PC Client shipments in free fall Q1.

[StrangerGuy]

Words

A lot of people are really bad looking at costs incurred beyond the initial sticker price and cashing in on resale value when going is still good.

If I had an OCed X58 platform in 2009 it would have easily ate at least 100W more at idle compared to Sandy Bridge which will cost me an additional $40 per year to run, 8 hours a day.

Just by going 2500K in 2011 would have saved me $200 in electricity costs by 2016 and still get significantly better CPU performance and 2 years more of warranty to boot. If I sold the X58/920 at a rather low $200 back in 2011, I would have spent nothing in the long term to upgrade. If the resale value was even more, I'm actually getting PAID to upgrade. Try doing the same with X58 in 2016.

Sidetrack: There was also a time when used S939 A64 X2s were going for $200+ when a new C2D/mobo/2GB RAM that runs circles around it can be had at the same price. You would be downright dumb if you had those chips and not cashed in.

Same for getting 4790K vs 4690K, 6700K vs 6600K and the latest iPhone 16GB vs 64/128GB, if the differential in ~$100 pricing can still be maintained in the resale value why would you not want to enjoy the benefits of the pricier model? Don't be penny-wise, pound-foolish.

 

[Zodiark1593]

If I was in your position and wasn't happy with current prices of GPUs, I would strongly consider a 6-core Xeon. A lot of these can be purchased for $100 USD and with a solid cooler can overclock to 4.4-4.6Ghz.

Intel Xeon Processor X5670 Six Cores 2.93 GHz/12M/6.40GT LGA1366 SLBV7 CPU = $89 USD

At the very least, you will get 2 more cores (and I am going to assume you want that) and higher IPC and clock speed than your i7 920 (again I am assuming since I don't know if your 920 is clocked at 4.2ghz+).

If you sell your 920 for even $50, that's a very cost effective upgrade. At least take full advantage of your X58 chipset/mobo. Burpo should be able to help you pick the best Xeon CPU and max it out. :thumbsup:



I had to dig deep into some old reviews that fairly compared an i7 Nehalem 920/930 overclocked against an overclocked i7 2600K. I remember that Sandy was my favourite architecture from Intel since I sold my i7 860 and I remember why. The upgrade from Nehalem to Sandy alone was HUGE.

Core i7 930 – 3.9GHz (stock 2.8GHz)
Core i5 2400 – 3.8GHz (stock 3.1GHz)
Core i5 2500K – 4.4GHz (stock 3.3GHz)
Core i7 2600K – 4.8GHz (stock 3.4GHz)


Pretty much anyone with a Core i5 2500K and below would benefit greatly from an upgrade to any modern Intel i7.

And if someone is a gamer with an AMD card, R9 390 or higher class, an i5 from SB era is already a bottleneck, which is to say nothing of the i7 920, an architecture with a solid 16-17% lower IPC vs. Sandy.

https://www.youtube.com/watch?v=frNjT5R5XI4

A 4Ghz overclocked i7 920 is only as fast as a stock i7 2600K.

Here is a stock i7 6700K vs. a stock i7 2600K. Absolute slaughter in CPU demanding games.
https://www.youtube.com/watch?v=VDo-j00vUtw

The reason a lot of Sandy owners are holding on is because some of them have i7 2600K OC to 4.5-5Ghz. That's totally understandable. Everyone else with an i5 2500K and below, including an i7 920, those CPUs are a major bottleneck now.

Oh sure, the Sandy beats down the Nehalem by 30 or so percent, but it's nowhere near the beat down Core gave the Pentium 4/D. In just the span of a night, the Core 2 Duo made the Pentium D feel like a slug by comparison.

In all seriousness, the accumulated architecture upgrades combined lead to some pretty impressive gains over Sandy, though I don't feel like I could call Sandy "Slow". Maybe just a downgrade to "Adequate" would be quite fair.

 

[intangir]

Oops, you are right, I 100% forgot about that CPU. Kind of sad looking at wiki, it's like Intel made great progress from 1, to 2, to 4 cores in 2007 and 2008, but started putting on the breaks in 2009.

Even the i7-920 was release late 2008 for $284, but here we are almost 8 years later and double the core count cost people about $1000 USD.... And Intel wonders why I am not giving them my money?

