"Intel in jeopardy" from insiders. Much much worse than what they tell.

Jandor

Gawd
Joined
Dec 30, 2018
Messages
665
Looks like there's no future for Intel.
Just read :

Looks like Intel would completely lose their market share everywhere without any other hope like GPU. That's no businness at all for 3 years. They may have some hope for 2023 but AMD will be much much more advanced. They may only get to AMD 2021 level in 2023, if everything works well.Not sure about that because they are lying and unable to deliver for years. They may be back in 2025, meaning at best competitive, if all goes well.
After reading this, it looks very real, puts everything together so one can understand what's happening. Means if you have money, don't put it in Intel. Intel is totally overpriced today. Not necessarily look at AMD as a replacement. Look how well Ampere is doing and put it in Nvidia (not for gaming but for the server market and self driving cars), if you feel the need to invest in that business. Also TSMC, Samsung have bright future, not Intel.
Also Apple has a great ARM architecture that may very well go into servers. So this could be the end of life of x86 as a global architecture. Because no competition between AMD and Intel is in no way good even for AMD. There has never been so much technological distance between Intel and AMD than now between AMD and Intel and this distance will grow until 2023 when Intel will at best keep that distance steady.
 
Last edited:
Charlie has been damn accurate the last few years while people laughed. That's definitely not good news for Intel as it's probably mostly correct. 10nm is obviously shit as shown by lack of desktop product and 'it's coming' for years...
I'll do some digging on DCG, know someone pretty high up.

I didn't think Intel was in that much trouble but if Charlie says that, then shit is 80-90% not going well at all.
 
If true how could of Intel fuck up so bad? Did they completely believe that AMD couldn't make a decent cpu again? All of Intel management needs to be fired.
 
Intel's server market looks worse than thought.

... Cooper Lake pushed to H2 2020
... Now 300w instead of 250w
... Intel "use water cooling" 6 months before launch
... OEMs "Fuck your face"

Paraphrasing a little
 
Last edited:
It's really going to get real for Intel once the OEMs and big volume customers start switching to AMD.

I sold my AMD stock early this year with over 100% gain. Crazy to think how I could have tripled or more if I held on longer.
 
This isn't Star Wars, some low percentage moment isn't going to topple the Empire on a coin flip event.
When your issued MacBook Pro is all AMD, and I have to specifically deploy Intel homed AMIs on public cloud then Intel is in trouble.

Otherwise a blip in the consumer market is nothing.
It's like assuming Toyota passenger truck sales affect Volvo and Peterbilt commercial.
 
It's really going to get real for Intel once the OEMs and big volume customers start switching to AMD.

I sold my AMD stock early this year with over 100% gain. Crazy to think how I could have tripled or more if I held on longer.

Yeah people don't realize that the effects of the problem will take years to accumulate. It's not like the entire server market is going to flip tomorrow, but as companies build there instructures through 2023, we will see how drastic of a flip there was at that point.

10nm was really a double whammy for Intel. First their was the yield issues that cost them years. Then came the performance issues that are so bad that it takes a new architecture to balance out what was lost. They basically lost 5+ years because of the mess.

They went from being years ahead of AMD to years behind.
 
This isn't Star Wars, some low percentage moment isn't going to topple the Empire on a coin flip event.
When your issued MacBook Pro is all AMD, and I have to specifically deploy Intel homed AMIs on public cloud then Intel is in trouble.

Otherwise a blip in the consumer market is nothing.
It's like assuming Toyota passenger truck sales affect Volvo and Peterbilt commercial.

The biggest deal with respect to marketshare at this point in time isn't performance (which affects price) but rather production capacity. AMD relies on third-party fabs, primarily TSMC, which has many greedy customers for its latest fabrication nodes.

Intel, on the other had, has to figure out how to continue to produce 14nm in volume because it works, while fixing 10nm for volume, if that's possible, while developing 7nm. The comment about farming out to Samsung makes sense, as Intel usually uses their oldest nodes for all of the other products they produce for themselves like chipsets and modems and networking ASICs and so on for others.

So yeah, while prices are going to fluctuate a bit, the biggest issue is what each company can actually deliver to their system integrators and retail customers. Demand is rising while both companies are in rolling transition phases, Intel with their fab struggles and AMD with record demand and without question a clear awareness that they must continue to meaningfully iterate on their current winning architectures, as while Intel cannot yet manufacture the designs that their engineers come up with, at some point in the future they most certainly will.


