Some updates on the Radeon Technology Group (RTG)

M

_mockingbird

Gawd
Joined
Feb 20, 2017
Messages
992
As you may remember, after Raja Koduri got kicked out, Lisa Su had to find someone to replace him as head of the RTG group.

The was particularly difficult because there were three strong contenders: lisa su, Lisa su, and lisa Su, but in the end, Lisa Su chose lisa su.

lisa su introduced a very different style of management. Instead of "shooting for the moon", su emphasizes "milestones" with realistic targets and tight deadlines.

It has been strongly hinted that the reason for the tight deadlines is that AMD's top brass have been pitching AMD's roadmap to OEM vendors, and emphasizing AMD's ability to deliver products on time. (basically, a jab at Intel)

I have said many months ago that RTG was laying the foundation for a new non-GCN architecture to succeed Navi. The work on this new architecture is now well underway.
 
Lisa Su has proven to be quite competent getting the CPU division in shape. I'm sure she can do the same with the GPU division. Here's to hoping that she has the command to shepherd both divisions without allowing one to suffer.
 
Algrim said:
Lisa Su has proven to be quite competent getting the CPU division in shape. I'm sure she can do the same with the GPU division. Here's to hoping that she has the command to shepherd both divisions without allowing one to suffer.
Click to expand...

She has been working hard to make AMD able to turn on a dime.

Metaphorically, she has been going around and straightening out the kinks in the hose.
 
_mockingbird said:
It's going to be a while.

...and when I said "a while", I mean ~"2021".
Click to expand...

Damn. I was hoping it wouldn't take that long. I have skipped Vega (mainly because of the insane price) and was holding out for more performance.
 
N4CR said:
So wait till GCN is over.. Gosh they are stringing it out.
Click to expand...

Well, that's what happens as a result of poor planning.

RTG should have started working on replacing GCN soon after it became clear with the Fury/Fury X that GCN was hitting its limits.

Maddness said:
Damn. I was hoping it wouldn't take that long. I have skipped Vega (mainly because of the insane price) and was holding out for more performance.
Click to expand...

Well, Navi would be the last stopgap before the next gen arrives, but that's probably not what you are looking for.
 
_mockingbird said:
Well, that's what happens as a result of poor planning.

RTG should have started working on replacing GCN soon after it became clear with the Fury/Fury X that GCN was hitting its limits.



Well, Navi would be the last stopgap before the next gen arrives, but that's probably not what you are looking for.
Click to expand...
If they do indeed decouple the geometry limitation in the architecture, navi might actually do ok.
 
N4CR said:
So wait till GCN is over.. Gosh they are stringing it out.
Click to expand...

they have no choice but to stick with what's already released and what has been taped out.. it takes anywhere from 3-5 years to design a gpu or cpu from the ground up and lisa took over at the end of last year so the 2021 prediction makes sense.

do you think volta was designed last year? that shit was started back when maxwell was released. that's why nvidia doesn't release things based on competition, it's not cost effective for them to hold onto something since R&D time and money is put into designing these gpu's years in advance.

there was actually a reddit ama a few years back with an amd engineer that went into a lot more detail about how the time tables for gpu's and cpu's release but i doubt i can find it again since searching on reddit is aid's.
 
sirmonkey1985 said:
they have no choice but to stick with what's already released and what has been taped out.. it takes anywhere from 3-5 years to design a gpu or cpu from the ground up and lisa took over at the end of last year so the 2021 prediction makes sense.

do you think volta was designed last year? that shit was started back when maxwell was released. that's why nvidia doesn't release things based on competition, it's not cost effective for them to hold onto something since R&D time and money is put into designing these gpu's years in advance.

there was actually a reddit ama a few years back with an amd engineer that went into a lot more detail about how the time tables for gpu's and cpu's release but i doubt i can find it again since searching on reddit is aid's.
Click to expand...
More than aware of how the timetables go but they have had gcn out since 2011 or so, when I bought a 7970 at launch. Amazing card and my favourite, was good for a 40% OC on a ref card. Absolutely ridiculous performance. But problems were already apparent with the Crysis 2 geometry fiasco and have only continued. So I'm curous why they didn't work on this further, earlier.
That said, GCN is great at everything else and especially at compute though.
 