The majority of the desktop and mobile market gains no performance going above 4 cores, and it actually hurts power efficiency. Only niche applications benefit from it, and it's going to cost you niche prices. Back in 2006, Dadi Perlmutter, Executive Vice President and General Manager of Intel Corporation's Intel Architecture Group, said this:

https://slashdot.org/story/06/07/27/1753255/intel---market-doesnt-need-eight-cores

Two are enough for now, four will be mainstream in three years and eight is something the desktop market does not need. ... I want everybody to go from a frequency world to a number-of-cores-world. But especially in the client space, we have to be very careful with overloading the market with a number of cores and see what is useful.

Ten years later, it is still true. There are rumors that Cannonlake in 2017 might go above 4 cores, but they're only rumors. There's just no widespread need for it, as you can see from AMD offering 8-core mainstream chips for years and years now, without hurting Intel's business one bit.

And looking into the foreseeable future, people who know what they're talking about believe there will never be a need for more than four cores.

http://www.realworldtech.com/forum/?threadid=146066&curpostid=146227
http://www.realworldtech.com/forum/?threadid=146066&curpostid=146238

So give up on parallelism already. It's not going to happen. End users are fine with roughly on the order of four cores, and you can't fit any more anyway without using too much energy to be practical in that space. And nobody sane would make the cores smaller and weaker in order to fit more of them - the only reason to make them smaller and weaker is because you want to go even further down in power use, so you'd still not have lots of those weak cores.

So the whole argument that people should parallelise their code is fundamentally flawed. It rests on incorrect assumptions. It's a fad that has been going on too long.

Parallel code makes sense in the few cases I mentioned, where we already largely have it covered, because in the server space, people have been parallel for a long time.

It does not necessarily make sense elsewhere. Even in completely new areas that we don't do today because you cant' afford it. If you want to do low-power ubiquotous computer vision etc, I can pretty much guarantee that you're not going to do it with code on a GP CPU. You're likely not even going to do it on a GPU because even that is too expensive (power wise), but with specialized hardware, probably based on some neural network model.

Give it up. The whole "parallel computing is the future" is a bunch of crock.

On the client side, there are certainly still workstation loads etc that can use 16 cores, and I guess graphics professionals will be able to do their photoshop and video editing faster. But that's a pretty small market in the big picture. There's a reason why desktops are actually shrinking.

Some things are cheap and easy. But those are done already. Building stuff in parallel? Trivial - except even there people tend to have build tools and Makefiles etc that make it not work that well in many cases. So even something really simple like a software workstation often doesn't scale that well (I'm happy to say that the kernel build scales exceptionally well, but most projects don't have tens of thousands of files and lots of effort on the build system.

The "more parallelism" people aren't new. This has been debated for decades, and the extreme parallelism people have been wrong for decades. I don't see anything that has really changed things. And if anything, scaling limits will take some of their arguments away.

I work on a project where we're doing extreme scaling. We're proud of it. But I also see the pain. For us, it makes sense to spend the effort. I don't see that being true in 99% of all cases.

Sorry to say, but if you're looking for more cores, there's just no justifiable reason to make a mass-market chip with that.

 

[StrangerGuy]

Oh sure, the Sandy beats down the Nehalem by 30 or so percent, but it's nowhere near the beat down Core gave the Pentium 4/D. In just the span of a night, the Core 2 Duo made the Pentium D feel like a slug by comparison.

In all seriousness, the accumulated architecture upgrades combined lead to some pretty impressive gains over Sandy, though I don't feel like I could call Sandy "Slow". Maybe just a downgrade to "Adequate" would be quite fair.

Sandy was remarkable not because of the huge IPC and clock gains over Nehalem, but also the first CPU where you can OC without wasting electricity at high clocks and voltages at idle, the chipsets themselves consume very little power, plus the mobos were dirt cheap compared to X58.

 

[RussianSensation]

I paid $550 CAD for both my i7-930 and my x58a-ud5 (dual lan, dual pci-e, and 10 sata ports) in 2010. Today getting just a quad core Intel (6700k) would cost me $500 CAD alone without a mobo.

OK, my bad then but then you didn't get the i7 930 when it was cutting edge tech. Frankly, i7 930 was always overpriced as the entire i7 920->960 stacked overclocked more or less the same. Besides the point, it's rather odd that you waited what 1.5 years to get an i7-930, in the process skipping the much cheaper 2008 i7 920, and then yet again cheaper September 2009 i7 860 but then decided not to wait for what was a huge performance increase in the January 2011 i7 2600K. You seem like a patient guy now having held on since 2010 but back then you couldn't wait what 1 more year for i7 Sandy? 2600K to 4.5-4.8Ghz absolutely demolishes an i7-930 @ 4Ghz in the context of 12 or so months it would have taken in your case. Just strange that you waited that long to get X58, but yet now you are calling X99 outdated. X99 chipset isn't actually outdated as boards have M.2/USB 3.1, and newer X99 updated boards will have all the latest features of Z170 too.