Biggest takeaway: AMD becomes credible, established competition across Intel's CPU product stack. Intel's fumble has created a wide opening for AMD to get their products into industry infrastructure and to continue to build on the massive mindshare Zen has built so far.
 
Intel's server market looks worse than thought.

... Cooper Lake pushed to H2 2020
... Now 300w instead of 250w
... Intel "use water cooling" 6 months before launch
... OEMs "Fuck you face"

Paraphrasing a little

The server market is both where Intel is in the most trouble, and at the same time, the one that will be slowest to change due to inertia. The question in my mind is just how much of a foothold AMD will gain; given that the market is expanding, where AMD loses sales to those organizations upgrading that are more resistant to change, just having the new hotness available to purchase in quantity makes them more attractive for new installations.

The longer Intel takes to get volume and performance back up, the more ground AMD will gain; I have an inkling that AMD won't gain as much as they rightly should given the quality of their products, but, I do expect them to have enough of a foothold to be able to maintain their presence and provide credible competition for the next decade at least.
 
The server market is both where Intel is in the most trouble, and at the same time, the one that will be slowest to change due to inertia. The question in my mind is just how much of a foothold AMD will gain; given that the market is expanding, where AMD loses sales to those organizations upgrading that are more resistant to change, just having the new hotness available to purchase in quantity makes them more attractive for new installations.

The longer Intel takes to get volume and performance back up, the more ground AMD will gain; I have an inkling that AMD won't gain as much as they rightly should given the quality of their products, but, I do expect them to have enough of a foothold to be able to maintain their presence and provide credible competition for the next decade at least.
Main issue for AMD isn't volume for datacentre, that's fine so far, it's the ecosystem and migration/software support which of course is rapidly shifting but it does take time.
For Intel, aside from volume which is highly constrained so far, thermals and inefficiency in high core applications will really hurt them - water cooling is a prick to implement for a datacentre that is air cooled. I think that will be a real pivot point moving forward and that's even if they can get the volume they need too which they can't so far.

Desktop they are hanging on by a few fingers.
Laptops we will see what happens Q1 2020. They're hemorrhaging $ subsiding stuff there already and shareholders will only tolerate that so much.
 
11 years of quad cores.
Yes but it seems like Intel put zero money into R&D over the past decade. I would though they would still put some effort into it. I wonder if Nvidia is doing the same thing now. They basically only a step ahead of AMD atm. I always though Nvidia had something on th.shelve ready to go to crush anything AMD r leases but I am doubting that now.
 
The biggest deal with respect to marketshare at this point in time isn't performance (which affects price) but rather production capacity. AMD relies on third-party fabs, primarily TSMC, which has many greedy customers for its latest fabrication nodes.

Intel, on the other had, has to figure out how to continue to produce 14nm in volume because it works, while fixing 10nm for volume, if that's possible, while developing 7nm. The comment about farming out to Samsung makes sense, as Intel usually uses their oldest nodes for all of the other products they produce for themselves like chipsets and modems and networking ASICs and so on for others.

So yeah, while prices are going to fluctuate a bit, the biggest issue is what each company can actually deliver to their system integrators and retail customers. Demand is rising while both companies are in rolling transition phases, Intel with their fab struggles and AMD with record demand and without question a clear awareness that they must continue to meaningfully iterate on their current winning architectures, as while Intel cannot yet manufacture the designs that their engineers come up with, at some point in the future they most certainly will.


Biggest takeaway: AMD becomes credible, established competition across Intel's CPU product stack. Intel's fumble has created a wide opening for AMD to get their products into industry infrastructure and to continue to build on the massive mindshare Zen has built so far.

It’s easy to stay agile but smaller, but once you have to focus on long term contractual enterprise delivery your feet are cemented down.
You don’t get to stray outside your lane in the enterprise.
Amd should stay where they are now, and Intel should cement down those 10+ year deliverables.
 
Yes but it seems like Intel put zero money into R&D over the past decade. I would though they would still put some effort into it. I wonder if Nvidia is doing the same thing now. They basically only a step ahead of AMD atm. I always though Nvidia had something on th.shelve ready to go to crush anything AMD r leases but I am doubting that now.

