Ivy Bridge-E or what happens when there's no competitions

[inf64]

1. I will guesstimate that it is because it is one humongous ship and the kind of planning and execution involved dictates that you from the get-go knows (need to know) what direction and speed you are sailing two years down the 'road'. And of course roi and margins and and and .... I have no problem imagining that intels execution would have been different if they knew, ahead of time, that bulldozer would be a massive fail.
So it is like, if intel gets the whole cake for them selves, but need 50% more time to eat it all .. they would be kind of nuts not to eat it .. or some of it at least, right?

2. Is excellent logic and applies just fine, imo, to the concept at hand.

A bit off topic,sorry,but I have to thank you for mentioning your sig . I had no idea it was shintai's "prediction" . I updated my sig

 

[Torn Mind]

Approximation is a viable way to analyzing stuff that you do not know the exact details of but knowing something is there. It only means your conclusion can't be as definitive as it would be with more information.

 

[Homeles]

1. "Uhm no. Your thread is an epic fail of misunderstanding."
- I call double negative.

That's not a double negative. In this case, the "fail" is comprised of misunderstanding.

 

[aigomorla]

now u see why i took my haitus from the CPU world...

things crawling compared to when i was active..

when i was active a new cpu would pop out every 6months..

Now i check back and see great... were still stuck in this generation....

 

[NTMBK]

now u see why i took my haitus from the CPU world...

things crawling compared to when i was active..

when i was active a new cpu would pop out every 6months..

Now i check back and see great... were still stuck in this generation....

Only in the PC world... take a look at how ARM is coming along

 

[cytg111]

That's not a double negative. In this case, the "fail" is comprised of misunderstanding.

hehe, right.

 

[Homeles]

hehe, right.

More conceit, as expected.

 

[ShintaiDK]

Only in the PC world... take a look at how ARM is coming along

I would say ARM moves slower than x86. Anytime ARM shows performance, the power consumption goes crazy. Else we wouldnt have the "octocore".

 

[Fjodor2001]

I would say ARM moves slower than x86. Anytime ARM shows performance, the power consumption goes crazy. Else we wouldnt have the "octocore".

The "octocore" is no proof of that. Instead it's a sign that there is no CPU core that will perform well perf/watt-wise over the complete performance range. You can always do better if you limit the performance range that the CPU will operate in. Hence having two different core types on the same die that cover separate performance ranges makes sense.

But this has already been discussed several times in other threads...

 

[ShintaiDK]

The "octocore" is no proof of that. Instead it's a sign that there is no CPU core that will perform well perf/watt-wise over the complete performance range. You can always do better if you limit the performance range that the CPU will operate in. Hence having two different core types on the same die that cover separate performance ranges makes sense.

But this has already been discussed several times in other threads...

It was a pure example that ARM couldnt deliver both worlds. And that performance increase for ARM means higher power consumption.

 

[Homeles]

Shintai is saying that ARM isn't the laws-of-physics-breaking design it's hailed to be. There seems to be this idea that ARM won't run into the same diminishing returns that x86 did, and the A15 is proof of that.

 

[tweakboy]

I don't know, it feels they are changing the market segmentation a bit. Or maybe this was what they were aiming from the beginning. The E series, dedicated to consumers, is more and more linked to the Xeon line, always a generation behind the "entry-level" desktop CPU's. But this way the E series resembles more a workstation CPU, or entry-level server CPU, dedicated to professionals, not gamers.

I am wondering, Haswell EP should have DDR4 support for example. Presumably, Haswell EP and E (xeon and consumer) will lunch in the same time, around the time the Haswell gets a dye shrink. Will the consumer version also get DDR4? Will also Broadwell get it?

DDR4 is in 2016 to 2017 .

 

[tweakboy]

Haswell for enthusiast is 2014 First Ivy Bridge E cash cow.

 

[cytg111]

But that is really not the potential of ARM is it? As I see it, the potential of ARM is the ecosystem it is bringing with it, linux / android and now windows rt ..
ARM may be successful, where others have failed, at bringing an alternative ISA (to x86) to the table, and that alone has monopoly-like-breaking potential. Time will tell.

 

[ShintaiDK]

But that is really not the potential of ARM is it? As I see it, the potential of ARM is the ecosystem it is bringing with it, linux / android and now windows rt ..
ARM may be successful, where others have failed, at bringing an alternative ISA (to x86) to the table, and that alone has monopoly-like-breaking potential. Time will tell.

Windows RT is already dead. And Android is just a base for the mostly OS independent Java apps. In terms of Android, just look at Samsung, they are backing Tizen now with several releases in the end of the year.

ARM, like the companies, suffers greatly from trend changes and can be replaced with anything else over night essentially. Its a brutal marketsegment because there is no foundation and no common baseline unlike the PC. You need x86, but you dont need ARM.

 

[cytg111]

While RT is not taking off like a rocket I think its too early for the last rites.
Either way you put it ARM has gained massively the last couple of years, has it lost momentum?
Android is here to stay, at least for a little while longer and I have a very hard time seeing Samsung turning its back on it.. Why would they? They dont like money?

