Full Skylake reveal result? Waiting for Zen.

[mahoshojo]

Now Intel's DCG has >3B revenue each Q, and >60% of Intel's overall profit. It's a super high margin market.
If AMD manages to take 1% of it with ZEN, that means >300M each Q. And most likely good enough for them to stop losing money.

 

[richierich1212]

Why do these claims never match what people and OEMs buy? And what sites report?

Do you really think people buy based on performance? Or brand? For most consumer products people buy based on name alone. There's only 2 major CPU manufacturers, and Intel will sell more crap J1000 boxes than AMD better spec'd boxes just because they see Intel Inside.

 

[cmdrdredd]

Boy are you in for a learning experience. Remember the Athlon64, XP and X2 days?

I do and those days are long gone. Once Intel went to Conroe and ramped up AMD couldn't compete even if they spent their entire budget on R&D for their CPU division it wouldn't come close to what Intel can and does spend. As I said before, the best we can hope for is that they can be competitive.

Also you can't call someone new and a troll because they poke holes in your argument(DX12 benchmarks currently show more gains from clock speed and IPC than number of cores).

Do you really think people buy based on performance? Or brand? For most consumer products people buy based on name alone. There's only 2 major CPU manufacturers, and Intel will sell more crap J1000 boxes than AMD better spec'd boxes just because they see Intel Inside.

That's not exactly a fair way of looking at. There are those people yes, but I know many who buy based on what gives them the best performance in their budget. There are lots of people on this forum who have owned GPUs and CPUs from every manufacturer depending on what is good at the time for their set budget. I have no problem buying AMD CPUs if they perform better than the Intel equivalent price wise. That hasn't happened in a decade.

 

[ZaphodBbrox]

Now Intel's DCG has >3B revenue each Q, and >60% of Intel's overall profit. It's a super high margin market.
If AMD manages to take 1% of it with ZEN, that means >300M each Q. And most likely good enough for them to stop losing money.

Learn to do math. 1% of 3B is 30M.

 

[Dave2150]

Technology moves forward, if you want to play Ashes Of Singularity in late 2016 with 1000s of units you will need more than a quad CPU. That doesnt mean the game will not be playable with a quad Core with lower unit count.

Its like the GPUs, vast majority of Gamers have $100-200 GPUs, and yet games can make a $650 GTX980Ti crawl to its knees. I havent seen anyone implying that Game devs will not make games more GPU bound because the majority of Gamers have $100-200 GPUs.

Clockrate and IPC seem to be more important for Ashes of Singularity, even in DX12.

 

[mahoshojo]

Learn to do math. 1% of 3B is 30M.

Sorry I mean 10%.

 

[AtenRa]

Clockrate and IPC seem to be more important for Ashes of Singularity, even in DX12.

How many units are present in the benchmark ??? Because i dont think i show more than 200-300 units at the screen and that was only for a few seconds. There were times when there were only 10-20 units flying alone and not a single fire was shot.

Any way, i believe it is too early to make any conclusions about DX-12 gaming simply by looking the AotS benchmark alone.

 

[ShintaiDK]

Why not just take a chess game now you are at it.

There is nothing pointing to that DX12 will do anything for more cores. DX11 games can use just as many.

 

[Phynaz]

OK yea, it seems your right ...

things slowed down after 2006 but, the stagnation really has only been since 2011 ...

2006 to 20011 we went from core 2 duo to sandy bridge. And essentially, CPU single threaded performance went up by like < 10% annually on average.

From 2011 to 2015... we have gained less than 5% performance annually on average.

It's funny how people use Intel to make the trend over time comparison. People never use AMD as an example. What's the annual single threaded performance increase for AMD?

Maybe AMD would be a better benchmark.

 

[Arachnotronic]

It's funny how people use Intel to make the trend over time comparison. People never use AMD as an example. What's the annual single threaded performance increase for AMD?

Maybe AMD would be a better benchmark.

Exactly.

 

[richierich1212]

It's funny how people use Intel to make the trend over time comparison. People never use AMD as an example. What's the annual single threaded performance increase for AMD?

Maybe AMD would be a better benchmark.

Why shouldn't they? Intel has been on top for the about a decade. They have had a much higher revenue which also means more R&D invested. I don't get why we shouldn't get angry at stagnation in the PC industry.

