Is mainstream desktop CPU development "completed"?

[Fjodor2001]

Well, don't say that, they could push for example the I7-4470 from 658
http://hwbot.org/submission/2516255_a7154ka_cinebench___r15_core_i7_3770_658_cb
all the way up to 1329,or any number in between
http://hwbot.org/submission/2784147_dancop_cinebench___r15_core_i7_4770k_1329_cb
"just by" increasing the TDP (FX-9590 style) ,the thing is if its viable as a business model or not since people would probably avoid them like the plague.

Yes, using liquid nitrogen at 6.4 GHz on a golden sample chip. And how much do you think the TDP increased? Only 32%, I don't think so...

 

[Fjodor2001]

You know this how?

Are you being serious? Check this out:

http://en.wikipedia.org/wiki/List_of_Intel_microprocessors#The_16-bit_processors:_MCS-86_family

Intel 80286:
[...]
3–6X the performance of the 8086

Do you really think Intel could design a new CPU in 2-3 years that is 3-6x as fast as the 4790K at 1.32x TDP (=88Wx1.32=116 W TDP)?

Note: 8086 is more or less the same as 8088, except that 8086 has 16 bits data bus instead of 8 bits. So if you compare the 80286 to 8088 instead, the difference is even greater!

 

[exar333]

Fjodor2001, I think you were too quick to jump to a defensive position by dismissing kimmel's point.

Why did desktop CPUs experience the annualized compound rate of performance improvements in the past that you now note is lacking?

Intel, AMD, Texas Instruments, Via, NexGen, WinIDT, etc. all built desktop CPUs and rushed about as frantically as possible for more than a decade trying to get faster and faster models out in front of the desktop consumer.

Consumer's bought the chips, justifying the development expenses and business risks that preceded the creation of those chips.

And then guess what happened to those desktop consumers? They started to care less and less about desktop performance. More and more of them started liking the idea of mobile computing, having a lighter and longer-lasting laptop was worth their consumer dollars.

So when you ask where did the development momentum go in the desktop market, momentum that was dependent on desktop consumers wanting to buy ever-higher performing desktop CPUs, you have to ask yourself where did the consumer's dollars themselves go?

And to kimmel's credit and astute synopsis, those revenue dollars went to Apple, they went to Samsung, and Intel's own mobile product offerings.

So why would Intel, or AMD for that matter, justify sinking ever higher R&D expenses into the development of uber faster desktop processors when the markets have spoken, voted with their wallets, and are buying up smartphones and tablets and silly thin netbooks/laptops instead of desktops?

The premise of your argument in the OP appears to be that you believe the pace of advancement stagnated, and thus the consumer had no choice but to migrate to other compute platforms and spend their money on mobile products. I don't buy that, the development money follows where decision makers think the markets are headed.

In circa 1880 there were a number of companies developing the next generation of leading-edge horse drawn carriages. And then the day came where they stopped investing in developing the next best horse carriage.

Did the end of an era of horse-drawn carriages come about because the pace of development stagnated? Or did it come about because consumers abandoned that market and pursued the acquisition of a more feature-compelling product called the automobile? Forcing horse carriage companies to allocate R&D appropriately in order to best survive the transition?

I think kimmel is right on the money. The focus is not on desktop performance improvements because the consumer markets are no longer focused on it either. Consumer markets shifted, they want mobility and other features (wireless charging, faster network speeds, etc), and these companies have shifted their R&D priorities accordingly.

This.

It is pretty obvious the focus has been on efficiency. If anything, server CPUs have seen an UNPRECEDENTED increase in capacity over the past few years. Just look at the number of cores you can get per socket now, vs. just a few years ago. Combine this with a mobile efficiency focus that also scales down to a very low TDP.

 

[Fjodor2001]

This.

It is pretty obvious the focus has been on efficiency. If anything, server CPUs have seen an UNPRECEDENTED increase in capacity over the past few years. Just look at the number of cores you can get per socket now, vs. just a few years ago. Combine this with a mobile efficiency focus that also scales down to a very low TDP.

As mentioned before, all of that is well known and accurate. However it is not the topic of this thread. Read the OP which clearly says:

"Is desktop CPU development more or less "completed" as far as performance goes?

PS. Note that I'm only talking about CPU performance increase, not lower TDP or iGPU performance improvements. "

 

[ClockHound]

"Is desktop CPU development more or less "completed" as far as performance goes?"

