chill and have a think.



Now your initial point was valid, Intel have some work ahead they need to get the driver right.
But bringing AMD driver's to the table is just toolish , pointless, trolly and distracting, but do ramble on off topic about your butt hurt.
You seem awfully upset at AMD's consistent pattern of screwing up drivers and leaving them broken until the media starts talking about it. They did it with evergreen, GCN (remember FCAT tests?), polaris, and rDNA. This is what we call a pattern. It was enough of a "trolly" response that AMD looked into the issues each time and FOUND ISSUES each time.
TheoneandonlyMrKIs it possible your experience isn't everyone's.


Yes, yes it is.
You should take your own advice.
Posted on Reply
#46
TheoneandonlyMrK
TheinsanegamerNYou seem awfully upset at AMD's consistent pattern of screwing up drivers and leaving them broken until the media starts talking about it. They did it with evergreen, GCN (remember FCAT tests?), polaris, and rDNA. This is what we call a pattern. It was enough of a "trolly" response that AMD looked into the issues each time and FOUND ISSUES each time.

You should take your own advice.
As for on topic?! ?
Posted on Reply
#47
Bomby569
TheoneandonlyMrKHow's about you put your big boy pants on , sit chill and have a think.

Is it possible your experience isn't everyone's.

Yes, yes it is.

Now your initial point was valid, Intel have some work ahead they need to get the driver right.
But bringing AMD driver's to the table is just toolish , pointless, trolly and distracting, but do ramble on off topic about your butt hurt.
That's a bit an asshole type of talking to someone even on the internet. Forgeting that, you could see the several driver releases where AMD itself listed the problems, including the disastrous black screen, that appeared as solved time and time again because they couldn't solve it.
Posted on Reply
#48
TheoneandonlyMrK
Bomby569That's a bit an asshole type of talking to someone even on the internet. Forgeting that, you could see the several driver releases where AMD itself listed the problems, including the disastrous black screen, that appeared as solved time and time again because they couldn't solve it.
Likewise, your comments sooooooo, and on topic ?!

I used AMD and Nvidia 20+ years sooo I will look up nothing I already know, now with big boy pants on wtaf does driver's have to do with news of Intel beating AMD to a 6nm node, nothing.

Pm your next insult at least then I can truly reply how I'D like.
Posted on Reply
#50
mtcn77
TheoneandonlyMrKLikewise, your comments sooooooo, and on topic ?!

I used AMD and Nvidia 20+ years sooo I will look up nothing I already know, now with big boy pants on wtaf does driver's have to do with news of Intel beating AMD to a 6nm node, nothing.

Pm your next insult at least then I can truly reply how I'D like.
Don't hold back, at least you get the thread removed, at most you get your comment removed: it is a %50-50 tossup.
Posted on Reply
Add your own comment
Dec 22nd, 2024 11:44 EST change timezone

New Forum Posts

Popular Reviews

Controversial News Posts