Intel Core Series 9000 Specifications published - merely 100 and 200 MHz speed bumps

#5562638 Posted on: 07/04/2018 08:37 PM

---

Ice lake architecture is the answer, complete redesign and no more roots from Sandy bridge..

The question is when?? there was talk this year in q4, but I'm not holding my breath.

#5562641 Posted on: 07/04/2018 08:40 PM

---

10nm is not going to offer that great of a performance difference. I'm sure they've actually had it working for a while, but it probably requires lower clocks to be stable, so they figure the small gains in efficiency are outweighed by the performance loss of the lower clocks. Nobody else seems to be able to reach 5GHz at 12nm or smaller (meanwhile, Intel isn't the only one to reach 5GHz at 14nm). So, if Intel is actually hoping to get the same clock speeds at 10nm as they do at 14nm, personally, I see that as a pipe dream. They're wasting too much time on this, and should instead be focusing development on a new architecture (which again, they have had PLENTY of time to do). They have the resources to do both.
Anyway, my point is they won't be squeezing more performance out of their existing process. It's pretty much maxed out, and the Spectre and Meltdown vulnerabilities have only worsened IPC. 100-200MHz will only mask those losses. Like I said before, the 9000 series doesn't have anything especially different about it. It doesn't have any leverage over Coffee Lake, except things like the rumored 8-core chip (and I don't think that's going to be a best seller).

Historically, sticking to your guns with a regressing plan practically never helps in the long run, especially when the rest of the entire industry has moved ahead. Historically, underestimating an opponent who isn't dead yet will backfire. Historically, investors get afraid when companies aren't being innovative, and back out. Intel wouldn't have had to resort to "drastic measures" if they had better future-planning.

Remember, Intel's net revenue in a single year is higher than all of AMD's assets combined. For them to have put only some of their eggs in just 1 basket while leaving the rest to rot is just asinine. When Nvidia is ahead, they don't slow down that much. They invest in new technologies. They attempt stuff like GPP (even though they didn't need to). They have definite plans in the event AMD pulls ahead by a wide margin. People whine about the 1100 series but they fail to realize everything else Nvidia has done with Volta.

I would like to point out AMD in general has very little to do with what I'm talking about here. They're not the only ones Intel should be worrying about. IBM, ARM, Nvidia, and RISC-V should all be putting on the pressure.

#5562647 Posted on: 07/04/2018 09:03 PM

---

From what i understand Intel can´t do much about this line up because this line wasn´t not supposed to exist! From what i understand Skylake or Kaby Lake were supposed to be the last CPUs using the 14nm process and current arch and then Ice Lake should have been released on a 10nm process with a brand new arch. The problem is that the 10nm process is broken and their new arch is tied to it making it very difficult to port it to 14nm. So with Ice Lake delayed until they can fix the 10nm mess, Intel only had one solution for their CPU releases, to carry on with their current CPUs with some small tweaks, like increased clocks or cores.

So in the end this all Intel can do for the moment until they sort out their 10nm so they can release a new CPU architecture. Good thing that their current CPUs are still good enough otherwise things would be problematic.

#5562648 Posted on: 07/04/2018 09:07 PM

---

This. If you like high framerates, then you need all the IPC you can get. This is less important the better games are made, but that really only applies to specific AAA releases. Many games are still limited by the power of a single thread. (And as below, even Intel's CPU still can't manage it without issue)
Crysis is an example of a game that needs a re-release to fix this.



Intel is like Nvidia, you'd like to purchase a competitor's product. But because of weaknesses in specific areas, it doesn't make it desirable in all scenarios.
I'm not one to talk as Ryzen 2 is clearly still a big improvement in IPC compared to my 1st generation i7. (Which in a lot of scenarios is easy to see the result of the weaker IPC. Even at only 60FPS).

I'm still heavily leaning towards a Ryzen 2 for my next CPU anyway.

#5562649 Posted on: 07/04/2018 09:34 PM

---

It's certainly true that AMD caught Intel with their pants down, but this isn't really a surprise - it's what happens when companies gain a monopoly for an extended period of time. Board meetings would focus on how to maximize quarterly profits - the odd person who stresses more R&D to counter a potential future threat from a small rival would be regarded with scorn ("crazy Bob is at it again with his Zen talk "). AMD was out of the loop for so long that Intel forgot how to compete.