I got no issues shelling out for HEDT when I see value. When I went for my current build I went from a dual core AMD to a quad core Intel. Not only was the IPC about 2x in some cases, but the number of cores was 2x, so in effect in certain apps I was getting 3-4x performance for a $300 CAD CPU.

I went from a 1998 Pentium II 233mhz MMX (never overclocked) to a 2001 Athlon XP1600+ overclocked to 1800+ speeds, or a 7.73X increase in performance in only 3 years. Then in 2003, just 2 years after my Athlon XP1600+, I got a Pentium 4 2.6Ghz "C" that overclocked to 3.2Ghz. That means from 1998 to 2003, or in 5 years, my CPU performance increased 13.73X. By 2006, I had E6400 @ 3.4Ghz (so more than 2X Pentium 4 "C" 3.2Ghz) and by August 2007, I upgraded that to Q6600 @ 3.4Ghz (double the cores). If I apply straight up linear logic, that means 13.73X * 2X for C2D and another 2X for 2 more cores = 54.92X from Fall 1998 to August 2007.

If I were to apply the same high standards, even for a small time-span of 1998 to 2003 where CPU speed for me went up almost 14X, I might as well never upgrade again for 30 more years then. See how flawed this logic is for CPUs? Unlike GPUs, CPUs perform work in a serial manner and you cannot just scale performance across 3840 Intel cores (Pascal CUDA cores).

Even looking at 6700k vs 930 doesn't even give me 2x improvements in certain task (like 7zip compression).

I really want HEDT, ram upgrade alone would is starting to sound good, but I am also crazy stubborn. I have been hoping broadwell-e would be the solution to all my problems (but the rumors of $1500 CPU cast doubt into that dream).

Why is it all or nothing for you? $1500 CAD for a 8/10-core or bust? So you'd rather use the i7 930 instead of finding something in-between with the upcoming Broadwell 6-core i7 6800/6820K? I mean I want a Porsche GT3 or a Cayman GT4 but realistically speaking, I also don't want to sacrifice the rest of my life so I have to kinda settle for a Camaro SS, or something. It doesn't mean it's either Porsche GT3 or Honda Civic (nothing wrong with a Honda Civic but you get the point).

I did not mean viable $1000 CPU. I meant if I wanted a 8 core CPU it would cost about $1000 USD. In reality a 5960x would cost me about $1500 CAD....

....
I need more cores, I want 8 or 10 cores. For a while ive been hoping Intel doesnt do anything stupid with the price, because a nice 8 or 10 core BW-E is what I want. I could even live with a 6 core if its dirt cheap.

Well I told you to consider X58 6-core Xeons then for $90-100 USD but you don't want that either, so you don't want to a get a Camaro SS/Mustang GT350 because you want a Porsche GT3/Nissan GTR but you don't want to spend the $ and so you end up driving the Honda Civic.

I find it VERY obsolete. I just don't like what Intel is offering.

What exactly are you running that needs 8-10 modern IPC cores but you seem to have been accepting running 2008-2009 Nehalem/Lynnfield 4-core tech for 6 years now?

There is no 6800 and the only 6820k is mobile, so I assume you are talking about 6700/6700k?

http://www.guru3d.com/news-story/intel-core-i7-6950x-broadwell-e-has-10-cores-and-20-threads.html

I set a 'bar' to reach for me to upgrade, and it seems each year either Intel releases something too weak, releases something too expensive, or this year looks like it will be the Canadian dollar that crushes my broadwell-e dreams (along with Intel probably introducing a $1500 CPU).

Intel prices have been pretty steady and they have actually lowered the price of entry for the 6-core in August of 2014 (!).

i7 4930K cost $583 and i7 5820K was $389 US.

I really hope broadwell-e will work out, if not maybe even a nice 8 core (or more) Zen...

If you prioritize more cores over single threaded throughput, perhaps wait for Zen and go from there. They should have an 8-core 16 thread offering but I doubt it'll be cheaper than $400-500 USD.

..... If I sold the X58/920 at a rather low $200 back in 2011, I would have spent nothing in the long term to upgrade. If the resale value was even more, I'm actually getting PAID to upgrade. Try doing the same with X58 in 2016.