Yup. Finally. Completed it today. Congrats all round to everyone for making the 2012(ish) level of desktop computing performance the eternal standard that our great grand children's grand children can enjoy. In previous times, jewelry and books were often handed down to the next generation. But, now our precious quad cores are the new family legacies. Well done!

Those future generations will look back at this era and chuckle, with its scifi desktop vision of 2 year performance doubling - flying cars of the 21st century.

And a big shout out to the cloud computing platforms! They have managed to emulate the user experience of a mid 1980s desktop computer. Right down to the mind-numbing latency and unresponsiveness. Amazing!

Now...moving to the slow, fast-throttling, crappy little tablet/smartyphone chips. When will their development be completed? Don't make us wait another 20 years.

 

[Ramses]

The thread topic is a bit on the simplistic side and not really warranting much discussion, that's why it's branched out a bit since there are plenty of related and ancillary topics directly related to the issue that help increase the understanding of it, such as TDP, iGPU, core count, packaging, market trends, software requirements, etc, etc. It would be better to ask why does cpu performance seem to have stalled for desktop enthusiasts lately. But the ensuing discussion/answer is basically the same either way. To answer the original post though: of course not.

 

[kimmel]

Chandler Carruth (google person) @ CPPCON
https://www.youtube.com/watch?v=fHNmRkzxHWs

His intro about performance and power is really good and how the compute landscape has changed in the last number of years. Go watch the first 5-10 minutes. (the rest of the talk is good too)

The best quote is around 7min in.

"I find it interesting that at the two extremes of computing platforms that we find convergence. When you drop to a phone, or go up to a data center all the sudden, power dominates."

Intel at a high impact level doesn't care about desktop CPUs because consumers don't care about desktop CPUs. Compute/watt is the dominating driving factor in today's space.

It is what it is. The desktop niche is not completed but the money has moved and so are the dollars to advance the product.

 

[VirtualLarry]

It is what it is. The desktop niche is not completed but the money has moved and so are the dollars to advance the product.

Why would companies spend money to out-maneuver themselves in the desktop market, when the "stupid" consumers are happy with Brazos and Atom derivatives, in a big empty case.

After all, if it's in a big case, and five years newer, it must be faster, right?

Edit: It's the companies pushing "power consumption", that resulted in that situation. Consumer desktops aren't really all that sensitive about power consumption. They handle 95W CPUs with no problems with power or cooling, and yet, companies are pushing 15W CPUs in those boxes, all the while pushing "power consumption" as the reason, while simultaneously destroying that segment of the market.

 

[mrmt]

Edit: It's the companies pushing "power consumption", that resulted in that situation. Consumer desktops aren't really all that sensitive about power consumption. They handle 95W CPUs with no problems with power or cooling, and yet, companies are pushing 15W CPUs in those boxes, all the while pushing "power consumption" as the reason, while simultaneously destroying that segment of the market.

I think most consumers don't care whether there is a 95W or a 15W chip inside this given machine. In fact, most consumers are quite happy with equivalent notebooks. Unless software companies come with a new kind of consumer application that demands high levels of CPU power, don't expect companies to spend money on the desktop market.

I was just thinking, since I'm not gaming these days it's been about a month since I touched a desktop, all my productivity and content consumption is being done on notebooks and mobile devices, and I'm not really missing the desktop at all.

 

[ehume]

Now, give me a laptop where the display unfolds to a viewable size, where the keyboard unfolds to be big enough to use -- and separates from the display so that it is close enough to me so I can type on it while keeping the display far enough away to focus on that -- then I might give up my desktop. But for us old folks, having the computer come in pieces so I can accommodate it to me -- rather than making me accommodate myself to the machine -- that's what I like. We still remember when "Do not fold, spindle or mutilate!" was an ironic statement that we would drive, and not be driven.

 

[VirtualLarry]

I think most consumers don't care whether there is a 95W or a 15W chip inside this given machine.

That was kind of my point. Most consumers don't care that much about their desktop's power-consumption, but they do care about performance, especially when their "new" PC is slower than their old.

And companies are NOT providing what the market demands, thus their market-share is shrinking. (Consumer desktop sales DOWN, PC gaming segment UP.)

 

[Ramses]

Now, give me a laptop where the display unfolds to a viewable size, where the keyboard unfolds to be big enough to use -- and separates from the display so that it is close enough to me so I can type on it while keeping the display far enough away to focus on that -- then I might give up my desktop. But for us old folks, having the computer come in pieces so I can accommodate it to me -- rather than making me accommodate myself to the machine -- that's what I like. We still remember when "Do not fold, spindle or mutilate!" was an ironic statement that we would drive, and not be driven.

