Intel is losing against AMD

Thank God Microcenter is a state away I probably would develop a disorder trying to shop there unhealthy PC parts binging with little or nothing to gain.

Ummm, what? Unhealthy PC parts? I don't get it.
 
Correct, and I'm confused by implications to the contrary. I have some familiarity with absolute beast of R&D that Intel has - do not underestimate them, or how things will go in the next couple years. I know Dr Su doesn't.

Intel's in a gunfight they have not experienced recently to be sure. A combination of corporate malaise and fab problems gave AMD a bay window opportunity to do exactly what they've done. They had a good product at a good price at a great time. They also have a CEO who is not only not clinically retarded, but actually one of the smartest people I've ever met. I'd admit I'm a huge fan, but also put forth this is based upon a lot of historical knowledge of her, not that I believe she's magically created Ryzen or something.

My adulation of Su is not meant to detract from any of AMD's victories and successes. Far from it. I'm thrilled they're in the game, and 50% of the machines my house run various Ryzens right now.

I am urging caution to the seemingly-unbridled exuberance and belief for some that Intel is hosed, or that they are somehow not pursuing responses with horrific intensity. I assure everyone - they are. As Dan_D has mentioned a few times, design cycles do not happen quickly. There will be competent responses. The only questions will be when, and what the rollout looks like.

I've been in the industry since the 90s, and both have ebbed and flowed. Both have come back, both have dominated at various times. This is normal business.

They Ebb and flow until they don't. You really can't use the past to predict much going forward.

Silicon Fabrication is ruthless. If you lose a step, you can get left behind. The amount of companies with Leading Edge Fabs keeps shrinking. AMD was one of those companies at one time. Then they spun them off as Global Foundries, the GF fumbled on the 7nm transition and they are out of the leading edge forever.

Intel had the process lead for decades, it could now be TSMCs turn for the next decade. I would not count on Intel regaining process superiority anytime soon, or ever.

Without that process lead, Intel is going have a much tougher competitive landscape against AMD, especially now that AMD is no longer shackled to GF for it's leading edge parts.

That's the process side. On the design side, it is useful to look at the core-for-core, clock-for-clock comparisons (contrary to what some argue). This gives us an idea of how the design performs when freed from process constraints.

Right now it looks like only some games have an edge for Intel with Coffee Lake vs Zen 2. But Intel does have Ice Lake which looks to tilt things back in favor of Intel, though it doesn't run on desktops.

On the design side I don't expect any big deltas to show up, like in the *Lake vs *Dozer days.

I don't see any big shifts coming. Things look better for AMD, because Intel has a hole to dig itself out of. It sounds like Intel desktop is stuck with Skylake 14nm derivatives until 2021 or more likely 2022. That is going to hurt more, the longer it continues.

Bottom line: We shouldn't count Intel out, but neither should count on Intel springing ahead anytime soon.
 
That was actually what I was trying to say, in my typical obtuse way. We're going to get competition. Nobody is dead.

The bottom line is more of a general comment about multiple posters arguments, some that basically say not only not to count Intel out, but as soon as Intel bounces back in a year or two, the best AMD can hope for is to hang on, relegated back to their proper second place status.
 
It sounds like Intel desktop is stuck with Skylake 14nm derivatives until 2021 or more likely 2022. That is going to hurt more, the longer it continues.

Not so much -- remember that most 'desktop' users have no need for more than four cores. A four core eight thread CPU at a decent price is overkill for desktop work, and it's probably the smallest die Intel is willing to build.
 
The bottom line is more of a general comment about multiple posters arguments, some that basically say not only not to count Intel out, but as soon as Intel bounces back in a year or two, the best AMD can hope for is to hang on, relegated back to their proper second place status.

With respect to volume, at the current rates of production AMD wouldn't overtake Intel in a decade.

With respect to performance, Intel is likely to leapfrog AMD.
 
That was actually what I was trying to say, in my typical obtuse way. We're going to get competition. Nobody is dead.

Normal folks know that no one is dead. Intel is losing against AMD though and they are not taking their pedal off the metal! :) This is not the AMD of the past and the Intel of the future has to deal with the deep, deep hole they dug themselves into, with quad core iterations adinfinium.
 