Sidetrack: There was also a time when used S939 A64 X2s were going for $200+ when a new C2D/mobo/2GB RAM that runs circles around it can be had at the same price. You would be downright dumb if you had those chips and not cashed in.

Ya, you are right. There are certain times when you almost have to resell old parts or their value will start moving closer to $50-100 at most. Also, in situations like you described where you can sell old tech for inflated prices, you go for it. At the end of the day, sometimes it works out better to buy $700 worth of parts, sell them for $450 in 3 years (loss of $250), then buy $700 worth of new parts and sell them again for $450 in 3 years (another loss of $250). In 6 years, you lose $500 of real $/value and over 6 years have a cutting edge system most of that time.

The alternative to that is spending $700 US on new parts and holding on to them for 6 years by which point let's say they are worth $200. Your real cost of ownership, or real loss in value, is still the same $500.

Once I grasped the concept of Total Cost of Ownership (TCO), upgrading actually makes more sense to me rather than not upgrading and waiting and waiting. In other words, there comes a time where it's more optimal to sell and upgrade rather than wait.

Same for getting 4790K vs 4690K, 6700K vs 6600K and the latest iPhone 16GB vs 64/128GB, if the differential in ~$100 pricing can still be maintained in the resale value why would you not want to enjoy the benefits of the pricier model? Don't be penny-wise, pound-foolish.

Exact same concept as above -- TCO. It removes all headaches.

Removing all other factors, let's say i7 6700K cost $350 and i5 6600K cost $230, let's say I can get 50% value from both in 3 years, I get:

($350 - $230) * 50% = $60 extra to own an i7 6700K over 3 years or $20/year. It doesn't get any simpler than that. :thumbsup:

Oh sure, the Sandy beats down the Nehalem by 30 or so percent, but it's nowhere near the beat down Core gave the Pentium 4/D. In just the span of a night, the Core 2 Duo made the Pentium D feel like a slug by comparison.

True but C2D was the greatest single leap Intel has ever done if we exclude mobile Merom. That's not even fair to compare.

In all seriousness, the accumulated architecture upgrades combined lead to some pretty impressive gains over Sandy, though I don't feel like I could call Sandy "Slow". Maybe just a downgrade to "Adequate" would be quite fair.

I am not calling Sandy slow per say. I was saying anything at i5 2500K and below can be much slower in certain apps. But again, I can wait another 2-3 years until my 2500K, mobo and RAM would have be worth $100 altogether. If I can sell them in 2016 for $200, I get a new fast system I use for 2-3 years. Alternatively, I have to wait 2-3 more years and then have to spend an extra $100 to get what 15-20% faster CPU in Icelake? My point is at some point or another, the opportunity cost of waiting starts to set in. Once you calculate the TCO and weigh in the resale value, at one point or another it will make sense to dump old parts. This will happen for people with i7 3930Ks and i7 3770K and later with i7 5820K/6700K, etc.

The reason it doesn't happen for the average person is because they don't really think of TCO. They'd rather buy a $1000 PC and let it become completely obsolete in 6 years. I'd rather buy $700 worth of platform parts, keep my core components like PSU, monitor, headphones, and sell those parts in 3-4 years, and buy new $700 parts. I already outlined above why financially this actually makes more sense even if the increases in CPU speed are nowhere near Pentium 4/D -> C2D -> Nehalem -> Sandy.

Sandy was remarkable not because of the huge IPC and clock gains over Nehalem, but also the first CPU where you can OC without wasting electricity at high clocks and voltages at idle, the chipsets themselves consume very little power, plus the mobos were dirt cheap compared to X58.

Another thing is HT on Sandy worked even better than on Nehalem so you got an even bigger benefit with i7 2600K vs. i5 2500K than i5 750/760 vs. i7 860. More importantly, with Sandy, you could achieve 4.4-4.5Ghz on $20 air coolers and 4.7-4.9Ghz on $50 air coolers. Overclocking headroom has decreased overtime and it took until i7 4790K and i7 6700K to barely catch up. That means Sandy's huge overclocking increase mitigated some of its IPC disadvantage. My i7 860 basically maxed out at 3.9Ghz and i7 2500k/2600K went to 4.8Ghz a lot of the time. That's a 23% increase in frequency and a 16-17% increase in IPC. If today Skylake i7 6700K could overclock 23% higher than i7 4790K, that would still be not even as good as i7 2600k vs. i7 860/920 because Sandy brought a bigger increase in IPC than Skylake did over Haswell. That is how amazing Sandy really was.