Last edited:
https://www.theinformation.com/articles/intels-manufacturing-glitch-opens-door-for-amd

Full text

Advanced Micro Devices’ battle with chip giant Intel has often seemed like a gnat fighting an elephant, with AMD struggling in recent years to gain even a tenth of the market for the chips that power PCs and data center servers. Forrest Norrod, a senior executive who joined AMD four years ago, says the company suffered from “little brother syndrome” where it tried and failed to compete with Intel on lots of different chips.

Now, though, AMD may have a shot at coming out with a faster, more powerful chip than Intel for the first time. Intel in April said it was delaying the release of a more advanced chip manufacturing process until sometime in 2019. AMD has its own new, advanced chip, which it will now be able to release earlier than Intel, potentially giving it an edge in the market for high-performance chips for PCs and data center computers.

It's a market opportunity worth around $50 billion. That’s what Intel makes from selling chips or PCs and data center servers, and it dominates both markets. The data center market is particularly important because of the growth of new technologies like artificial intelligence–related applications, much of which is handled in the cloud. Companies that buy chips for data centers or PCs could gravitate to AMD chips as a result of Intel’s delay.

“I think we have a year lead now,” said Mr. Norrod. AMD now has “an unexpected [manufacturing] advantage for the first time ever,” he added.

AMD wasn’t in a position to take advantage of Intel’s product delays until recently. CEO Lisa Su, who started in 2014 and hired Mr. Norrod shortly afterwards, has focused on a fewer number of chips where it had more of an edge, such as high-performance processors at lower prices than Intel. That was a reversal of AMD’s past “little brother” strategy of competing more broadly.

Mr. Norrod freely admits that when he joined AMD, it was “in really bad shape.” Intel was sucking up all its traditional business of selling processing chips for PCs and data centers. Nvidia, meanwhile, was establishing a lead in selling graphics processors into the nascent artificial intelligence market.

There are signs of progress. AMD’s revenue was up 23% last year and it turned a small profit of $204 million, after losing money the previous two years. AMD’s market share in the central processors for PCs and data centers has risen to 8.6% from 7.9% between the fourth quarter of last year and the first quarter of this year, according to Mercury Research.

Still, any fight between AMD and Intel is a David-and-Goliath struggle. Intel is a giant, with $64 billion in revenue, roughly $10 billion in net income and a market capitalization of $260 billion. AMD, on the other hand, has revenue of a little above $5 billion, a market capitalization of $14.4 billion and a history of struggling to even turn a profit.

Intel long had a manufacturing advantage because it makes its own chips. AMD contracts its manufacturing out to companies like TSMC and GlobalFoundries, which originally spun out of AMD in 2009.

But Intel has been having problems making chips at a 10 nanometer density—a measure of how closely it packs transistors together on a chip, which has traditionally been important for continuing to improve chip performance. As a result, Intel continues to make chips based on its older 14 nanometer-density. AMD, in contrast, announced this week it is testing the first version of a graphics processing chip using 7-nanometer density technology, being made by TSMC, which is comparable to what Intel is trying to make.

Intel didn’t respond to a request for comment. Last month, at Intel's annual shareholder meeting, CEO Brian Krzanich said there was "absolutely not" a competitive threat posed by its manufacturing difficulties. “We think we have one of the best roadmaps that Intel’s had in its history moving forward,” he said.

The manufacturing glitch is the latest in a series of challenges for Intel. It missed the mobile market and has struggled to diversify into areas like cybersecurity and wearables such as smart glasses. But it dominates the market for PC and data center server chips.

Graphics Counterattack

Intel is now mounting a counterattack on graphic chips, where it has historically been weak. Intel recently nabbed the chief of AMD’s graphic chip unit, Raja Koduri, to head a new high-end graphics chip unit. Intel has tried and failed to build graphics chips several times in the past. But graphics processing units, also called GPUs, have become a more important market in recent years. Nvidia, for instance, which specializes in graphics chips, and is selling them in new areas like artificial intelligence and cryptocurrencies.

Intel also recently hired well-regarded processor architect Jim Keller, who led the design of the current processing chips of AMD before he left in late 2015. That hire was a sign that “Intel is focusing more on chip architecture,” said Bernstein Research analyst Stacy Rasgon. “When you can’t have [manufacturing] advantage, architecture becomes more important.”