Remember that AMD stock was cheaper than a happy meal at one point, with many people predicting bankruptcy or liquidation due to the massive debt and endless quarterly losses. The market had almost entirely given up on the company, and only a few would have imagined that they would come back with a viable product. In some ways, it's similar to how people talk about AMD regarding GPUs today - most people would say that Nvidia is so far ahead that AMD will never be able to catch up. Can any of us really imagine AMD coming out with a product that can match or exceed Nvidia's best?

I don't really blame Intel for being in their current situation. Zen was a truly ambitious project and there was no telling how it would do against Intel CPUs, which had the benefit of market dominance and years of software optimizations (we saw this in the first Ryzen 7 reviews, where it trailed badly in some games). My guess is that Intel thought any new architecture that AMD created could be countered by their Core chips, and to a large extent this was true. The challenge for them is to continue the fight when AMD moves on to 7nm with Zen 2, which depends largely on how quickly they can get their 10nm process going.

#5562650 Posted on: 07/04/2018 09:38 PM

---

Your brain is pile of crap.
How many cores should (the new) i7 have by that twisted logic then?

8, obviously, the way it should be without Intel deciding to try and rip people off every time they get.

#5562651 Posted on: 07/04/2018 09:56 PM

---

Yea? Who says it won't? I lived under the impression the upcoming CPU will have 8 cores. The guy was babbling some crap about the i5 variant only having six. Wtf? That makes no sense.

#5562665 Posted on: 07/04/2018 11:21 PM

---

Yea? Who says it won't? I lived under the impression the upcoming CPU will have 8 cores. The guy was babbling some crap about the i5 variant only having six. Wtf? That makes no sense.

He was "blabbing" about the fact that the i5 next gen still only has 6 threads instead of 12, which is rightfully so.

Of which you asked how many cores should the new i7 have, which doesn't make any sense, since they did not specifically talk about cores, they talked about threads.

As well, if you think the new i7 series is going to have 8 cores and 16 threads even though it is seeming they are going to have 6 core 6 thread i5s, i don't really know what logic you are using to come to that conclusion.

If they make the new i5 be 6 core 6 thread, and the new i7 be 8 cores 16 threads, where are the 6 core 12 threads going to (what is currently the i7s)?

Or are you implying they are just not going to have 6 core 12 thread processors?

Both scenarios make zero sense.

You seem very new to PC electronics and/or simply not paying attention to what is actually said, and deciding to insult people by saying their "brain is pile of crap". I think you'll go a long ways on this forum, for sure.

Here i'll leave you with what you quoted and what you stated so you can figure out what it is you didn't read and what you stated because of what you didn't read

Core i5 still only 6 threads? What a pile of crap.

Your brain is pile of crap.
How many cores should (the new) i7 have by that twisted logic then?

#5562733 Posted on: 07/05/2018 09:11 AM

---

Oh, wait. He wrote threads and I read cores.
Ok, I apologize then!

#5562735 Posted on: 07/05/2018 09:18 AM

---

Honestly, still not seeing me buying an Intel CPU next year, sorry. Let's see what Ryzen 2 brings to the table.

People probably would be happier with a 100-200MHz bump on Ryzen clocks than on Intel's...

#5562749 Posted on: 07/05/2018 10:34 AM

---

Sadly enough, Intel is gonna win again. The bigger brother of 8700K holding 8c/16t will cost at average 33.3eur more per core plus 16.67eur per thread. Mathematically that means ((33.3x2)+(16.67x4))=66.6+66.68=133.28eur.

With other words, it is gonna cost exactly as much as i7 7820X. But it will crush this server wannabe PoS. We are talking about a real gaming cpu and a new generation. Well everyone has eyes for 9 series of intel and ryzen 2.

Wrong. 9 series is slower in gaming than i7/i5. The 7980XE performs far worse in gaming than an 8600k. It's simply not meant for gaming.

#5562764 Posted on: 07/05/2018 11:26 AM

---

9 series is the 9XXX not the i9. Irrelevant.

#5562811 Posted on: 07/05/2018 02:12 PM

---

As apparent as that may be, this is precisely what I'm criticizing. However, you do have an interesting point when it comes to "crazy Bob".

Slight tangent, but I think people heavily over-exaggerate how bad the GPU situation is (I'm not pointing fingers at you, I'm just saying in general). Gamers (and really just gamers, nobody else) complain that the 1080Ti has nothing to trounce it and act like the competition is non-existent, as though the fastest performing GPU is the only thing worth considering. This isn't like the FX-9000 series vs Haswell. Yes, GCN/Vega is obviously behind, but not by a wide enough margin to warrant the complete lack of hope people have.
So that being said, I think it is perfectly reasonable for AMD to come out with a product that can match Nvidia's best, with Navi. On the other hand, I also think it's perfectly reasonable to have doubts that it will outperform Nvidia's best (I personally have such doubts). GPU competition could be better, but it could be a lot worse too.