I bought a usb docking station thingy for mine, one plug and it becomes a desktop. Not portable but it solves home pretty well.

 

[ninaholic37]

That was kind of my point. Most consumers don't care that much about their desktop's power-consumption, but they do care about performance, especially when their "new" PC is slower than their old.

At the libraries here, all the "Internet Computers" (around 70 at the main location alone) are all Atom D2500 desktops, with 1920x1080 screens. They are so cute and tiny! Way smaller than the Core 2 ones at the libraries where I used to live (maybe 6 times smaller?). They are a bit slower, but not overly noticeable (at least in Firefox), since they're mostly just used to surf the internet anyway. Once in a while, I hear people screaming about how they're **** slow, but people even did that with the Core 2 desktops, especially after the admin "upgraded" all those computers to Windows 7. My instinct tells me that most of them are just attention seekers or overly impatient and just enjoy complaining about everything. :biggrin:

 

[VirtualLarry]

At the libraries here, all the "Internet Computers" (around 70 at the main location alone) are all Atom D2500 desktops, with 1920x1080 screens. They are so cute and tiny! Way smaller than the Core 2 ones at the libraries where I used to live (maybe 6 times smaller?). They are a bit slower, but not overly noticeable (at least in Firefox), since they're mostly just used to surf the internet anyway. Once in a while, I hear people screaming about how they're **** slow, but people even did that with the Core 2 desktops, especially after the admin "upgraded" all those computers to Windows 7. My instinct tells me that most of them are just attention seekers or overly impatient and just enjoy complaining about everything. :biggrin:

Or maybe it's the pig of software packages that they run on those library computers. I don't think I've ever seen one that I would label "fast". (I run very lean + mean OS setups at home.)

 

[TidusZ]

Sorry, not seeing much revolution there in terms of CPU performance increase, given the time that has passed between those CPU models.

not to mention the 2600k will o/c from 3.4 to 5 ghz while the new 4 ghz processor wont hit 5 ghz

 

[kimmel]

Edit: It's the companies pushing "power consumption", that resulted in that situation. Consumer desktops aren't really all that sensitive about power consumption. They handle 95W CPUs with no problems with power or cooling, and yet, companies are pushing 15W CPUs in those boxes, all the while pushing "power consumption" as the reason, while simultaneously destroying that segment of the market.

Maybe you didn't actually watch the first part of the video. In the datacenter (cloud) they care immensely about power. In phones/tablets/laptops they care about power. Supercomputing and HPC care massively about power. Anyone who is driving growth in unit volume and profit cares about power. Only in the desktop arena do we have consumers who don't generally care about power consumption and do care about performance. Look at unit volume growth in phones, tablets, and servers, look at profit margin compared to servers. Now you know why desktop market can't drive large investments.

You see that many people are still buying desktops and maybe even desktop is growing slowly, but that data in a vacuum is almost meaningless. Likewise, much of the growth of desktop that companies report is in AIO computers, which again still care about power.

 

[StrangerGuy]

I think most consumers don't care whether there is a 95W or a 15W chip inside this given machine. In fact, most consumers are quite happy with equivalent notebooks. Unless software companies come with a new kind of consumer application that demands high levels of CPU power, don't expect companies to spend money on the desktop market.

I was just thinking, since I'm not gaming these days it's been about a month since I touched a desktop, all my productivity and content consumption is being done on notebooks and mobile devices, and I'm not really missing the desktop at all.

The moment a power cord is plugged, the CPU power consumption becomes a meaningless statistic for the average consumer PC. A 95W Intel CPU is really a 5W chip at idle virtually 100% of the time, dwarfed by display and the rest of the components.

 

[VirtualLarry]

You see that many people are still buying desktops and maybe even desktop is growing slowly, but that data in a vacuum is almost meaningless. Likewise, much of the growth of desktop that companies report is in AIO computers, which again still care about power.

No-one I know is concerned about power consumption in a desktop (Edit: Besides my own adventures with low-power rigs, although I don't really care about desktop power-consumption anymore myself either, now that my G3258 + 7950 rig is drawing the same wattage almost as my older OCecd Q9300 + GT630), much less an AIO. However, I do know people that are concerned about performance. One of my clients bought a very nice (IMHO) 20" AIO off of me, cheaply, and it had an E1-2500 APU in it, which is admittedly slightly slow. They were so unhappy, they returned it. Which was a fairly ridiculous notion, given how cheaply I sold it.