And I agree with your last section, and that is the entire point, continue to grow the ecosystem around arm and windows rt and you will soon have a viable alternative to x86 on the desktop. Nice ! Double Nice !

 

[ShintaiDK]

A last note for today. Android runs on x86 too.

 

[TuxDave]

A last note for today. Android runs on x86 too.

Intel has so many software/OS/compiler guys that they probably could get any OS to run on x86.

 

[Torn Mind]

Intel has so many software/OS/compiler guys that they probably could get any OS to run on x86.

A compiler is a piece of software used to turn "code" that you see into machine code for the machine to execute. There is no person doing the compiling. Microsoft Visual Studio, for example, comes with an x64 compiler, 32-bit compiler, and a ARM compiler. Intel is not pulling the strings here, as you can cross compile using gcc if you're using Linux.

For linux distros or things built on the Linux kernel like Android, x86 compatibility is simply a mere matter of using a compiler to compile the code to work for the Atom or whatever is on the phone.


Lol, developers use emulators so that can test out their Android apps on a Mac or PC.

 

[tweakboy]

June 2013 Ivy Desktops 6 and 8 core. 12 core for xeon. gl

 

[colonelciller]

The same goes for all the tinfoil stories about higher prices.

I cannot wait for the day that the use of the "tinfoil" comment in place of an actual coherent argument dies permanently in the fire pits of internet shame, never to return.

...pulling out "tinfoil" as a retort is just not doing it right.

 

[james1701]

now u see why i took my haitus from the CPU world...

things crawling compared to when i was active..

when i was active a new cpu would pop out every 6months..

Now i check back and see great... were still stuck in this generation....

I miss those days.

 

[colonelciller]

Intel probably sows the seeds of distrust when they manipulate chip prices as they do.
In order to see the blatant price manipulation you just need to look at the chips, their specs, and the prices.

An excel spreadsheet for Intel CPUs with:

• the chips listed in a vertical stack according to chip model
• all chip attributes arrayed to the right of said stack

... make such a spreadsheet (I have) and BAM, there it is clear as day. The insane price manipulation (for the exact same product... with cheaper "models" having extra work put in to deactivate cores and features) becomes clear as crystal within the Xeon product line.

The only explanation for the price structure is "milking" the market, which they can do because they have no competitor... they have a market monopoly.
Behavior like this by the one company who completely dominates the market breeds distrust... and thus this thread. If you wonder why people are skeptical of Intel's motives, Intel's past behavior is your answer.

 

[frozentundra123456]

Intel probably sows the seeds of distrust when they manipulate chip prices as they do.
In order to see the blatant price manipulation you just need to look at the chips, their specs, and the prices.

An excel spreadsheet for Intel CPUs with:

• the chips listed in a vertical stack according to chip model
• all chip attributes arrayed to the right of said stack

... make such a spreadsheet (I have) and BAM, there it is clear as day. The insane price manipulation (for the exact same product... with cheaper "models" having extra work put in to deactivate cores and features) becomes clear as crystal within the Xeon product line.

The only explanation for the price structure is "milking" the market, which they can do because they have no competitor... they have a market monopoly.
Behavior like this by the one company who completely dominates the market breeds distrust... and thus this thread. If you wonder why people are skeptical of Intel's motives, Intel's past behavior is your answer.

With a 3570k, you get amazing performance in a 230.00 chip that is unlocked for overclocking. To me that is a great, great deal. Or you can get a pentium or celeron for low end performance for 50 to 80 dollars. What do you want, for them to give them away?

 

[colonelciller]

With a 3570k, you get amazing performance in a 230.00 chip that is unlocked for overclocking. To me that is a great, great deal. Or you can get a pentium or celeron for low end performance for 50 to 80 dollars. What do you want, for them to give them away?

missing the point completely.

anyhow, I was specifically talking about the Xeons as that's where the market manipulation is at its most absurd. The manipulation is also easily viewable in an excel document, just as described. There's this nifty little feature in excel:

Highlight a column such as Memory Bandwidth, or Core Clock, or Max Turbo, or Threads per core, or Cores, or QPI link speed, or Max Supported Memory, or Max Supported Memory Speed, or L3 Cache, etc... Highlight a column across all of the i7 CPUs, E5-Xeons (1 socket xeons, 2s Xeons, etc)... and then with the highlighted column use the Excel feature:

conditional formatting > data bars > gradient fill

that puts these nifty little horizontal bar graphs in each cell according to the cell value. Do this separately for each column of data and you've got a real nice (Crystal clear) picture of ascending "CPU model", ascending "CPU price" and the extremely well-thought-out manipulation (deactivation/diminishment) of features from the baseline model which has an ultra-insane pricetag to those models which have had extra work done to them to the features so that each Lower "CPU Model" has a bit less and less of something here, something else there... and so on and so on... cheaper and cheaper as more work is put in to cut off features.