Intel put more money into R&D in the last decade than AMD has ever made -- but they also committed to their worst mistake since Netburst.
 
Yes but it seems like Intel put zero money into R&D over the past decade. I would though they would still put some effort into it. I wonder if Nvidia is doing the same thing now. They basically only a step ahead of AMD atm. I always though Nvidia had something on th.shelve ready to go to crush anything AMD r leases but I am doubting that now.
Well they have and it's called Ampere, and it will deliver much better performance. AMD too has RDNA2 but I kind of think Ampere will be rather ahead. They also managed to create competition between TSMC and Samsung to get the better price which AMD seems to be out of time to organize.
 
Before reading this article, I would have considered a Rocket Lake next year (maybe in December 2020). Of course this makes me pause but I wonder if AMD will simply raise prices.
 
I've been pricing out Dell servers lately. I can get 24-core AMD based servers cheaper than I can get 8-core Intel based servers. Other specs are pretty much the same.

Only problem with that is.. I can't really go over 16-cores per server as the Microsoft Server licencing goes sky-high once you go over 16-cores.

So I will probably end up getting lower core count AMD servers.

I still think that this is their comeuppance for their dirty and illegal tactics during the Athlon 64 time frame when they tried to block AMD out of the market.
 
I think AMD got a bit lucky when they were essentially forced to go for a chiplet design in the face of Intel's heavily financed monolithic 10nm.

Or I spose you could argue they simply innovated in a highly competitive market, and Intel just simply didn't.
 
Intel put more money into R&D in the last decade than AMD has ever made -- but they also committed to their worst mistake since Netburst.
Yes but they kind of invest to show shareholders they invest, not to move ahead and win. So this is how AMD got ahead with little investment. Also one has to take into account that some of the major investment is at the founder level, TSMC.
 
Yes but they kind of invest to show shareholders they invest, not to move ahead and win.

Intel romeo-foxed their 10nm release schedule, yet still managed to continue to innovate and push 14nm further than anyone had thought possible. When I bought my 6700k, I never thought I'd be able to get anything close to that in a laptop short of a DTR boat anchor, let alone the ~2lb ultrabook I've had for almost two years.

AMD still isn't there yet despite a clear process advantage, more recent architecture, more recent security research (and recent research overall) to base that architecture on.

And I agree, by using TSMC, AMD is leaning on on TSMCs customer base (Apple etc.) for their fab advances, and this round, that has worked out very well for them.
 
Some people say that there is no such thing as a coincidence. I am of the same thought. AMD made some very smart decisions after the Bulldozer debacle. Infinity Fabric was one of them. The new scalable Ryzen Architecture is another. The Ryzen to Ryzen2 improvements are close to perfect. Teaming with TSMC and perfect transition from 14nm to 7nm to 7nm FF+ and soon 5nm.

Luck? No, Intel sat on their ass and created what we call a window of opportunity. Intel spent about the same amount on R&D last year as stock buy backs. Billions that could have been used for doubling the R&D budget. For a company that has fallen behind, they sure don't act like it.
 
Infinity Fabric was one of them.

...is a bus. Giving it a fancy name is an example of marketing, not engineering.

The new scalable Ryzen Architecture is another.

It's okay. It's still years behind Ice Lake.

The Ryzen to Ryzen2 improvements are close to perfect.

This is mostly down to debugging Zen / Zen+, and upon implementing chiplets and 7nm, being able to afford copious amounts of cache to cover for the disadvantages of their architecture.

Teaming with TSMC and perfect transition from 14nm to 7nm to 7nm FF+ and soon 5nm.

This is really a stroke of luck. TSMC is a coin flip away from fumbling a process family; they're famous for it, and it's bitten their customers in the ass before. AMD having something worth making and a fab to make it at really is a coincidence.

Luck? No, Intel sat on their ass and created what we call a window of opportunity.

Intel's fabs have caused all kinds of trouble, and given their history, a four-year slip really is 'lucky' for AMD. Had Intel's fabs not slipped, Zen would have looked a lot like Bulldozer on release. Zen 2 would be just 'okay'.

And yeah, reaching Skylake IPC but not clockspeeds with a node advantage is 'just okay'.
 
...is a bus. Giving it a fancy name is an example of marketing, not engineering.



