Intel Broadwell-K & Skylake (non-K) desktop CPUs to launch in Q2-2015

[Enigmoid]

A lot of people holding on to dinosaurs... 2006 was 8 years ago, yet I still see Core 2 machines everywhere. Pentium 4 era machines are very rare, though. I see the occasional Athlon 64... probably the oldest line I see on a somewhat regular occurrence. Everything else has croaked, or is too slow to use.

Come to my university, Core i is practically non-existent, core 2 is pretty rare and the majority are Athlon 64s and Pentium 4s.

Computers in my lab include an old core 2 duo system, an ancient p4 computer and an even more ancient and slow celeron desktop from the p4 era (single core) -- not fun for matlab and statistical computations. Newer machines include a i3 (came with equipment so doesn't count as its not being used for general usage) and an i7 (currently not in use because you know, logic). A P3 with win 98 drives most of the equipment.

The library computers are better, lots of core 2 system with the occasional pentium, i3, i5 or i7 some p4s (it just seems that they throw systems out there, you will see i7s, p4s and core 2s side by side in the same chassis). However around campus core 2 and p4 rule the roost. Win xp everywhere.

I am honestly surprised at how snappy those old systems are. Running a bare xp install, browsing performance is very good and heavy excel usage isn't a problem.

 

[Homeles]

Yikes. At mine, half the kids are rocking Macs.

The school computers are either Phenom IIs, or some sort of Core i series. Lot of Macs there too.

The monitors are pretty much all 1080p too.

 

[Enigmoid]

Yikes. At mine, half the kids are rocking Macs.

The school computers are either Phenom IIs, or some sort of Core i series. Lot of Macs there too.

The monitors are pretty much all 1080p too.

You are probably going to a much wealthier university.

Macs (for students) are popular but there are only a few on the campus itself.

University, with only the athlon64 exception, is intel based.

1080p monitors are rare but at least there are no CRT screens.

 

[Homeles]

I'm at the University of Utah. Not drowning in cash, but they're pretty well off.

 

[ShintaiDK]

Not at all. Most "average persons" I know were on 3-4 year upgrade cycles prior to around 2006.

And the world crashed financially. And people are still holding on to their money. Thats pretty much easy to display by looking at average savings since.

Now I am sure you wish to play the tablet and smartphone card. Smartphones tho are essentially just replacing older regular cellphones. And both of them are forced into a steep replacement cycle due to being crappy products and the manufactors on purpose outdating them by not offering software upgrades and/or releasing updates that makes them crawl. Use and throw away mentality.

Example from Denmark:

Consumer Quota:

Average account holdings:

Then I am sure someone will make up more rubbish about x86 is dying, Intel is in trouble and so on. It just have no hold in reality. Companies and consumers are still sitting on their money.

And I would like to note, Denmark is one of the countries that have handled the economic crisis the best. Companies are even buying danish bonds with negative interest instead of investing in development.

 

[ithehappy]

So I had a plan to upgrade in Q1 in 2015, what should I aim? 4790k or 4930k or wait for Skylake?

 

[ShintaiDK]

So I had a plan to upgrade in Q1 in 2015, what should I aim? 4790k or 4930k or wait for Skylake?

In Q1 2015 the 4930K have been replaced, and Skylake is not out for another 3 months.

 

[Fjodor2001]

And the world crashed financially.

True, but we had economic downturns in the early 90s and 00s too, but it did not affect upgrade cycles as much then. So the economic downturn is not the primary reason why people are not upgrading their desktop PCs as frequently nowadays.

The major reason is instead that the benefit of upgrading to a later desktop PC model is not as great as it was before. Back in the P2/P3/P4 days, performance improvements per CPU generation were much greater, and higher performance was also of definitive benefit to the average user. E.g. going from P2/233 (May 7, 1997) to P3/800 (December 20, 1999) could seriously be felt, and there was only ~2.5 years between those two processor releases.

 

[ShintaiDK]

True, but we had economic downturns in the early 90s and 00s too, but it did not affect upgrade cycles much then. So the economic downturn is not the primary reason why people are not upgrading their desktop PCs as frequently nowadays.

Can you provide documentation for that? The relation between shipped CPUs and peoples saving accounts? Not to mention any parameters relating to the ever more bloated software of the time?

The major reason is instead that the benefit of upgrading to a later desktop PC model is not as great as it was before. Back in the P2/P3/P4 days, performance improvements per CPU generation were much greater, and higher performance was also of definitive benefit to the average user. E.g. going from P2/233 (May 7, 1997) to P3/800 (December 20, 1999) could seriously be felt, and there was only ~2.5 years between those two processor releases.