Normal folks know that no one is dead. Intel is losing against AMD though and they are not taking their pedal off the metal! :) This is not the AMD of the past and the Intel of the future has to deal with the deep, deep hole they dug themselves into, with quad core iterations adinfinium.

Intels biggest mistake was sitting at 4 cores forever... they have successfully convinced the masses that CPUs have been as fast as they will ever need for a long time now. That might be true for the most part... there was a time though that Intel did a good job of getting software vendors to make the case for hardware upgrades. There really isn't anything software wise today that won't run on a 10 year old Intel chip just fine... that is Intels fault.
 
Not so much -- remember that most 'desktop' users have no need for more than four cores. A four core eight thread CPU at a decent price is overkill for desktop work, and it's probably the smallest die Intel is willing to build.

I didn't see this one earlier. I agree that a 4c/8T is overkill for the majority of home users.

I was thinking more in terms of enthusiasts.

Enthusiast will buy more cores where they need them. Say for Gaming, or if they do a lot of Video encoding (after that it gets more niche).

But enthusiasts will also buy more cores because they want them, and they just like overkill, because they might need it, or find a use for it. I wouldn't be surprised if more enthusiasts sales are want based rather than need based.

Either way, it will hurt Intel with enthusiasts, as it is doing now, and will continue to do.

With Ryzen 1000 series, AMD made a big dent with Enthusiast aftermarket, but it was more of a mixed bag so they didn't pull ahead.
With Ryzen 2000 series, AMD made a massive dent with Enthusiast aftermarket, and AMD actually pulled ahead in aftermarket sales, because 2000 series got a bit more well rounded.
With Ryzen 3000 series, AMD made a took over Enthusiast aftermarket, because 3000 series shrunk Intel benefits to near insignificant, and delivered cores that some need, and even more cores that some just wanted.

So it only looks worse if Ryzen 4000 is even better, and Intel is still serving up Skylake on 14nm rehashes, and even worse if for Ryzen 5000, AMD is still facing 14nm Skylake refreshes....

During all this time, the inertia will be moving in enterprise and OEM contracts as well..

From what I have seen it will probably be 2022 before Intel moves desktop off 14nm, so then Intel will finally have a new core to compete with Zen 4? We can't assume Intel will leapfrog. AMD might have the process advantage, and AMD is not standing still in the interim either.

Once Intel digs out of the process hole, leaps don't seem so likely to me. I expect edging one way or the other.
 
Intels biggest mistake was sitting at 4 cores forever... they have successfully convinced the masses that CPUs have been as fast as they will ever need for a long time now. That might be true for the most part... there was a time though that Intel did a good job of getting software vendors to make the case for hardware upgrades. There really isn't anything software wise today that won't run on a 10 year old Intel chip just fine... that is Intels fault.

I would say that only the software from that era is going to run fine on that 10 year old quad core, in my opinion, for the most part. New stuff will tend to run slow in comparison.
 
I didn't see this one earlier. I agree that a 4c/8T is overkill for the majority of home users.

I was thinking more in terms of enthusiasts.

Enthusiast will buy more cores where they need them. Say for Gaming, or if they do a lot of Video encoding (after that it gets more niche).

But enthusiasts will also buy more cores because they want them, and they just like overkill, because they might need it, or find a use for it. I wouldn't be surprised if more enthusiasts sales are want based rather than need based.

Either way, it will hurt Intel with enthusiasts, as it is doing now, and will continue to do.

With Ryzen 1000 series, AMD made a big dent with Enthusiast aftermarket, but it was more of a mixed bag so they didn't pull ahead.
With Ryzen 2000 series, AMD made a massive dent with Enthusiast aftermarket, and AMD actually pulled ahead in aftermarket sales, because 2000 series got a bit more well rounded.
With Ryzen 3000 series, AMD made a took over Enthusiast aftermarket, because 3000 series shrunk Intel benefits to near insignificant, and delivered cores that some need, and even more cores that some just wanted.

So it only looks worse if Ryzen 4000 is even better, and Intel is still serving up Skylake on 14nm rehashes, and even worse if for Ryzen 5000, AMD is still facing 14nm Skylake refreshes....

During all this time, the inertia will be moving in enterprise and OEM contracts as well..