I hate that Intel sells a dinky 4-core i7 6700K with sub-125mm2 die with mostly useless for gaming iGPU for $350 but I cannot argue with the numbers.

i7-6700K@4.6Ghz vs i7-4790K@4.9Ghz in 10 Games (Fury X)
https://www.youtube.com/watch?v=f5lfMogcrPU

At least Intel has been improving their CPUs. With AMD, if you overclocked your FX8150 to 4.6Ghz nearly half a decade ago, well you are lucky enough to get anything 12-13% faster as of now.

 

[StrangerGuy]

Total Cost of Ownership (TCO)

Well to be fair, there probably won't be another time where a successive platform is going to perform better for much cheaper like SB vs Nehalem.

But it's still baffling to see people how people obsessing over the initial price sunk cost they will never recover, hold onto aging gear until it becomes worthless while being slow during this entire period of time, than to upgrade for free with the existing resale value of old parts when this option actually works.

 

[coercitiv]

But it's still baffling to see people how people obsessing over the initial price sunk cost they will never recover, hold onto aging gear until it becomes worthless while being slow during this entire period of time, than to upgrade for free with the existing resale value of old parts when this option actually works.

The upgrade isn't free: it requires time and knowledge (== more time) in order to make a proper resale. The same time can be used to make money. Money get new computer parts in seconds

That's not to say I don't agree with you and RS on this, someone complaining about lack of revolutionary performance increase as incentive to upgrade while not taking advantage of market opportunities to upgrade for little/no cost is clearly doing it wrong.

 

[frozentundra123456]

I said it before, and ill say it again, I feel intel is not giving people a reason to upgrade... Same old 4 cores (4 cores have been out since 2009), barely any IPC gain per generation, and OC potential stays the same or is made worse.... There is nothing for me to sink my teeth into unless I want to pay high prices for HEDP using old architecture.

Each generation I think 'nah, ill pass, Intel can do better'.

First, I adamantly believe Intel should introduce a mainstream hex core, and make the HEDT platform 8 and 10 cores, at maybe 500 and 800 dollars.

However, it simply is not realistic for users to expect to continue the kind of generation to generation gains in clockspeed and IPC that we used to see. The easy gains have been gotten long ago, and it is becoming more and more difficult to increase performance. One is certainly free to set whatever criteria he wants to upgrade, but it is going to be a long, long time between upgrades if one insists on the kind of gains we used to see.

 

[DidelisDiskas]

I don't see PC's going anywhere until either computing in small devices has enough power in games to make the pc obsolete, or the number of gamers starts to decrease (currently only increasing).

http://www.pcgamer.com/there-are-711-million-pc-gamers-in-the-world-today-says-intel/

“We're really focused around PC gaming and enthusiasts. This is the one area of PCs that has kept growing,” Intel's Lisa Graff told me at the show. “These are our most loyal customers: PC gamers. They want as much performance as we can throw at them. We're going to bring Intel's best technologies to bear for PC gaming.”

Steam alone has 130 million users and constantly growing. These numbers are not insignificant.

 

[Denly]

Go grab any adult -- ask them, would you like a nice dell desktop computer with a monitor, mouse, keyboard, speakers, etc, set up in their home office with wires everywhere? Or would they like a dell XPS 13 laptop and wireless speaker? They will pick the laptop.

I know many adult prefer to sit comfy in front of their PC with a nice 22-24" screen or two with a nice keyboard and mouse to paybill, email, shopping, social network, read news, photo work, office work, reddit, movie and game(light). It is the personal quite time that they need to finish the day. Myself included.

 

[2is]

I'm an adult... I rarely use my laptop unless i'm on the go. If I'm home and I need to do a quick email check or look something up I use my phone most of the time. If I need to do any length of computing, I would much prefer sitting in the comfort of a nice chair in front of my desk with a pair of large monitors and a full size keyboard and mouse vs putting a laptop on my lap while sitting on the couch contributing to my already bad posture.

 

[nerp]

I know many adult prefer to sit comfy in front of their PC with a nice 22-24" screen or two with a nice keyboard and mouse to paybill, email, shopping, social network, read news, photo work, office work, reddit, movie and game(light). It is the personal quite time that they need to finish the day. Myself included.

You are the shrinking minority. The numbers play this out. Tag along with a cable installer, plumber, electrician for a few weeks. You do not see desktop computers anymore. You might see an old Dell, but rarely a modern machine.