Intel apparently enticed Mr. Koduri by giving him oversight of a huge new organization, the Core and Visual Computing Group, and making him chief architect at Intel. Mr. Koduri had clashed with AMD management over the future direction of the graphics chip unit, which he wanted to spin off and make independent, said two people familiar with the situation. (The clashes were previously reported by PC hardware website HardOCP.

Mr. Norrod played down the significance of the hiring. “Raja is a brilliant guy, a brilliant architect, a brilliant marketer, but he lives in the future,” he said. “He’s running a massive organization, and based on my perceptions of Raja and my experience working with him, that doesn’t play to his strengths. We’ll see what happens.”

He also claimed that AMD’s graphic chip business is doing a little better since Mr. Koduri’s departure. Under Mr. Koduri, AMD changed product plans for new graphic chips several times. “We had some wastage on the GPU roadmap where he changed his mind,” he said. “I think that since we’ve really locked that down, we’re making a little bit more rapid progress.”

Mr. Koduri didn’t respond to a request for comment.

Because of the time it takes to build a new chip, Mr. Norrod doesn’t expect Messrs. Koduri and Keller to have any impact at Intel for at least another three years. He argues that gives AMD, for now, an historical advantage it’s never had before.

“We didn’t expect Intel to stumble as they have on the manufacturing side,” said Mr. Norrod. “We never expected that. This is a real tailwind for us. I liked our previous plan, but it’s a gift that Intel has given us. We just need to make sure we don’t screw it up.”
Click to expand...

Interesting article which includes statements from Forrest Norrod, head of AMD's Datacenter and Embedded Group, of which I found most interesting were his comments on RTG and Raja Koduri.
 
I think AMD will miss Keller, but i'm happy Koduri has left the GPU department.
 
Underestimating Intel and gloating over their misfortune, isn't that how last time AMD management screwed up badly ?
 
Rizen said:
I'm not sure how you got that perception from his comments.
Click to expand...
When you assign blame to just one person that went to Intel (they don't have money problems) is not only unrealistic but also far from the truth when you seen the personal changes in RTG.

He also claimed that AMD’s graphic chip business is doing a little better since Mr. Koduri’s departure. Under Mr. Koduri, AMD changed product plans for new graphic chips several times. “We had some wastage on the GPU roadmap where he changed his mind,” he said. “I think that since we’ve really locked that down, we’re making a little bit more rapid progress.”

Mr. Koduri didn’t respond to a request for comment.

Because of the time it takes to build a new chip, Mr. Norrod doesn’t expect Messrs. Koduri and Keller to have any impact at Intel for at least another three years. He argues that gives AMD, for now, an historical advantage it’s never had before.
Click to expand...

It is clear from https://semiaccurate.com/2018/05/29/is-intels-upcoming-10nm-launch-real-or-a-pr-stunt/ and the whole GPP situation that Intel dictates a lot of things maybe you find it logical that their best 14nm parts are not used but their under performing 10nm are by OEM.

When you take 3 years to think that Intel would be back on track you are badly mistaken. Do I need to remind you of how Dell did not sell any volume of AMD products last time around when AMD had a good lead?

If you look at the following AMD marketing got caught with their pants down at Computex. The talk now is how Intel needed a chiller that can not be bought by the desktop crowd. But all the media reported on how Intel 28 core reached 5ghz and did 7743 at Cinebench. And that story should have been about AMD not Intel. Long winded youtube about this .

And yes that means that AMD management still underestimates Intel.

“We didn’t expect Intel to stumble as they have on the manufacturing side,” said Mr. Norrod. “We never expected that. This is a real tailwind for us. I liked our previous plan, but it’s a gift that Intel has given us. We just need to make sure we don’t screw it up.”
Click to expand...

Yeah but that has been the case for the last 3 years when Intel failed to deliver on 10nm. It is not about screwing up it is about to maintain a healthy lead and when Intel does come back they are still 3 years behind. And that is what AMD should focus on.
 
Pieter3dnow said:
When you assign blame to just one person that went to Intel (they don't have money problems) is not only unrealistic but also far from the truth when you seen the personal changes in RTG.