From what I have seen it will probably be 2022 before Intel moves desktop off 14nm, so then Intel will finally have a new core to compete with Zen 4? We can't assume Intel will leapfrog. AMD might have the process advantage, and AMD is not standing still in the interim either.

Once Intel digs out of the process hole, leaps don't seem so likely to me. I expect edging one way or the other.

A lot of what you typed here I can agree with. However, 4c 8t processors are no where near overkill for the majority of home users, since this is 2019 and not 2009.
 
This is not the AMD of the past and the Intel of the future has to deal with the deep, deep hole they dug themselves into, with quad core iterations adinfinium.

What hole?

However, 4c 8t processors are no where near overkill for the majority of home users, since this is 2019 and not 2009.

The majority of home users need a tablet with a keyboard.

A quad core Skylake or Zen will run circles around that.
 
A lot of what you typed here I can agree with. However, 4c 8t processors are no where near overkill for the majority of home users, since this is 2019 and not 2009.

The year doesn't matter. You don't need more than 4C, for Internet/media consumption and home office productivity.
 
I would say that only the software from that era is going to run fine on that 10 year old quad core, in my opinion, for the most part. New stuff will tend to run slow in comparison.

But it runs and that is the point. Sure perhaps it takes 20 seconds to load instead of a 2... but it runs. And most people will notice more uplift from a faster SSD then a new system if they have anything in the 4-5 year old range. There was a time where if you didn't have the latest hardware you couldn't run the latest software >.<

I mean as an example SW Jedi Fallen min requirements....

AMD FX-6100 release date 2011-10-12
Intel Core i3-3220 release date 2012-09-01

So the latest star wars games min requirement is a 8 year old CPU. Sure the recommended list hits 3-4 years out. Still if you told teenage me that games in 2019 would run on CPUs from 2010 I would have been very sad.

Those same processors handle more mundane things with ease. Software folks stopped pushing the envelope cause there are billions of machines in the world with no more then 4 cores. Anything with more then 4 cores is a niche part of the market from the past 2 years at most. Intels fault for sitting still. They almost deserve it for letting AMD make a come back. Intel VS AMD is like a boxing match where the blue team scored 2 knockdowns in everyone of the first 8 rounds... and after 2 technical knockouts in the 9th the blue team decided to start show boating and put on a foot work display. They almost deserve it if AMD knocks them right out of the server market over the next couple years. With even a half hearted push they should have put AMD under for the count 4-5 years ago. (not that that would have been good) before Lisa Su come on though AMD was one bad quarter from being toast.
 
But it runs and that is the point. Sure perhaps it takes 20 seconds to load instead of a 2... but it runs. And most people will notice more uplift from a faster SSD then a new system if they have anything in the 4-5 year old range. There was a time where if you didn't have the latest hardware you couldn't run the latest software >.<

I mean as an example SW Jedi Fallen min requirements....

AMD FX-6100 release date 2011-10-12
Intel Core i3-3220 release date 2012-09-01

So the latest star wars games min requirement is a 8 year old CPU. Sure the recommended list hits 3-4 years out. Still if you told teenage me that games in 2019 would run on CPUs from 2010 I would have been very sad.

Those same processors handle more mundane things with ease. Software folks stopped pushing the envelope cause there are billions of machines in the world with no more then 4 cores. Anything with more then 4 cores is a niche part of the market from the past 2 years at most. Intels fault for sitting still. They almost deserve it for letting AMD make a come back. Intel VS AMD is like a boxing match where the blue team scored 2 knockdowns in everyone of the first 8 rounds... and after 2 technical knockouts in the 9th the blue team decided to start show boating and put on a foot work display. They almost deserve it if AMD knocks them right out of the server market over the next couple years. With even a half hearted push they should have put AMD under for the count 4-5 years ago. (not that that would have been good) before Lisa Su come on though AMD was one bad quarter from being toast.

The fact is, quite a bit of stuff does not run on 10 year old machines, anymore. Also, people very much care about how quickly things happen on their computers or any other computing device they may own.
 