 

[ComputerGuy2006]

The majority of the desktop and mobile market gains no performance going above 4 cores, and it actually hurts power efficiency. Only niche applications benefit from it, and it's going to cost you niche prices. Back in 2006, Dadi Perlmutter, Executive Vice President and General Manager of Intel Corporation's Intel Architecture Group, said this:

https://slashdot.org/story/06/07/27/1753255/intel---market-doesnt-need-eight-cores


Ten years later, it is still true. There are rumors that Cannonlake in 2017 might go above 4 cores, but they're only rumors. There's just no widespread need for it, as you can see from AMD offering 8-core mainstream chips for years and years now, without hurting Intel's business one bit.

And looking into the foreseeable future, people who know what they're talking about believe there will never be a need for more than four cores.

http://www.realworldtech.com/forum/?threadid=146066&curpostid=146227
http://www.realworldtech.com/forum/?threadid=146066&curpostid=146238


Sorry to say, but if you're looking for more cores, there's just no justifiable reason to make a mass-market chip with that.

Linus says a lot of things people don't agree with. Even in his own post he says parallelism works on the server space because its been there for a while.

Also faster and faster single cores do help, but certain task scale almost forever with more cores, and its those task that tend to be the slowest (encoding, rending, AI, compression, computation, emulation, etc).

OK, my bad then but then you didn't get the i7 930 when it was cutting edge tech. Frankly, i7 930 was always overpriced as the entire i7 920->960 stacked overclocked more or less the same. Besides the point, it's rather odd that you waited what 1.5 years to get an i7-930, in the process skipping the much cheaper 2008 i7 920, and then yet again cheaper September 2009 i7 860 but then decided not to wait for what was a huge performance increase in the January 2011 i7 2600K. You seem like a patient guy now having held on since 2010 but back then you couldn't wait what 1 more year for i7 Sandy? 2600K to 4.5-4.8Ghz absolutely demolishes an i7-930 @ 4Ghz in the context of 12 or so months it would have taken in your case. Just strange that you waited that long to get X58, but yet now you are calling X99 outdated. X99 chipset isn't actually outdated as boards have M.2/USB 3.1, and newer X99 updated boards will have all the latest features of Z170 too.

Back then I was just finishing school (or just got out) so I had limited funds and was in debt. I had wanted a PC upgrade for a while, and figured x58 would last me a while. I think I was also hoping more cores would come thereby giving me cheap'ish 6 or 8 core drop in upgrade replacements. I also liked the idea of triple channel ram with two pci-e slots (and the dual lan ports are/were awesome).

If I were to apply the same high standards, even for a small time-span of 1998 to 2003 where CPU speed for me went up almost 14X, I might as well never upgrade again for 30 more years then. See how flawed this logic is for CPUs? Unlike GPUs, CPUs perform work in a serial manner and you cannot just scale performance across 3840 Intel cores (Pascal CUDA cores).

The thing you need to remember is, I never planned to take it this far. Like I said previously, I felt I was always boxed out of my next upgrade (and had money ready for a long time now for another upgrade). I always felt Intel either did not give me a decently priced 6 core, or the upgrade in IPC was not worth it for another 4 core. These days the IPC improvement is massive, and I agree with you, upgrading would be highly beneficial.... Skylake is too expensive when I factor in its only 4 cores, broadwell-e is is around the corner, and the Canadian dollar.

Why is it all or nothing for you? $1500 CAD for a 8/10-core or bust? So you'd rather use the i7 930 instead of finding something in-between with the upcoming Broadwell 6-core i7 6800/6820K? I mean I want a Porsche GT3 or a Cayman GT4 but realistically speaking, I also don't want to sacrifice the rest of my life so I have to kinda settle for a Camaro SS, or something. It doesn't mean it's either Porsche GT3 or Honda Civic (nothing wrong with a Honda Civic but you get the point).

Well its not so much all nothing, but my fear is that Intel will continue with crazy prices which will once again push me away from upgrade another generation.

Well I told you to consider X58 6-core Xeons then for $90-100 USD but you don't want that either

Actually that is a good point. When you said that I was thinking you were suggestion I buy a new mobo, but looks like that CPU might be a drop in replacement in my current mobo... Might be a good upgrade. especially if I end up turning this computer into a server later...

What exactly are you running that needs 8-10 modern IPC cores but you seem to have been accepting running 2008-2009 Nehalem/Lynnfield 4-core tech for 6 years now?