It is clear from https://semiaccurate.com/2018/05/29/is-intels-upcoming-10nm-launch-real-or-a-pr-stunt/ and the whole GPP situation that Intel dictates a lot of things maybe you find it logical that their best 14nm parts are not used but their under performing 10nm are by OEM.

When you take 3 years to think that Intel would be back on track you are badly mistaken. Do I need to remind you of how Dell did not sell any volume of AMD products last time around when AMD had a good lead?

If you look at the following AMD marketing got caught with their pants down at Computex. The talk now is how Intel needed a chiller that can not be bought by the desktop crowd. But all the media reported on how Intel 28 core reached 5ghz and did 7743 at Cinebench. And that story should have been about AMD not Intel. Long winded youtube about this .

And yes that means that AMD management still underestimates Intel.
Click to expand...


Sometimes it's best to let a competitor back themselves into a corner, in this case it will be something else that Intel cant deliver like 10nm chips and no the broken ones they are selling in China really dont count. Marketing also would not help to get Dell to use them, that had everything to do with Intel and their anti-competitive practices some years ago. Also Raja caused his own problems by over stating his own products, he was his own worst enemy and he had full control of the project, just budgeting was not within his control. A bad leader can ruin a whole project, will just have to see how that all turns out.
 
Pieter3dnow said:
Underestimating Intel and gloating over their misfortune, isn't that how last time AMD management screwed up badly ?
Click to expand...

Underestimating? You must be reading something else. Talk about blowing a statement out of proportion lol! I am not sure how you got that lol.

He was talking about missteps in the GPU roadmap. Did Raja feel like he knew what the f he was doing? Kyle stated clearly he had his own goals, he was trying to spin RTG. How about his presentations? Looked like a guy who over promised and under delivered each time. A great engineer doesn't make a great product manager. Honestly I am tired of everyone making excuses for Kaduri. I am glad he had the balls to call out Kaduri. He is no longer at the company so he can say whatever he wants. It is clear Kaduri was someone they couldn't control. Looked he was all Intel.

It wasn't even about the shitty performance and power. It was how RTG was ran. It was a failure from marketing to all the hype.
 
Last edited:
NKD said:
Underestimating? You must be reading something else. Talk about blowing a statement out of proportion lol! I am not sure how you got that lol.

He was talking about missteps in the GPU roadmap. Did Raja feel like he knew what the f he was doing? Kyle stated clearly he had his own goals, he was trying to spin RTG. How about his presentations? Looked like a guy who over promised and under delivered each time. A great engineer doesn't make a great product manager. Honestly I am tired of everyone making excuses for Kaduri. I am glad he had the balls to call out Kaduri. He is no longer at the company so he can say whatever he wants. It is clear Kaduri was someone they couldn't control. Looked he was all Intel.

It wasn't even about the shitty performance and power. It was how RTG was ran. It was a failure from marketing to all the hype.
Click to expand...
Between Kyle,Raja and RTG is a story which people thought was worth mentioning but the real story is about not getting ahead of the design problems where you end up using HBM2 because the design at that core speed performance just can not function any other way (not with GDDR).

And the way Raja saw RTG was needing cash flow more then anything else. No surprise he went to Intel basically because they have had so many trail runs that it would never amount to anything marketable.

What Raja could not do is buy $40 worth of groceries with only $15 to spend. The problems that were created are done by himself and no one else. But the designs in the AMD/RTG gpu department were already in progress.

If you see the message I quoted above the part where :
Because of the time it takes to build a new chip, Mr. Norrod doesn’t expect Messrs. Koduri and Keller to have any impact at Intel for at least another three years. He argues that gives AMD, for now, an historical advantage it’s never had before.
Click to expand...

I would say that Norrod needs to think about the choice of his words again. These are the people that in the past did very impressive things. If AMD did not have problems with cash flow I would even like to believe that these people still would be working for AMD.

And I'm not blowing this out of proportion, find 5 industry veterans which will concur with the statement Norrod made, Intel won't be doing anything significant for 3 years ....

Gideon said:
Sometimes it's best to let a competitor back themselves into a corner, in this case it will be something else that Intel cant deliver like 10nm chips and no the broken ones they are selling in China really dont count. Marketing also would not help to get Dell to use them, that had everything to do with Intel and their anti-competitive practices some years ago. Also Raja caused his own problems by over stating his own products, he was his own worst enemy and he had full control of the project, just budgeting was not within his control. A bad leader can ruin a whole project, will just have to see how that all turns out.
Click to expand...