Again, these companies are pushing power consumption (or lack thereof), OVER performance. And that's not what the desktop / AIO-buying population cares about. Sure, they might say that they do, if asked, but then when left with the PC that is slower than their old one, I think that they would say otherwise, once they realize the implications of choosing a "power saving" (AKA SLOW) PC.

Edit: I'm not trying to argue that datacenter, or tablet/phone users don't care about power consumption, don't get me wrong. Just that it isn't the primary focus of the vast majority of "big box" desktop users. (Mini-PC users might care, because more power usage translates directly into more fan noise.)

Edit: I will also add, "desktop" used to get "server" leftovers. With the advent of Intel and AMD's "low-power" line of CPU/APUs, it seems that for a large swath of the market, "desktop" now got "tablet" leftovers, which are significantly slower chips. All because the OEMs and the industry is greedy. Sure, they label that greed "power efficiency", but that's what it is.

 

[cbn]

Passmark scores of some CPUs:

E2140 (the slowest consumer desktop Core 2 duo ever made): 878 cpu marks http://www.cpubenchmark.net/cpu.php?cpu=Intel+Pentium+E2140+@+1.60GHz&id=1133

E1-2500: 892 cpu marks http://www.cpubenchmark.net/cpu.php?cpu=AMD+E1-2500+APU

Sempron 2650: 869 cpu marks http://www.cpubenchmark.net/cpu.php?cpu=AMD+Sempron+2650+APU+with+Radeon+R3&id=2214

I think AMD could do a better job (especially on the AM1 since it has a 25 watt TDP).

In contrast, Intel keeps the single thread performance of the J1800 Celeron almost the same as the J2900 Pentium. AMD should at least do the same with the dual core AM1 Sempron and the top quad core AM1 Athlon.

 

[ninaholic37]

Or maybe it's the pig of software packages that they run on those library computers. I don't think I've ever seen one that I would label "fast". (I run very lean + mean OS setups at home.)

They're pretty clean, for what I can tell (looking at the CPU and RAM usage when idle). Lots of programs and browser addons installed by default on these Atom D2500s though, and they still run ok to me anyway, but I guess they're free to use so it's kind of hard to complain. Some people don't even realize the Firefox icon though, and just immediately click on the IE or Chrome ones. Suckers :sneaky:

 

[Fjodor2001]

not to mention the 2600k will o/c from 3.4 to 5 ghz while the new 4 ghz processor wont hit 5 ghz

Yes, from an OC perspective it's even worse.

 

[ShintaiDK]

Funny how 5Ghz seems to be some kind of OC standard now for SB. Just a shame it wasnt so, it was very rare.

 

[Revolution 11]

I would argue that there is still quite a bit of innovation going on in desktop CPUs. But as Idontcare pointed out, the R&D is following consumer demands. Consumers want more convenience, more usability, even in their desktop products.

Look at the upcoming standards/connectors for M.2, SATA Express, and the increasing availability of USB 3.0/3.1 ports. Boot support for NVME SSDs is also coming in future CPUs/motherboards.

Rapidly increasing access to fast and plentiful I/O is both more appealing and useful to consumers than chasing heat/power/cost limited options to increase clock speeds or going down the path of diminishing returns for IPC gains. Intel will continue to do both, but only when it makes sense and there is enough demand to pay for it.

It doesn't help that AMD is completely out of the race and unlikely to re-join in the near future.

 

[jpiniero]

I want Nonvolatile memory dammit. And onpackage for high speed.

 

[Fjodor2001]

I would argue that there is still quite a bit of innovation going on in desktop CPUs. But as Idontcare pointed out, the R&D is following consumer demands. Consumers want more convenience, more usability, even in their desktop products.

Look at the upcoming standards/connectors for M.2, SATA Express, and the increasing availability of USB 3.0/3.1 ports. Boot support for NVME SSDs is also coming in future CPUs/motherboards.

Rapidly increasing access to fast and plentiful I/O is both more appealing and useful to consumers than chasing heat/power/cost limited options to increase clock speeds or going down the path of diminishing returns for IPC gains. Intel will continue to do both, but only when it makes sense and there is enough demand to pay for it.

It doesn't help that AMD is completely out of the race and unlikely to re-join in the near future.

Sure, all of those out of topic aspects are valid. But as far as desktop CPU performance goes, would you consider that development to be more or less "completed"?