Everything I need more cores works right now with 4 cores, its just slow... I do compression a lot, so faster compression would be nice (I also have plans on writing a program that will do a lot more compression). I also been playing around with stockfish (a chess AI), I love to analyze many lines in very deep dept, so more cores would help there. From time to time I also mess around with x264 encoding and sometimes even photoshop (I plan on learning illustrator soon).... On top of this veryacrypt is a little slow at triple encryption on my current 4 core 930.... And all this is without talking about me messing around with vmware...

 

[RussianSensation]

You are the shrinking minority. The numbers play this out. Tag along with a cable installer, plumber, electrician for a few weeks. You do not see desktop computers anymore. You might see an old Dell, but rarely a modern machine.

Facts prove you wrong. While it is true that the popularity of overall unit volume for desktops is decreasing, Intel is recording record i7 sales quarter after quarter. This means the types of people who are interested in desktop PCs are gravitating towards cutting edge performance. Just because junky $300-700 OEM desktops are no longer selling as well doesn't mean anything. The market we care about -- the enthusiast CPU and GPU market segments -- are very strong and both are experiencing growth despite the overall decline in PC shipments.

This would be akin to high end headphone sales growing or after-market DIY car upgrades growing despite an overall decline in headphone or car sales across the general population. Since this is an enthusiast tech forum, we don't generally care for low end stuff not selling well as we are not the target market.

In practice it means if there are 100 people on Core 2 Duos and 10 people on Core i7 Nehalem, if 20 of those upgrade to Broadwell-E/Skylake i7, and 50 buy Celerons/AMD, it shows that the desktop market is shrinking but the high-end enthusiast market has just doubled from 10 to 20 sales of the past generation.

All it means is people who appreciate the desktop know it and will invest into this specialized device. As long as AMD, Intel, NV can pump out new cutting edge enthusiast tech, I really couldn't care less if the desktop market shrinks even more. It's like the majority of people do not buy a Porsche GT4, Cayman S or say Camaro SS. Instead they drive average Joe cars like Civics, Camries and Accords. Even if 100 million of those stale cars are sold but Porsche can sustain pumping out just 2500 GT4s, that's what matters to me. I view the PC market the same way and there are no signs that Intel, AMD and NV are about to close shop. That means on the whole, unless you are a shareholder of Intel, who cares that the average Joe doesn't want i7 6700K or i7 6950K because there are now more people than ever interested in those CPUs which is enough to get Intel RECORD i7 sales growth.

As mentioned already, 130+ million Steam gamers won't be cruising on outdated Core and Nehalem architectures for much longer. Sooner or later, tens of millions of them will open their wallets and get Skylake, Icelake, Tigerlake because once PS5/XB2 come out in 2019/2020, their current 2006-2010 CPUs will be not even meet the minimum spec for next gen games. Right now, a lot of these outdated CPUs can still game because the CPUs in PS4/XB1 are weak. Popular games like League of Legends, Starcraft 2, WOW, Dota 2, Counter Strike, Rocket League, etc. are still built around outdated/non-demanding game engines. Once successors of these games launch, or new popular games launch in 2019 and beyond using next gen DX12, there will be waves of gamers upgrading. There will be even more demand due to VR.

If the overall PC market was comprised of 325-350 million shipments and that shrinks to 225-250 million shipments but the high margin/expensive parts more than offset the declining sales of low-end junk, then the enthusiast PC market would actually be healthier than ever despite an overall collapse of PC shipments. For gamers and high-end users, it's way better if there are 75 million i5/i7, Zen 6-8 cores sold per annum in a 225 million market than if there are 50 million of the same CPUs sold in a 350 million units market. With AMD, NV and Intel all acknowledging that the high-end/enthusiast segments are strong, they will actually spend even more effort and R&D on high end parts, making them faster. This is perfect since I don't want $50 G4400 CPUs. Give me $350-500 6-10 cores and I'll keep upgrading. Now how many of these crappy $50 CPU rigs does Intel have to sell to make as much money as selling a single $500 Broadwell-E and an expensive $250 X99 board? They probably need to find 4-5 customers, if not more. Alternatively, by targeting us strategically, once they meet our needs, they only need to generate 1 sale. It's not surprising then Intel/AMD/NV are starting to realize it's better to double-down on markets that are willing to pay than cater to consumers who won't buy a $50 CPU even if it's 5X faster than their C2D since they'd rather use a smartphone/tablet.