You must have missed the part where a good deal of OEM signed the GPP but not Dell .....
 
Last edited:
Pieter3dnow said:
Between Kyle,Raja and RTG is a story which people thought was worth mentioning but the real story is about not getting ahead of the design problems where you end up using HBM2 because the design at that core speed performance just can not function any other way (not with GDDR).

And the way Raja saw RTG was needing cash flow more then anything else. No surprise he went to Intel basically because they have had so many trail runs that it would never amount to anything marketable.

What Raja could not do is buy $40 worth of groceries with only $15 to spend. The problems that were created are done by himself and no one else. But the designs in the AMD/RTG gpu department were already in progress.

If you see the message I quoted above the part where :


I would say that Norrod needs to think about the choice of his words again. These are the people that in the past did very impressive things. If AMD did not have problems with cash flow I would even like to believe that these people still would be working for AMD.

And I'm not blowing this out of proportion, find 5 industry veterans which will concur with the statement Norrod made, Intel won't be doing anything significant for 3 years ....



You must have missed the part where a good deal of OEM signed the GPP but not Dell .....
Click to expand...

he didn’t have the best budget. But he also didn’t help with constant hype train and stupid marketing. I think there was too much other shit going on at rtg rather than just cash problems. Sometimes you gotta learn to work with what you got. AMD doesn’t have intel budget either but did Lisa not tighten the belt and got them going? Sometimes execution means a lot too. I am sure raja learned too. He just wasn’t someone who could lead rtg. His first go at it was a failure. There is no 2 ways about it. I hope he doesn’t better at Intel.
 
NKD said:
he didn’t have the best budget. But he also didn’t help with constant hype train and stupid marketing. I think there was too much other shit going on at rtg rather than just cash problems. Sometimes you gotta learn to work with what you got. AMD doesn’t have intel budget either but did Lisa not tighten the belt and got them going? Sometimes execution means a lot too. I am sure raja learned too. He just wasn’t someone who could lead rtg. His first go at it was a failure. There is no 2 ways about it. I hope he doesn’t better at Intel.
Click to expand...
For us consumers yes, he over did it. But he also landed the AMD gpu deal with Intel and at that point in time was a good move. It can still be a good move if Intel can manage to sell these into the market now dominated by Nvidia.

I get where you are going but I think there is more to it.
 
Pieter3dnow said:
For us consumers yes, he over did it. But he also landed the AMD gpu deal with Intel and at that point in time was a good move. It can still be a good move if Intel can manage to sell these into the market now dominated by Nvidia.

I get where you are going but I think there is more to it.
Click to expand...

Yea I agree. But I do think he may have lost his way a little though. Intel deal was decent but I do believe the rumors he was trying to spin RTG off, I wouldn't be surprised honestly if one of the reasons he got let go was because he was trying to push RTG to intel. I am very confident that is what happened. We will never know though. But I think after intel using vega was confirmed I started believing the rumors that he might have wanted AMD to sell of RTG. I think intel was probably highly interested.
 
Does anyone else find it a bit sad that we're talking more about company infighting, politics and gossip of engineers (both on the AMD and nV side) than... you know... actual technology?
 
harmattan said:
Does anyone else find it a bit sad that we're talking more about company infighting, politics and gossip of engineers (both on the AMD and nV side) than... you know... actual technology?
Click to expand...

Not that weird there is no new product (consumers) for this year and RTG has been shuffling some personnel around.
 
harmattan said:
Does anyone else find it a bit sad that we're talking more about company infighting, politics and gossip of engineers (both on the AMD and nV side) than... you know... actual technology?
Click to expand...

Takes years to make these things anymore, easy node shrinks are gone. Plus gossip and rumors can be interesting sometimes, you would be surprised how much infighting their can be at a company.
 
Gideon said:
Takes years to make these things anymore, easy node shrinks are gone. Plus gossip and rumors can be interesting sometimes, you would be surprised how much infighting their can be at a company.
Click to expand...
There is one up side at least everyone is tone deaf now towards exaggerated claims on the AMD side ;)
 
Pieter3dnow said:
There is one up side at least everyone is tone deaf now towards exaggerated claims on the AMD side ;)
Click to expand...