 

[TheProgrammer]

Now Intel's DCG has >3B revenue each Q, and >60% of Intel's overall profit. It's a super high margin market.
If AMD manages to take 10% of it with ZEN, that means >300M each Q. And most likely good enough for them to stop losing money.

This. Best post in the entire thread. You've put some actual thought into what's going to happen next year. AMD is the only possible x86 player that could be relevant at all other than Intel, and they'll scratch back share.

There's no reason not to wait if you have a somewhat recent Intel CPU or even an AMD FX, they're all "good enough".

Who has the time to actually educate themselves?!?! It's way easier just to repeat the same tired arguments and misinformation repeatedly.

AMD's CPUs today are better than good enough, when I checked into what would be the best choice for the money for a server handling 900 clients per second, AMD won with a Bulldozer derivative.

These review sites literally work for Intel. The benchmarks are meaningless unless it's exactly what you intend to do with the chip.

Take your budget, then buy the best chip at that price.. since about 1995 if you do that comparison and the research, AMD has almost always won. Again, you have to do some hard work, Anandtech does not do that work for you... unless you do nothing with your computer and could be using a potato.

Ohh well, they look not so good now, probably should dump it and use the money at the local mexican eatery but instead I will hold it until they go defunct, or until Zen comes out and either is a success or they go defunct.

I am hopeful, but not optimistic.

Prepare for another golden era. AMD is bringing the fight next year.
But I don't mind the exclusive Intel guys too much, the saddest bunch are the Nvidia diehards. They don't even have x86. Pumping out garbage like Tegra as their CPU then riding Intel so they have something x86 to use.

But I agree, Zen better do well in servers, because despite all the salivating by AMD fans on these forums, it will have an *extremely* limited market in the consumer segment without an igp.

Mainstream consumer market is solidly 4core APUs at this point. 8cores are for overpowered gaming rigs and servers. No one buys a 5820K to use the IGP. Same for server-grade Zen.

I have no problem buying AMD CPUs if they perform better than the Intel equivalent price wise. That hasn't happened in a decade.

Suuurreeee you do. Yeah, because you read Anandtech reviews with your 6700K, when your needs are met with an etch-a-sketch.

I do and those days are long gone. Once Intel went to Conroe and ramped up AMD couldn't compete even if they spent their entire budget on R&D for their CPU division it wouldn't come close to what Intel can and does spend. As I said before, the best we can hope for is that they can be competitive.

Oh boy another peasant "tech enthusiast" R&D circle jerker. That worked out with Skyfail so well right? Intel is stuck with it for another ~4 years. They're done. Expect more power efficiency improvements like you've been seeing.
Oh, and make sure to keep buying.

Also you can't call someone new and a troll because they poke holes in your argument(DX12 benchmarks currently show more gains from clock speed and IPC than number of cores).

What I said was that at 4ghz the clock speed gains level off with anything sitting at SB IPC or newer. 8C/16T wins there. Zen is that. You're hopeless though and just want to win the argument. Ok you win, now go away troll.

Why not just take a chess game now you are at it.
There is nothing pointing to that DX12 will do anything for more cores. DX11 games can use just as many.

So much misinformation on this forum. Have you ever developed a AAA DX11 game? I have.
DX11 is not the future BTW.

 

[2is]

Why would I wait for something that MIGHT be almost as good as Sandy Bridge???

 

[drcurran]

There have been reports/rumors of Broadwell-E's cancellation:

"According to an alleged slide from Intel’s roadmap, which was published by Hardware.fi, the company intends to cancel release of its “Broadwell-E” in Q1 2016, but to unveil its “Skylake-E” chip in the Q3 2016. It is unclear whether the upcoming CPU will be compatible with the Intel X99 core-logic and LGA2011-3 platforms, but it is a likely scenario."

And Intel's lack of communication on the matter tells me they're still trying to make up their minds. My personal guess would be that Broadwell E does get canned if they can get Skylake E out by Q2 or very early Q3.

Broadwell-E isnt cancelled. And Skylake-E isnt moved anywhere.

 

[jpiniero]

AMD's CPUs today are better than good enough, when I checked into what would be the best choice for the money for a server handling 900 clients per second, AMD won with a Bulldozer derivative.

For servers, it's about TCO... namely power consumption/efficiency and heat. Bulldozer got wiped out in servers because it's so bad at perf/W moreso than raw performance. Even if AMD manages to hit the supposed SB IPC, they'd have to do it at a competitive perf/w to even have a shot at regaining server share.