It's okay. It's still years behind Ice Lake.



This is mostly down to debugging Zen / Zen+, and upon implementing chiplets and 7nm, being able to afford copious amounts of cache to cover for the disadvantages of their architecture.



This is really a stroke of luck. TSMC is a coin flip away from fumbling a process family; they're famous for it, and it's bitten their customers in the ass before. AMD having something worth making and a fab to make it at really is a coincidence.



Intel's fabs have caused all kinds of trouble, and given their history, a four-year slip really is 'lucky' for AMD. Had Intel's fabs not slipped, Zen would have looked a lot like Bulldozer on release. Zen 2 would be just 'okay'.

And yeah, reaching Skylake IPC but not clockspeeds with a node advantage is 'just okay'.


Really there is no substance here as usual. Relying on groundless inflammatory comments in hope of provoking a response is pathetic.
 
Really there is no substance here as usual. Relying on groundless inflammatory comments in hope of provoking a response is pathetic.
Fanboy at work again. You can't really think Ryzen is all that much better than Skylake refreshes. Clock for clock they are within error most of the time. AMD just caught up to a five year old architecture. I am glad they did, but Ryzen is not an Intel beater without the massive cache and core count, thus is still situatioinal advantagious. Ice Lake is showing massive IPC advantages over current tech, but it's vaporware ATM.

Where is your substance? If you are provoked by that post, it's a problem with your brain, not our remarks.
 
Fanboy at work again. You can't really think Ryzen is all that much better than Skylake refreshes. Clock for clock they are within error most of the time. AMD just caught up to a five year old architecture. I am glad they did, but Ryzen is not an Intel beater without the massive cache and core count, thus is still situatioinal advantagious. Ice Lake is showing massive IPC advantages over current tech, but it's vaporware ATM.

Where is your substance? If you are provoked by that post, it's a problem with your brain, not our remarks.

The sales and popularity speaks for itself. Funny how the true Fanbois whip that flag out all the time. The only live thread in the Intel section is over how dead Intel really is and whip out the Fanboi flag in defense - because that's all they have. Try looking over at the AMD section where people are building boxes.
 
The sales and popularity speaks for itself. Funny how the true Fanbois whip that flag out all the time.
Ryzen is certainly a better value, in multithread for sure, but you are dodging the point. AMD just caught up, they are not beating Intel in performance architecture, yet. I am not a Intel fanboy at all, as I have said. I also hope AMD continues to improve the core architecture, not just adding cores. I hope they pull another Athlon 64 innovation on Intel, and force the end of stagnation of CPU design we have suffered. I am just not a delusional AMD fanbot either.
 
Last edited:
Funny how AMD flopped for a decade with no viable CPU against Intel and still survive... Now Intel flopped for a year or 2, maybe 4 if they really messed up their "next gen" and everybody is telling stories how "It is the end for Intel". LOL

I'm glad that AMD finally put some pressure on Intel since we're now getting 8+ cores in the mainstream. However, if AMD become complacent (remember Netburst vs K8 and then Conroe vs Phenom?), intel will crush them again with their (dirty) tactics and money.

And where most of the money is made (enterprise), the decision maker is usually the business guy, not the technical guy. This is simply a case of "nobody ever got fired for buying Cisco network equipment" with Intel being Cisco here.
 
Last edited:
Funny how AMD flopped for a decade with no viable CPU against Intel and still survive... Now Intel flopped for a year or 2, maybe 4 if they really messed up their "next gen" and everybody is telling stories how "It is the end for Intel". LOL
What's really funny is how people think OEMs will stop using Intel, and that Intel only sells consumer grade CPUs, or only CPUs.
 
Fanboy at work again. You can't really think Ryzen is all that much better than Skylake refreshes. Clock for clock they are within error most of the time. AMD just caught up to a five year old architecture. I am glad they did, but Ryzen is not an Intel beater without the massive cache and core count, thus is still situatioinal advantagious. Ice Lake is showing massive IPC advantages over current tech, but it's vaporware ATM.

Where is your substance? If you are provoked by that post, it's a problem with your brain, not our remarks.