At least they didnt borrow that chiller from Intel just to show a overclock no normal user will ever be able to achieve. Intel over hypes their shit often they just dont get called out as often for it. How many years ago did they promise a easy 5GHz overclock and they just now got a cherry picked one that can boost to that. Intel and AMD have both been guilty of hyping something to a unrealistic expectation. I personally like the way Lisa Su is running the company, She has achieve a lot with AMD with far less resources then others had. I am guessing you will see quite a bit more focus at RTG and less hype with her at the helm now.
 
  • Like
Reactions: Zuul
like this
Gideon said:
At least they didnt borrow that chiller from Intel just to show a overclock no normal user will ever be able to achieve. Intel over hypes their shit often they just dont get called out as often for it. How many years ago did they promise a easy 5GHz overclock and they just now got a cherry picked one that can boost to that. Intel and AMD have both been guilty of hyping something to a unrealistic expectation. I personally like the way Lisa Su is running the company, She has achieve a lot with AMD with far less resources then others had. I am guessing you will see quite a bit more focus at RTG and less hype with her at the helm now.
Click to expand...

The funny thing is (more sad then funny) that all of those articles and youtube reports still are up this day with the 5ghz claim. If AMD had such a chiller and ran Cinebench it would have been murder :) .
 
Gideon said:
At least they didnt borrow that chiller from Intel just to show a overclock no normal user will ever be able to achieve. Intel over hypes their shit often they just dont get called out as often for it. How many years ago did they promise a easy 5GHz overclock and they just now got a cherry picked one that can boost to that. Intel and AMD have both been guilty of hyping something to a unrealistic expectation. I personally like the way Lisa Su is running the company, She has achieve a lot with AMD with far less resources then others had. I am guessing you will see quite a bit more focus at RTG and less hype with her at the helm now.
Click to expand...

intel have "5ghz overclocked" processor since 2011 with the i7 2700K and it was somewhat easy even on Air cooling.. AMD have "5ghz overclocked" processor since 2012 with the FX 8350.. 5000Mhz it's a great marketing number, which it's why both companies always call for it in desperate situations, if not ask to the 900$ FX 9590, which it's exactly what you are claiming about the i7 8086K and even worse, as at least was binned enough to keep the same 95W TDP..

intel it's actually in a damage control situation acting as if doesn't have AMD's dick in their ass but as Pieter3dnow intel can't be underestimated even if they are now in a bad position...
 
Araxie said:
intel have "5ghz overclocked" processor since 2011 with the i7 2700K and it was somewhat easy even on Air cooling.. AMD have "5ghz overclocked" processor since 2012 with the FX 8350.. 5000Mhz it's a great marketing number, which it's why both companies always call for it in desperate situations, if not ask to the 900$ FX 9590, which it's exactly what you are claiming about the i7 8086K and even worse, as at least was binned enough to keep the same 95W TDP..

intel it's actually in a damage control situation acting as if doesn't have AMD's dick in their ass but as Pieter3dnow intel can't be underestimated even if they are now in a bad position...
Click to expand...

Many 2700K did not make it to 5 GHz, some did tho and it was not as easy as Intel tried to imply. The 9590 was no doubt a desperation move by AMD basically a overclocked chip from the factory, but Intel was just as desperate with this 28 core display they put on. Intel is screwed for the next 3 years at least and they know it and why they went and got the man Jim Keller to build a new architecture. They know their current design will not scale to the same level as AMD's and with AMD about to move to 7nm it's only a uphill battle for Intel from there even if you want to say it's only equivalent to Intel 10nm that is still better then AMD has ever been able to do before. While I agree Intel should not be underestimated they are also in a bad position and they know it. AMD will have some breathing room now which is what they really need, takes time to rebuild a customer base in the corporate world. Id prefer AMD marketing right now as being on target rather then live in fantasy land and talking numbers or product it cant deliver. At least the cpu market is interesting again.
 