You make it sound so rosy. But what you havent told is that back in those days, TDP went from 20W to 130W. Also you forgot to tell that you had 300Mhz CPUs on may 7 1997 as well. And it went from 350nm to 180nm in the timeframe as the key driver.

 

[Fjodor2001]

Can you provide documentation for that? The relation between shipped CPUs and peoples saving accounts? Not to mention any parameters relating to the ever more bloated software of the time?

Can you provide documentation of the contrary?

You make it sound so rosy. But what you havent told is that back in those days, TDP went from 20W to 130W. Also you forgot to tell that you had 300Mhz CPUs on may 7 1997 as well. And it went from 350nm to 180nm in the timeframe as the key driver.

P2/233: 34.8 W (May 7, 1997)
P2/300: 43 W (May 7, 1997)
P3/800: 20.8 W (December 20, 1999)

For P2 see this, and P3 this.

And yes, P2/300 -> P3/800 could be seriously felt too... :biggrin:

 

[ShintaiDK]

Can you provide documentation of the contrary?

So I assume you cant back your statement up with facts. I am somehow not surprised.

P2/233: 34.8 W (May 7, 1997)
P2/300: 43 W (May 7, 1997)
P3/800: 20.8 W (December 20, 1999)

For P2 see this, and P3 this.

And yes, P2/300 -> P3/800 could be seriously felt too... :biggrin:

You listed the P2 to the P4 as a time where everything was just rosy and perfect. Also your selection contains a 2 node jump from 350nm to 180nm. The end of one and the start of the other for maximum difference. Its about equal to me saying that in Q1 2010 you could buy a 2.93Ghz dualcore with HD IGP at 73W. While in Q2 2012 I could buy an i7 3770 with HD3000 thats 3x faster CPU wise at 77W. AMAZING PROGRESS!

The point that you dance around to avoid is that outside your very narrow selection. It simply doesnt look as good as you try to portrait to fuel the delusion about todays progress. On the same 180nm process you also had 72W 2Ghz P4 from August 2001. And before that you had 200Mhz PPro from 1995.

If you make a chart and plot in performance/W. You would notice todays CPUs advance rather well. And thats even with the continual faster added IGPs and other integration.

 

[witeken]

But it's not nearly as fast as before.

 

[Fjodor2001]

So I assume you cant back your statement up with facts. I am somehow not surprised.

And you can't back up your claims saying it's not true either. I am somehow not surprised.

Do you really believe we'd see 3-4 year desktop upgrade cycles for the average Joe nowadays if it wasn't for the economic downturn? People would spend $600+ on a new computer to get 20% better CPU performance and 50% better iGPU performance, which most won't have much benefit of anyway? Seriously? 🙄

You listed the P2 to the P4 as a time where everything was just rosy and perfect. Also your selection contains a 2 node jump from 350nm to 180nm. The end of one and the start of the other for maximum difference.

Yes, that was the whole point. It was one of the golden periods. We've seen both ups and downs since then. But to be honest, the trend has been going downwards consistently for a long time. And looking at the current state and a few years ahead, the desktop CPU arena is looking very dull indeed.

Just imagine if we'd see the same improvements now as back in 1997-1999. Then we'd go from a 3.4 GHz @ 77 W 3770K to a similar one at ~9 GHz and 35 W in around 2.5 years. We are so far off from that it's ridiculous unfortunately. :\

 

[Idontcare]

This got a little off-topic...

 

[cytg111]

True, but we had economic downturns in the early 90s and 00s too, but it did not affect upgrade cycles as much then. So the economic downturn is not the primary reason why people are not upgrading their desktop PCs as frequently nowadays.

There is economic conjectures and there is full blown knockouts and 2008 was a knockout, counted out, carried off stage on a stretcher. People stopped buying anything, the very foundation of consumption on which our "democracy and way of life" is based upon was nullified. Worst threat to common society? The total independent bio-dynamic farmer that never has reason to buy anything ever again.
To think that this crisis did/do not have an substantial impact on PC sales? Common.

 

[ShintaiDK]

Some good news:
http://newsroom.intel.com/community...ll-year-revenue-and-gross-margin-expectations

As a result of stronger than expected demand for business PCs, Intel Corporation now expects second-quarter revenue to be $13.7 billion, plus or minus $300 million, as compared to the previous range of $13.0 billion, plus or minus $500 million.