That's the only thing that bugs me about your post; saying "Ryzen is not an Intel beater without the massive cache and core count" to me would be like saying "Intel wouldn't have beat AMD if it weren't for higher IPC and the mhz advantage and etc etc" or "the only reason a V8 has more power than a straight 4 is because of the extra cylinders. Uhh yeah, that's kinda the point. You're right, AMD wouldn't be competitive with Intel if it weren't for the cache and core count but that's part of the architecture that has put AMD in the place it is. The cache and cores is also why AMD has seemed to show such relatively large performance gains when paired with faster tighter memory and Intel CPU's don't care as much about memory speed. You could say Intel's IMC is better than AMD's since it doesn't care about memory speed but you could also say the AMD CPU is so hungry for data to process (and therefore the Intel isn't) that we need faster memory to keep it fed. I guess it comes down to how you want to spin the situation. BTW I'm also being a bit of a devil's advocate here so take my post with a grain of salt, so to speak.

Edit: Not trying to pick a fight, just that it bug's me when someone says "the only reason X is ahead/better than Y is because they did Z different". Doesn't matter why they're competitive/ahead; fact of the matter is they did something that worked and that's that. And don't say "well if Intel hadn't stagnated then AMD wouldn't be competitive". That's not productive conversation because the fact of the matter is Intel f*cked up during the same time that AMD made some good decisions and came up with an architecture that is working for them with a seemingly good future and that's that. (y)
 
Last edited:
Edit: Not trying to pick a fight, just that it bug's me when someone says "the only reason X is ahead/better than Y is because they did Z different". Doesn't matter why they're competitive/ahead; fact of the matter is they did something that worked and that's that. And don't say "well if Intel hadn't stagnated then AMD wouldn't be competitive". That's not productive conversation because the fact of the matter is Intel f*cked up during the same time that AMD made some good decisions and came up with an architecture that is working for them with a seemingly good future and that's that. (y)

To respond in a similar light, the issue with adding cache --> winning is that that's a one-time deal. AMD cannot add more cache and get more performance, and Zen 2 looks like 'peak Zen'. Now they have to go back to the drawing board and actually innovate for more IPC, like Intel has already done with Ice Lake.

On the one hand, kudos to AMD for being ready for an opening from Intel, and on the other, AMD has to continue to execute with Nvidia-like precision if they're going to keep pace with Intel.
 
Funny how AMD flopped for a decade with no viable CPU against Intel and still survive... Now Intel flopped for a year or 2, maybe 4 if they really messed up their "next gen" and everybody is telling stories how "It is the end for Intel". LOL

I'm glad that AMD finally put some pressure on Intel since we're now getting 8+ cores in the mainstream. However, if AMD become complacent (remember Netburst vs K8 and then Conroe vs Phenom?), intel will crush them again with their (dirty) tactics and money.

And where most of the money is made (enterprise), the decision maker is usually the business guy, not the technical guy. This is simply a case of "nobody ever got fired for buying Cisco network equipment" with Intel being Cisco here.

Yeah I'm just happy we have some good competition now so prices will drop. Much needed in the GPU space!
 
That's the only thing that bugs me about your post; saying "Ryzen is not an Intel beater without the massive cache and core count" to me would be like saying "Intel wouldn't have beat AMD if it weren't for higher IPC and etc etc" or "the only reason a V8 has more power than a straight 4 is because of the extra cylinders. Uhh yeah, that's kinda the point. You're right, AMD wouldn't be competitive with Intel if it weren't for the cache and core count but that's part of the architecture that has put AMD in the place it is. The cache and cores is also why AMD has seemed to show such relatively large performance gains when paired with faster tighter memory and Intel CPU's don't care as much about memory speed. You could say Intel's IMC is better than AMD's since it doesn't care about memory speed but you could also say the AMD CPU is so hungry for data to process (and therefore the Intel isn't) that we need faster memory to keep it fed. I guess it comes down to how you want to spin the situation. BTW I'm also being a bit of a devil's advocate here so take my post with a grain of salt, so to speak.