Gideon said:
Many 2700K did not make it to 5 GHz, some did tho and it was not as easy as Intel tried to imply. The 9590 was no doubt a desperation move by AMD basically a overclocked chip from the factory, but Intel was just as desperate with this 28 core display they put on. Intel is screwed for the next 3 years at least and they know it and why they went and got the man Jim Keller to build a new architecture. They know their current design will not scale to the same level as AMD's and with AMD about to move to 7nm it's only a uphill battle for Intel from there even if you want to say it's only equivalent to Intel 10nm that is still better then AMD has ever been able to do before. While I agree Intel should not be underestimated they are also in a bad position and they know it. AMD will have some breathing room now which is what they really need, takes time to rebuild a customer base in the corporate world. Id prefer AMD marketing right now as being on target rather then live in fantasy land and talking numbers or product it cant deliver. At least the cpu market is interesting again.
Click to expand...
Nonsense. Intel got caught with their pants down -- temporarily. But they could stomp AMD's current Threadripper offering with a current gen Xeon rebranded for Enthusiast level. It just wouldn't be the particular Xeon they false started with in this instance.

2013 news: Intel has 72 cores:
https://www.extremetech.com/extreme...ights-landing-cpu-for-exascale-supercomputing

Intel Phi type chips......You don't think Intel has stuff in the engineering labs that they could dust off in a generation or two and bring to market?

They wanted to show off the Mhz crown to keep the spotlight - but if you want to just do an all out torque war on cores - Intel still wins. Their prices are too high to be competitive with threadripper - but they can drop them as needed to be competitive.
 
Last edited:
Archaea said:
Nonsense. Intel got caught with their pants down -- temporarily. But they could stomp AMD's current Threadripper offering with a current gen Xeon rebranded for Enthusiast level. It just wouldn't be the particular Xeon they false started with in this instance.

2013 news: Intel has 72 cores:
https://www.extremetech.com/extreme...ights-landing-cpu-for-exascale-supercomputing

Intel Phi type chips......You don't think Intel has stuff in the engineering labs that they could dust off in a generation or two and bring to market?

They wanted to show off the Mhz crown to keep the spotlight - but if you want to just do an all out torque war on cores - Intel still wins. Their prices are too high to be competitive with threadripper - but they can drop them as needed to be competitive.
Click to expand...
If Intel knee-jerk released a 72 core processor for the HEDT crowd how could they price it reasonably to compete with Threadripper 2 without simultaneously killing demand and sales of their more expensive Xeon parts? Keep in mind they're already dropping that 28 core part to 1/4 of the original Xeon price. Yes they *can* drop prices but we don't know exactly how much it will hurt them to do so.
 
DedEmbryonicCe11 said:
If Intel knee-jerk released a 72 core processor for the HEDT crowd how could they price it reasonably to compete with Threadripper 2 without simultaneously killing demand and sales of their more expensive Xeon parts? Keep in mind they're already dropping that 28 core part to 1/4 of the original Xeon price. Yes they *can* drop prices but we don't know exactly how much it will hurt them to do so.
Click to expand...

Doesn't have to be a knee jerk - they should stop doing knee jerks and calmly rebrand. Nobody in the Server market is going to be upset if a chip they bought in 2013 is rebranded and sold for 1/4 or even 1/8 the price in 2018 -- even if it has some enhancements/tweaks. Intel could divide their product line into a gaming enthusiast line with faster, but fewer cores, vs an enthusiast workstation/productivity line for consumers with greater/slower cores. The work has already been done on the high end Xeon's and Phi chips from 5 years back!!! They just have to rebrand it --- No need for hail mary passes and bad press.

They also may have to face facts that the new market, competition driven, means that perhaps they won't get outrageous profit margins on their medium to high core count chips going forward.

I'm not saying Threadripper in 16 core and 32 core formats isn't a market changer or isn't disruptive -- it absolutely is. I'm just saying Threadripper isn't an Intel killer, and clearly doesn't have Intel's tech beat if productivity is the new challenge theme AMD is chasing.

Frankly, Intel still wins both ends of the spectrum. It's the consumer price spectrum they lose in. And since they've already done the research years ago - they could afford to lower prices ----- even if they don't want to --- and still retain the crowns with existing tech.
 
Last edited:
I was surprised I didn't see any wood screws from Intel at computex.
 
Last edited:
She is a very competent CEO for AMD. I have been enjoying her performance for some time now.
 
You must log in or register to reply here.
Back
Top