  • Like
Reactions: ChadD
like this
Intels biggest mistake was sitting at 4 cores forever... they have successfully convinced the masses that CPUs have been as fast as they will ever need for a long time now. That might be true for the most part... there was a time though that Intel did a good job of getting software vendors to make the case for hardware upgrades. There really isn't anything software wise today that won't run on a 10 year old Intel chip just fine... that is Intels fault.

everyone used to say "why should intel compete against itself"

now amd is competeing against itself and they are beating themselves to a pulp.

from 16 to 64 cores on HEDT and 8 to 16 on mainstream.

intel was just milking people.
 
It not only wouldn't have been 'good' for competition, but Intel would likely have fallen under regulatory scrutiny and been broken up.

If there ever was a case for breaking up intel, it would have happened three years ago before zen. But how would you do that? Cpu division East, west and central? Cpu divisions consumer and datacenter? Besides we have custom CPUs all over the place, arm..

Intel will never be broken up. Besides, they are selling off all their divisions themselves now so sort of “self break up”.
 
The fact is, quite a bit of stuff does not run on 10 year old machines, anymore. Also, people very much care about how quickly things happen on their computers or any other computing device they may own.

Most stuff runs. I am using a 10+ year old C2Q as my only machine. Works fine for everything I want it to do except modern games.
 
  • Like
Reactions: ChadD
like this
If there ever was a case for breaking up intel, it would have happened three years ago before zen. But how would you do that? Cpu division East, west and central? Cpu divisions consumer and datacenter? Besides we have custom CPUs all over the place, arm..

Intel will never be broken up. Besides, they are selling off all their divisions themselves now so sort of “self break up”.

Uh, spinning off the fans etc.? Graphics, CPU, chipsets and so on?

Why is this hard?
 
Uh, spinning off the fans etc.? Graphics, CPU, chipsets and so on?

Why is this hard?

Because the issue is the CPUs. Spinning off everything else does not address a cpu monopoly, now does it? It would be a pointless exercise.
 
It not only wouldn't have been 'good' for competition, but Intel would likely have fallen under regulatory scrutiny and been broken up.

For what selling CPUs. I'm not suggesting they should have sent a hit man over to the AMD offices. lol

My point is if Intels reaction to FX chips... was instead of just laughing in the profits with refreshes for 8+ years and instead, releasing a real 8 core Intel chip for high end mainstream sale. They would have crushed any investment AMD picked up to make Zen happen in the first place.

Regulation is a joke... and even if the Gov came knocking, Intel would rightfully have pointed at ARM and easily made the case that Intel and x86 isn't even #1 anymore.

I agree it would have been bad for consumers. But really intel took a victory lap for 6 or 7 years... I mean core counts didn't go up, and there are multiple Skylake vs Sandybridge showdowns online that show Intel gained what 10% in that time. I know Intel developed stuff and that isn't the case in Servers... and Intel has invested and has some interesting things going on in regards to ARM ect. Still they basically had their opponent on the ropes and said woo lets hold up and make some bank here. Good for their investors short term... but they miscalculated AMDs ability to rebound and it looks like they are going to pay long term.
 
Last edited:
The fact is, quite a bit of stuff does not run on 10 year old machines, anymore. Also, people very much care about how quickly things happen on their computers or any other computing device they may own.

As others have pointed out... you, I and the types of folks on these forums care. Average people don't even care if there running a PC anymore for the most part. Intel made ARM possible as well by not pushing Desktop hardware hard enough to make a desktop a requirement for much of anything. People have accepted their handheld computing devices... but I do wonder how many people would still have a desktop as well if it actually offered them a reason to have one. Over the last 10 or so years a lot of people that used to have desktops, simply never bothered replacing them. Why when there phone does everything they where using it for anyway.

Most gamers are happy with their consoles. (no there not better, just good enough for the masses)
Most regular folks are happy to web surf on their ARM powered phones.
Even artistic types are starting to use phones and tablets.

Perhaps Intel couldn't have done much about that... and frankly I'm not sure what would have happened in software with more powerful PCs. But we'll never know. I tend to think software developers would have found more compelling uses for more horsepower if it was widely available. Who knows perhaps I'm overstating and software developers would have ignored all the extra horsepower if even shitty dells where shipping with 6+ cores.
 
  • Like
Reactions: Mega6
like this
Uh, spinning off the fans etc.? Graphics, CPU, chipsets and so on?

Why is this hard?

Because the issue is the CPUs. Spinning off everything else does not address a cpu monopoly, now does it? It would be a pointless exercise.