Intel now expects some revenue growth for the year as compared to the previous outlook of approximately flat. The change in outlook is driven mostly by strong demand for business PCs.

Full-year R&D plus MG&A spending is expected to be $19.2 billion, plus or minus $200 million, higher than the prior expectation of $18.9 billion, plus or minus $200 million, driven mostly by revenue- and profit-dependent items.

x86 in growth again.

 

[witeken]

"Some good news." Understatement of the day I guess. INTC reached $30 for the first time since 20 Feb 2004, 10 years ago.

$13.4 - $14.0 billion is quite a dramatic change from $12.5 - $13.5 billion.

 

[Ajay]

OT, but: That's a good sign for the IT industry in general, if corporations are finally feeling comfortable enough to step up PC purchases. I hope that companies will start loosening their pocketbooks and start spending the money they have been banking.

 

[TuxDave]

"Some good news." Understatement of the day I guess. INTC reached $30 for the first time since 20 Feb 2004, 10 years ago.

$13.4 - $14.0 billion is quite a dramatic change from $12.5 - $13.5 billion.

There are pigs flying outside my window. What's happening? 😛

 

[witeken]

Intel updated its Q2 2014 forecast from $13B plus or minus 500M to $13.7B plus or minus 300M, killing all the death of PC conspiracies, followed by a $2.00 or 7.20% increase in share price.

 

[Homeles]

Intel updated its Q2 2014 forecast from $13B plus or minus 500M to $13.7B plus or minus 300M, killing all the death of PC conspiracies, followed by a $2.00 or 7.20% increase in share price.

So either Intel just won a bunch of design wins, or they're expecting Devil's Canyon to sell like hotcakes. Probably the former.

 

[Biggiesized]

Hi, guys. I'm looking for some advice.

I've been following the Broadwell debacle for months with the news of its release changing all the time.

I'm sitting on an old computer. Really old! I built my computer in January 2007. I have a Conroe/Allendale E6400 and GeForce 450 GTS (lifetime warranty replacement for my GeForce 8800 GTS 640 MB).

Suffice it to say, this computer isn't cutting it anymore. I wanted to grab Haswell last year, but Broadwell was supposed to be around the corner and then didn't materialize. When I heard earlier this year that Broadwell might make a year end release, I had hope. Now, with the Chinese leak of the slides, it looks like it Q2 2015 for sure.

Here's my dilemma. Should I grab a Devil's Canyon when it hits the streets or grab a Haswell Xeon (e. g. E3-1231v3) since I want a discrete GPU? Also, other things to consider are the EOL cycle for DDR3 (I jumped on DDR2 right when it went out of style) with the DDR4 Haswell-E later this year.

I heard Maxwell might land in July. I hope it is true because a new GPU would likely be necessary.

 

[witeken]

So either Intel just won a bunch of design wins, or they're expecting Devil's Canyon to sell like hotcakes. Probably the former.

Isn't it clear enough? Maybe this is better written. It has nothing to do with design wins or DC; Q2 has only 17 days left.

 

[Ajay]

Hi, guys. I'm looking for some advice.

I've been following the Broadwell debacle for months with the news of its release changing all the time.

I'm sitting on an old computer. Really old! I built my computer in January 2007. I have a Conroe/Allendale E6400 and GeForce 450 GTS (lifetime warranty replacement for my GeForce 8800 GTS 640 MB).

Suffice it to say, this computer isn't cutting it anymore. I wanted to grab Haswell last year, but Broadwell was supposed to be around the corner and then didn't materialize. When I heard earlier this year that Broadwell might make a year end release, I had hope. Now, with the Chinese leak of the slides, it looks like it Q2 2015 for sure.

Here's my dilemma. Should I grab a Devil's Canyon when it hits the streets or grab a Haswell Xeon (e. g. E3-1231v3) since I want a discrete GPU? Also, other things to consider are the EOL cycle for DDR3 (I jumped on DDR2 right when it went out of style) with the DDR4 Haswell-E later this year.

I heard Maxwell might land in July. I hope it is true because a new GPU would likely be necessary.

You are listing allot of parameters - this isn't really the thread for that. You may want to try the general hardware thread http://forums.anandtech.com/forumdisplay.php?f=4 or ask about each specific component in the appropriate hardware forum. Good luck!

 

[bullzz]

@Biggiesized - please start a new thread with info like budget, typical workloads, when u plan to upgrade and mention other components in ur current build to see if it can be reused