These review sites literally work for Intel. The benchmarks are meaningless unless it's exactly what you intend to do with the chip.

The review sites are mainly about games, and of course AMD does much worse in games than Core.

So much misinformation on this forum. Have you ever developed a AAA DX11 game? I have.
DX11 is not the future BTW.

Whether developers will actually make DX12 games is still very much in the air considering it requires W10.

 

[mrmt]

AMD's CPUs today are better than good enough, when I checked into what would be the best choice for the money for a server handling 900 clients per second, AMD won with a Bulldozer derivative.

Mind share a scenario where TCO for Bulldozer server make sense?

 

[cmdrdredd]

This. Best post in the entire thread. You've put some actual thought into what's going to happen next year. AMD is the only possible x86 player that could be relevant at all other than Intel, and they'll scratch back share.




AMD's CPUs today are better than good enough, when I checked into what would be the best choice for the money for a server handling 900 clients per second, AMD won with a Bulldozer derivative.

These review sites literally work for Intel. The benchmarks are meaningless unless it's exactly what you intend to do with the chip.

Take your budget, then buy the best chip at that price.. since about 1995 if you do that comparison and the research, AMD has almost always won. Again, you have to do some hard work, Anandtech does not do that work for you... unless you do nothing with your computer and could be using a potato.



Prepare for another golden era. AMD is bringing the fight next year.
But I don't mind the exclusive Intel guys too much, the saddest bunch are the Nvidia diehards. They don't even have x86. Pumping out garbage like Tegra as their CPU then riding Intel so they have something x86 to use.



Mainstream consumer market is solidly 4core APUs at this point. 8cores are for overpowered gaming rigs and servers. No one buys a 5820K to use the IGP. Same for server-grade Zen.


Suuurreeee you do. Yeah, because you read Anandtech reviews with your 6700K, when your needs are met with an etch-a-sketch.



Oh boy another peasant "tech enthusiast" R&D circle jerker. That worked out with Skyfail so well right? Intel is stuck with it for another ~4 years. They're done. Expect more power efficiency improvements like you've been seeing.
Oh, and make sure to keep buying.



What I said was that at 4ghz the clock speed gains level off with anything sitting at SB IPC or newer. 8C/16T wins there. Zen is that. You're hopeless though and just want to win the argument. Ok you win, now go away troll.



So much misinformation on this forum. Have you ever developed a AAA DX11 game? I have.
DX11 is not the future BTW.

I think you need to stop posting entirely. Your entire argument revolves around insults and ignoring facts.

You are no developer, if you were you wouldn't resort to insults when someone challenges your argument based on feelings with facts.

Again Intel has offered better performance within my budget (and looking at the signatures on the forum, a good many others) than AMD for a decade.

 

[looncraz]

they can't even do a smaller SoC design apple gave them so i think they will screw up.

Actually, while it was true their ramp-up was looking bad at first, they have fixed those issues completely. Apple just couldn't afford to take the risk and wanted to make sure they could reach the necessary chip volume. Don't blame them, obviously.

They had 30% yield in the middle of April, but apparently managed 50% by the end of April. They are now "on par" with Samsung's yields (as should be expected) - those yields are claimed to be about 70%, though there is some doubt about the accuracy of that claim. They are now focused on 14nm LPP and should already be trying to create prototypes for AMD (there is also an advantage for AMD with the process wetting its beak on AMD parts :thumbsup: ).

As for Zen, it should fairly easily reach 40% more than Excavator with what we already know about it. Excavator has almost exactly the same IPC as Core 2, so 40% would put it right around Haswell's IPC... we just have to see how that is distributed - since not all instructions are equal.

 

[Abwx]

Why would I wait for something that MIGHT be almost as good as Sandy Bridge???

Yet another one of those blank statements that seems to be the specialty of some experts in everythings with arguments in about nothing..

I could say that it will be twice as fast and i wouldnt be more wildly speculating that you are..

 

[frozentundra123456]

How many units are present in the benchmark ??? Because i dont think i show more than 200-300 units at the screen and that was only for a few seconds. There were times when there were only 10-20 units flying alone and not a single fire was shot.

Any way, i believe it is too early to make any conclusions about DX-12 gaming simply by looking the AotS benchmark alone.

Wow did those goalposts move fast.

 

[tential]

On one hand, skipping broadwell and going straight to skylake-e rocks because I want skylake-e and not broadwell-e. On the otherhand... I would have had more time to save! Well, I hope Skylake-E impresses me although with 1 year away, I doubt I'll care.