Ya its not like Intel wasn't the dominate player when other companies where making Intel compatible chipsets. Really the US gov isn't going to break up their darling Intel. Might as well just ask the Chinese to please take over the world. Look what the US did a few years back forbiding the sale of Intel server parts to China. (which backfired horribly) Intel is as close to state sponsored as it gets in the US. Breaking up Intel would weaken the US tech position. Some plucky American startup or even AMD wouldn't be the winners of an Intel breakup... companies like Zhaoxin, Sunway, FeiTeng; or some other Chinese Gov backed player would be. Companies like Zhaoxin and Sunway frankly are already going to be players in 10 years even if Intel remains massive. Its only a matter of time before some of those Chinese made chips start showing up in consumer devices and perhaps even servers in the west. Even if the US and other western Gov pass laws to stop that.... the Chinese are already starting wiggle into other world wide markets that Intel (and AMD) rely on to hit their numbers. AMD has been wisely working with China to co develop stuff... Intel well I imagine the US Gov would put a stop to them doing the same.

Anyway long answer as always to make a short point. The US isn't breaking up Intel for nothing... even if AMD knuckled under tomorrow.
 
  • Like
Reactions: Mega6
like this
As others have pointed out... you, I and the types of folks on these forums care. Average people don't even care if there running a PC anymore for the most part. Intel made ARM possible as well by not pushing Desktop hardware hard enough to make a desktop a requirement for much of anything. People have accepted their handheld computing devices... but I do wonder how many people would still have a desktop as well if it actually offered them a reason to have one. Over the last 10 or so years a lot of people that used to have desktops, simply never bothered replacing them. Why when there phone does everything they where using it for anyway.

Most gamers are happy with their consoles. (no there not better, just good enough for the masses)
Most regular folks are happy to web surf on their ARM powered phones.
Even artistic types are starting to use phones and tablets.

Perhaps Intel couldn't have done much about that... and frankly I'm not sure what would have happened in software with more powerful PCs. But we'll never know. I tend to think software developers would have found more compelling uses for more horsepower if it was widely available. Who knows perhaps I'm overstating and software developers would have ignored all the extra horsepower if even shitty dells where shipping with 6+ cores.

absolutely. No one wants to be tied to an x86/64 desktop. Convenience rules. Wasted opportunity? Maybe.

40% black Friday sales by online phone.
 
  • Like
Reactions: ChadD
like this
absolutely. No one wants to be tied to an x86/64 desktop. Convenience rules. Wasted opportunity? Maybe.

40% black Friday sales by online phone.

Not sure if that article is scary funny or just amazing.

Its a different world ya... I feel sometimes our geeky back and forth about PC parts. Is analogous to record geeks arguing Vinyl and record player needle quality in 1995. Where not wrong but the world just doesn't care. :)
 
  • Like
Reactions: Mega6
like this
Ya its not like Intel wasn't the dominate player when other companies where making Intel compatible chipsets. Really the US gov isn't going to break up their darling Intel. Might as well just ask the Chinese to please take over the world. Look what the US did a few years back forbiding the sale of Intel server parts to China. (which backfired horribly) Intel is as close to state sponsored as it gets in the US. Breaking up Intel would weaken the US tech position. Some plucky American startup or even AMD wouldn't be the winners of an Intel breakup... companies like Zhaoxin, Sunway, FeiTeng; or some other Chinese Gov backed player would be. Companies like Zhaoxin and Sunway frankly are already going to be players in 10 years even if Intel remains massive. Its only a matter of time before some of those Chinese made chips start showing up in consumer devices and perhaps even servers in the west. Even if the US and other western Gov pass laws to stop that.... the Chinese are already starting wiggle into other world wide markets that Intel (and AMD) rely on to hit their numbers. AMD has been wisely working with China to co develop stuff... Intel well I imagine the US Gov would put a stop to them doing the same.

Anyway long answer as always to make a short point. The US isn't breaking up Intel for nothing... even if AMD knuckled under tomorrow.

Even a greedy monopolistic acting intel is good for America, until it hurts AMD. The military is having a hell of a time weeding Chinese chips out of military systems in the interest of national security. And the joint Amd / china venture has been blocked/shut down by US because of tech transfer fears. The Chinese will tell you straight up - they want to rule the world. Huh, who said that before last? CCCP.
 