Edit: Not trying to pick a fight, just that it bug's me when someone says "the only reason X is ahead/better than Y is because they did Z different". Doesn't matter why they're competitive/ahead; fact of the matter is they did something that worked and that's that. And don't say "well if Intel hadn't stagnated then AMD wouldn't be competitive". That's not productive conversation because the fact of the matter is Intel f*cked up during the same time that AMD made some good decisions and came up with an architecture that is working for them with a seemingly good future and that's that. (y)
Well my point is, core to core, clock for clock, AMD is about equal now, not surpassing Intel, which the post I waas quoting from the fanboy was alluding. Sure, you can say "AMD still does better chip for chip", but you can just run more Intel cores as well, if you want to waste your money. The true test of the architecture's performance is single core IPC, the test you are refering to is scalability. Scalability is important too, jsut not as indicative of advancment. There are lots of ways to get to the finish line first, but what you are driving and what track you are on will be the largest factors (driver skill bieng equal). Similarly, a I4 can produce more power than a V8, it all depends on how much air and fuel it is pushing, and what load it has to pull. The design of the motor is the contributing factor to how efficient it can achieve the goal you set for it. So if your workload will benefit from core/ thread count, then AMD is going to be the value bar none. But if you say want the best framerate in games at 1080p like me, and do nothing else, the lowly 6c/6t 9600K has slightly more value even at stock.

AMD has improved IPC to match Intel, as well as developed a platform for easy scalability on the package. But again, Intel has done nothing since Skylake, and not much more than that since Sandybridge, to improve IPC outside of specific instrustion sets. You get DDR4 and some iGPU advances, but that's about it. More cores and cache is the route they took as well, just before AMD did it, and AMD beat them at that game due to a better node and package design. But in the end, if power and money are not an issue, you can throw more consumer grade Intel cores in a machine, and beat a machine with the same number of AMD cores. Not a good idea, but it can be done.

Now the new Epyc and TR parts are in a class by themselves due to that afformentioned scalabilty reaching it's peak, Intel has nothing to match AMD there, certainly not within a single deployable machine.
 
There are leaks, that Intel is ready to launch Comet Lake. Another Skylake version with 2 more cores on a new socket LGA 1200. Seems like the speed is a little under what was expected so this is new bad news.
Fact is nobody may expect for LGA1200 to live longer than AM4, and eventually will have a shorter life since AM4 is expected to support Zen 3 in 2020 and 2021. 10 core will need 140 TDP support on the motherboard.
In fact the launch of Comet Lake for Desktop looks like as awkward as Skylake X HEDT. Hope they drop their price since one still won't have ECC RAM would need to go for 490 chipset for Raid feature. All this is offered by AMD on all their CPU line (and APU from the Pro line). Not a big deal for Intel to put those features ON on all their products. Intel is so unattractive. Not sure someone at Intel can see that. One would have to really love Intel to go their way.
 
  • Like
Reactions: N4CR
like this
There are leaks, that Intel is ready to launch Comet Lake. Another Skylake version with 2 more cores on a new socket LGA 1200. Seems like the speed is a little under what was expected so this is new bad news.
Fact is nobody may expect for LGA1200 to live longer than AM4, and eventually will have a shorter life since AM4 is expected to support Zen 3 in 2020 and 2021. 10 core will need 140 TDP support on the motherboard.
In fact the launch of Comet Lake for Desktop looks like as awkward as Skylake X HEDT. Hope they drop their price since one still won't have ECC RAM would need to go for 490 chipset for Raid feature. All this is offered by AMD on all their CPU line (and APU from the Pro line). Not a big deal for Intel to put those features ON on all their products. Intel is so unattractive. Not sure someone at Intel can see that. One would have to really love Intel to go their way.
Not everyone cares about RAID. Intel still games well. But beyond that, AMD has the market, if not the name. So Intel can keep putting out rehashed silicon and it will sell.
 
I think it's possible maybe even likely that AMD switches positions with Intel for a few years. But even that statement is a drastic oversimplification.

Intel is entrenched in OEM and Server markets and both of those groups are slow at best to change.
On the enthusiast end, it's possible or even likely that AMD will continue to take 90% of purchases. But that's an infinitesimally small part of the pie in comparison to OEM and Server.

Intel just has to weather the storm. They will have the worse part for a few years. It's not like this has never been the case before. I've lived through P4 Netburst architecture. Even "if" it takes until 2025, they have plenty of cash and can continue funding themselves for years to come. And it's also not like they'll be making zero money just because they aren't dominant. It just means "less" income. Similar to what AMD has been doing all these years.
Making "less" won't immediately doom them and thinking it will is ignoring more or less every competitive market that exists and instead just spreading FUD and doom and gloom.