I still disagree. This is no different than a cheesy action movie, where character A is fighting character B, and A gets knocked to the ground, face down. B then just turns his back to A, slowly walking away, wrongfully assuming A is unconscious (or dead) without confirming, and then gets backstabbed. If you're fighting for your life, the last thing you should do is presume your assailant won't get up again, and willingly put yourself in a weak position. And that's exactly what Intel did.
The only difference between this scenario vs Intel/AMD is AMD accidentally punched themselves in the face, got winded from it, and Intel was kinda just like "ok... I guess I'm just going to continue on my merry way". Then, AMD caught their breath and didn't miss at the 2nd attempt.
Remember, AMD isn't the only competition - Intel isn't the best in every market they're involved in. They should be used to not underestimating others.

#5562816 Posted on: 07/05/2018 02:21 PM

---

Perhaps there's something important that I don't know (which is a good possibility) but what exactly prevents Intel from releasing the new arch on 14nm? Worst-case scenario, they could just sell 8-core products and smaller. Sure, it's somewhat counter-intuitive to have mainstream and low-end Xeons get the latest tech, but real progress is better than none at all.


Their current CPUs are good enough, but ultimately my point was the 9000 series has nothing compelling to offer, unlike previous generations.

#5562853 Posted on: 07/05/2018 03:46 PM

---

Slight tangent, but I think people heavily over-exaggerate how bad the GPU situation is (I'm not pointing fingers at you, I'm just saying in general). Gamers (and really just gamers, nobody else) complain that the 1080Ti has nothing to trounce it and act like the competition is non-existent, as though the fastest performing GPU is the only thing worth considering. This isn't like the FX-9000 series vs Haswell. Yes, GCN/Vega is obviously behind, but not by a wide enough margin to warrant the complete lack of hope people have.
So that being said, I think it is perfectly reasonable for AMD to come out with a product that can match Nvidia's best, with Navi. On the other hand, I also think it's perfectly reasonable to have doubts that it will outperform Nvidia's best (I personally have such doubts). GPU competition could be better, but it could be a lot worse too.
Well, that's the thing. Based on recent history, how seriously would you take the idea that AMD will match Nvidia's best? If you had to bet your house on this, which way would you bet? Frankly, I doubt that Huang is losing sleep over the possibility of AMD matching or snatching back the performance crown anytime soon. My guess is that he is keeping an eye on it, but focusing most of his R&D on other projects (AI, cloud computing, self-driving). He certainly wouldn't see it as a mortal threat to his company.

I still disagree. This is no different than a cheesy action movie, where character A is fighting character B, and A gets knocked to the ground, face down. B then just turns his back to A, slowly walking away, wrongfully assuming A is unconscious (or dead) without confirming, and then gets backstabbed. If you're fighting for your life, the last thing you should do is presume your assailant won't get up again, and willingly put yourself in a weak position. And that's exactly what Intel did.
The only difference between this scenario vs Intel/AMD is AMD accidentally punched themselves in the face, got winded from it, and Intel was kinda just like "ok... I guess I'm just going to continue on my merry way". Then, AMD caught their breath and didn't miss at the 2nd attempt.
Remember, AMD isn't the only competition - Intel isn't the best in every market they're involved in. They should be used to not underestimating others.
Well, I don't think Intel is in a life-and-death struggle with AMD here (even during the height of the Athlon days, Intel was never seriously threatened). Recent sales data from mindfactory shows that Intel has retaken the lead (although Zen+ has narrowed the gap once again) and their struggle with Ryzen so far is in trying to maintain that lead. In short, although Ryzen presents a formidable challenge, Intel is still managing to hold their ground.

Intel still has time to come up with a suitable response. They may lose some market share to AMD during that time but it's not a gloom-and-doom scenario by any stretch. They currently can't do much more than tweaking and refining their current products but there is still hope that they can get their next-gen fab tech going and that's where their focus is at the moment.

I do agree that the 9000 series is kind of stupid. It would make more sense to brand these as 8000 series chips, but you could make the same argument with the Radeon RX 500 series. I guess they think it will generate more sales if people think of it as a next-gen product ("This system features Intel's latest 9th generation Core i7 chip for blazing fast performance!").