  • Like
Reactions: ChadD
like this
Not so much -- remember that most 'desktop' users have no need for more than four cores. A four core eight thread CPU at a decent price is overkill for desktop work, and it's probably the smallest die Intel is willing to build.

I think you need to extrapolate more on "most users". Entry level mainstream gamers need a 4/8 at the least today but everyone beyond trying to just budget game and try get 120FPS is needing little more than an athlon 200G or g series apu which is ample for any other need including admin style work or granny emailing family. The new Athlon is dirt cheap for AMD to turn profit on.

If Intel only do 4/8 around 200$, they are in big trouble
 
Even a greedy monopolistic acting intel is good for America, until it hurts AMD. The military is having a hell of a time weeding Chinese chips out of military systems in the interest of national security. And the joint Amd / china venture has been blocked/shut down by US because of tech transfer fears. The Chinese will tell you straight up - they want to rule the world. Huh, who said that before last? CCCP.

Long term China is the threat not Intel to AMD or AMD to Intel.

The problem with blocking China is they just fire up the home grown stuff.

The US under Obama (only mention to point out the time frame) stopped Intel server chip sales to China... and it backfired hard. That is when China poured billions into chip R&D and sunway built the then fastest Super computer in the world. Couldn't buy Intel fine we'll drop a couple 100 billion and cut them out, and they did.

The most hilarious part of china's rise though is that most of it is fueled by US money. The amount of American venture capital investment in the Chinese tech sector is obscene. US venture capital spends about equal money in China as they do in the US right now.
 
  • Like
Reactions: N4CR
like this
Long term China is the threat not Intel to AMD or AMD to Intel.

The problem with blocking China is they just fire up the home grown stuff.

The US under Obama (only mention to point out the time frame) stopped Intel server chip sales to China... and it backfired hard. That is when China poured billions into chip R&D and sunway built the then fastest Super computer in the world. Couldn't buy Intel fine we'll drop a couple 100 billion and cut them out, and they did.

The most hilarious part of china's rise though is that most of it is fueled by US money. The amount of American venture capital investment in the Chinese tech sector is obscene. US venture capital spends about equal money in China as they do in the US right now.
China can fire up the homegrown but it’s still only a pentium level equivalent. brute force is how they get it done. Let’s face it- a billion population that graduates 8 stem college degrees to 1 in USA. It’s only a matter of time before they copy everything. Plus the unfair trade practices that force tech transfer to China to do business there. Damn right we are funding china’s rise via venture capital, washing machines, refrigerators..
 
Last edited:
I really don't. I said most users because I meant most users.

I kind of broke it down for you anyways. for anyone not using the computer for 120FPS gaming $50 -$120 all in ones are plenty value to do anything you want. APU's are playing esport titles at 1080P these days rather easily. I am hoping Renoir addresses the iGPU limitations of Vega, basically it is not robust enough for the CPU performance now.
 
I kind of broke it down for you anyways. for anyone not using the computer for 120FPS gaming $50 -$120 all in ones are plenty value to do anything you want. APU's are playing esport titles at 1080P these days rather easily. I am hoping Renoir addresses the iGPU limitations of Vega, basically it is not robust enough for the CPU performance now.
Renoir is Zen2 and Vega 10.
 
Last edited:
The year doesn't matter. You don't need more than 4C, for Internet/media consumption and home office productivity.
People said the same about dual cores...
You have ancient C2Q so you should know this, no one who keeps their PC for 3+ years is going to buy a quad core right now unless they want a dirt cheap computer, or a secondary PC.
 
no one who keeps their PC for 3+ years is going to buy a quad core right now unless they want a dirt cheap computer

This is one reason why I am so frustrated with AMD and their APUs. I don't want a quad core anything in 2019.

Thankfully the wait should be over in Q1 2020.
 
People said the same about dual cores...
You have ancient C2Q so you should know this, no one who keeps their PC for 3+ years is going to buy a quad core right now unless they want a dirt cheap computer, or a secondary PC.

There is a difference between what people buy, and what people need.

My argument is that MOST people don't need more than a modern 4c/8t machine today.

Show me some evidence that MOST people need more than this today.
 
Back
Top