To respond in a similar light, the issue with adding cache --> winning is that that's a one-time deal. AMD cannot add more cache and get more performance, and Zen 2 looks like 'peak Zen'. Now they have to go back to the drawing board and actually innovate for more IPC, like Intel has already done with Ice Lake.

On the one hand, kudos to AMD for being ready for an opening from Intel, and on the other, AMD has to continue to execute with Nvidia-like precision if they're going to keep pace with Intel.

Current info says that Zen 3 will offer a 17% increased IPC gain over Zen 2, which was already a 13% IPC gain from Zen.
https://www.thefpsreview.com/2019/1...l-reportedly-offer-an-average-ipc-gain-of-17/
I would also argue that Intel hasn't been pushing or innovating for roughly 7 years. Their track record is right there. This has been the subject of conversation for sometime. It hasn't been worth it for basically any quad core i7 owner since 3000 series CPU's to upgrade until Hex and Quad core's came out. Which only came out due to AMD pushing them to and only offered gains due to core count not increased IPC per core.

Still, to Intel's "credit" if you want to call it that, they were so dominant 10 years ago they didn't need to really compete all this time as they haven't had anyone to compete with. It took basically 10 years of Intel doing nothing to find the crack for AMD to get the wedge in. But now that the wedge is in, it's ignorant to flatly deny that it's happening.

AMD has been calling their shots for the past 3 years. They said that things were coming and then they delivered them. I have no reason to believe that Zen 3 will not arrive on time. And although 17% increased IPC is currently not confirmed by third parties, what can be said with certainty is that they will be faster than current Zen 2 parts. If Intel continues to sit as they have, it is likely that Zen 3 will finally eclipse all the last bastions that Intel has held onto, such as gaming.
 
Last edited:
Ryzen is certainly a better value, in multithread for sure, but you are dodging the point. AMD just caught up, they are not beating Intel in performance architecture, yet. I am not a Intel fanboy at all, as I have said. I also hope AMD continues to improve the core architecture, not just adding cores. I hope they pull another Athlon 64 innovation on Intel, and force the end of stagnation of CPU design we have suffered. I am just not a delusional AMD fanbot either.
They have just pulled an a64, beaten Intel everywhere except a few games and using significantly (or many times) less power. And at lower clocks, thus AMD IPC is higher than Intel. Only thing keeping them relevant is clocking 14nm to the limit.

I guess we have not seen the same benchmarks?
I had multiple a64/opteron rigs and they were much the same. Few games and certain video editing niches Intel were competitive in.
This is no different.

Ice lake IPC is going to be a little higher but the clocks also lower on 10nm. Certianly not going to see the fabled 20-30% claims. Only time you will see that is comparing unmitigated cpus to next gen.

AMD however has a newly established reputation of making very accurate IPC claims. Intel marketing currently has almost zero trustworthiness.
 
Funny how AMD flopped for a decade with no viable CPU against Intel and still survive... Now Intel flopped for a year or 2, maybe 4 if they really messed up their "next gen" and everybody is telling stories how "It is the end for Intel". LOL

I'm glad that AMD finally put some pressure on Intel since we're now getting 8+ cores in the mainstream. However, if AMD become complacent (remember Netburst vs K8 and then Conroe vs Phenom?), intel will crush them again with their (dirty) tactics and money.

And where most of the money is made (enterprise), the decision maker is usually the business guy, not the technical guy. This is simply a case of "nobody ever got fired for buying Cisco network equipment" with Intel being Cisco here.
Intel has sold the same cpu four times in 3 platforms
 
They have just pulled an a64, beaten Intel everywhere except a few games and using significantly (or many times) less power. And at lower clocks, thus AMD IPC is higher than Intel. Only thing keeping them relevant is clocking 14nm to the limit.

I guess we have not seen the same benchmarks?
I had multiple a64/opteron rigs and they were much the same. Few games and certain video editing niches Intel were competitive in.
This is no different.

Ice lake IPC is going to be a little higher but the clocks also lower on 10nm. Certianly not going to see the fabled 20-30% claims. Only time you will see that is comparing unmitigated cpus to next gen.

AMD however has a newly established reputation of making very accurate IPC claims. Intel marketing currently has almost zero trustworthiness.
Most recently...


Intel still has the advantage in most games.
 
Back
Top