Edit:
Zen is still an unknown... I don't think a SINGLE person wants it to fail. We ALL want choices and if an 8core/16thread CPU came out and was a good alternative to intel we'd hop on it in a heartbeat.

The issue is, we haven't seen anything worthwhile from AMD in awhile. My 4770k didn't have a better choice to get at the time.

I PRAY Zen changes that because I don't want a 1 player market, but I'm realistic since AMD has slashed theri R&D budget a LOT and is competing against 2 companies that are capable of outspending them in Nvidia/Intel.

At the same time though, moves like Freesync are great moves that cost AMD a little, but can convert customers. Although I have the HD7950, I would have bought a GTX980Ti, but Freesync UHD650 monitor is out, there is no Nvidia alternative (Gsync being closed and all), so now I go AMD.

So I think that AMD has made some smart moves recently, I just hope Zen comes out on time and is a competitive alternative. At the very least AMD needs to assure it has it's core fanbase as right now I've even seen AMD diehards with intel CPUs. AMD needs to at LEAST fix that with Zen.

 

[looncraz]

Things have been almost stagnant since the first core 2 duo when it comes to single threaded peformance. You are lucky to get 10%.

Skylake's nominal IPC is 56.12% higher than Penryn (Core 2 Quad 9xxx) according to Intel.

Accrued nominal IPC for Intel CPUs:

Penryn: 100%
Nehalem: 115%
Sandy Bridge: 127.65%
Ivy Bridge: 132.76%
Haswell 139.39%
Skylake: 156.12%

Currently, Excavator is about 0.93% faster than Penryn. Zen is supposed to be 40% faster, which should put it right around Haswell.

If they manage that, then they only need clock speed parity and decent energy efficiency to be competitive.

Interestingly, I think AMD can get 33% IPC improvement almost for free.

The module design is suppose to be about 85~90% efficient (each core performing about 85~90% of its optimum). The cache system is horrendous, and cost performance from phenom II's cache system (about 5~6% it seems, load dependent). Bulldozer integer cores are massively narrow, with only two ALUs and two AGUs.

Zen, then, should gain 10~15% from dumping CMT. They should gain 5~6% in many loads just from using a write-back cache like what we saw with Phenom II - at 14nm, with the reduced latency that would bring, that should be closer to 8~10%. Going back to the Phenom II layout, with Bulldozer's improved ALUs and AGUs, would bring an easy 10% improvement.

That is a 30~42% improvement right there, with no other changes. Of course, these aren't all fully cumulative, so AMD will have had to make more changes to fully flesh out a 40% improvement. I just think many people underestimate just how much of a burden the front end in a Bulldozer CPU has to process the load for two cores. Sure, it can do it, which is amazing, but it can't do it without adding retirement latency, which reduces performance in MANY different types of workloads.

 

[2is]

Yet another one of those blank statements that seems to be the specialty of some experts in everythings with arguments in about nothing..

I could say that it will be twice as fast and i wouldnt be more wildly speculating that you are..

You could. And I could come back and ask you to put your money where your mouth is and see who's predictions will be closer to the truth, and you won't do it. Why? Because you know as well as I do you aren't going to win.

 

[nismotigerwvu]

Personally, I think Zen for consumers without any igp is because they simply did not have time/resources to develop and validate one (an igp), or they are waiting for HBM2.

But I agree, Zen better do well in servers, because despite all the salivating by AMD fans on these forums, it will have an *extremely* limited market in the consumer segment without an igp.

I think you're fairly close to the truth here. They likely saw the timetable HBM2 fed APUs as more in line with Zen+ (or whatever the revision will be called) and cobbled together a plan to launch Zen 1.0 as a straight CPU for the server and high end desktop market. Server lines have never really placed a huge value on iGPUs and can likely offer up an attractive bundle of Zen CPUs and dedicated GPUs for OEMs. Really Zen only needs to be competitive on either of these fronts to keep the ship afloat until the APUs come around and next gen console contracts arrive.

 

[nismotigerwvu]

Clockrate and IPC seem to be more important for Ashes of Singularity, even in DX12.

Something seems a bit off with these numbers, why would running on high detail increase the FPS on AMD hardware over running on low? I get that we're looking at CPU bound cases (just like with the i3) but the performance should stay the same, if not decrease slightly there.