 

[PhIlLy ChEeSe]

When given a choice between an useless by its own i3 chip and a more than good enough smartphone at the same price that also has a ton more real world utility than a PC, it's pretty clear which options the poor folks would pick.

I got an Obama phone

 

[Ketchup]

I think my parents are a good example of this. Most of the parts in the pc are from 06. I put Windows 7 on it when they got it, and it is now on Windows 10. But about all it does now is hold data for mac devices. It still plays many of the games our kids like when they come over, but the folks are happy with their phones and tables.

I don't look at this as Intel losing money, just moving it to different departments. As for AMD, I think they would be headed downhill either way.

 

[Denly]

You are the shrinking minority. The numbers play this out. Tag along with a cable installer, plumber, electrician for a few weeks. You do not see desktop computers anymore. You might see an old Dell, but rarely a modern machine.

Visit any small mid business and you will see 99% are on PC with 2-3 screen, a lot of them are still on C2D or 1st/2nd gen core I. They're still good enough for them with the legacy program they depends on.

Same apply to home user, they just don't have a NEED to upgrade YET. But they will soon.

Mobile devices just isn't enough for productivity, let said you planning for the up coming road trip, you need to take notes, plan route, POI, book hotel and rental a car. Would you(or anyone) rather to do it on your phone/tab/laptop or on a PC with 2 big screen/kb/ms? I am sick of taking note on phone, I use evernote on PC and let it sync to my phone.

Cable installer - most home use wifi, a $10 USB stick is much cheap than cable installer
Plumber?
Electrician?

 

[harobikes333]

Not to derail... but plumbers & electricians will be around nearly forever.....

 

[Socio]

Or they just have a system that is already good enough. That is becoming true even in smartphones.

I think this is true, technological leaps have become baby steps; I got almost 8 years out of my last PC with a ram upgrade and CPU upgrade and a couple video card upgrades over the duration.

 

[Denithor]

As mentioned already, 130+ million Steam gamers won't be cruising on outdated Core and Nehalem architectures for much longer. Sooner or later, tens of millions of them will open their wallets and get Skylake, Icelake, Tigerlake because once PS5/XB2 come out in 2019/2020, their current 2006-2010 CPUs will be not even meet the minimum spec for next gen games. Right now, a lot of these outdated CPUs can still game because the CPUs in PS4/XB1 are weak. Popular games like League of Legends, Starcraft 2, WOW, Dota 2, Counter Strike, Rocket League, etc. are still built around outdated/non-demanding game engines. Once successors of these games launch, or new popular games launch in 2019 and beyond using next gen DX12, there will be waves of gamers upgrading. There will be even more demand due to VR.

And you expect this to change drastically? I would assume that Sony/MS will stick with APU for next gen consoles, not cost effective to go to separate CPU+GPU instead. And, thus far, AMD has shown no signs of improving CPU power at all (certainly not within a power envelope usable in a console).

Many people, faced with updating a PC to be able to specifically play a game, won't. They will instead spend the same money to just buy a console.

What we really need, is keyboard/mouse included with consoles, along with a fully featured browser. That could, by itself, kill the desktop market for all but extreme gamers.

 

[Boze]

Or they just have a system that is already good enough. That is becoming true even in smartphones.

Yeap, this has been the case for me. I have an LG G3 right now, I was considering the G4 when it came out and then the G5, but the difference in real world performance is just not there, nor is battery life even an issue, because the G series has replaceable batteries that are around $9 on eBay.

For home computers, this is doubly true. I bought my Mom a used HP ProBook with a 3rd gen Core i7 2620M, 2 cores, 4 threads, 2.7 gHz base, 3.4 gHz turbo. I bought an 8 GB RAM upgrade kit for $29, and put her OCZ Trion 100 240 GB drive that I had replaced in her old laptop in there. Upgraded it to Windows 10.

For her, it screams. All she does is watch YouTube, use Facebook, and browse websites.

She constantly tells me how much faster it is than her old Samsung based on an AMD A4 chip, and since its using the same SSD, I know for sure its just a function of the extra 4 GB of RAM and the Core i7 processor.

EDIT: The point of this story is that she's using a Core i7 chip from 3 generations ago, around 4-5 years I think. Using five year old hardware and having no real reason to upgrade even in the far future is actually a testament to how well engineered the processors are, as well as how much more effective we're becoming at developing good code (probably to a much lesser degree though, since there's no reason to optimize for performance, most programmers won't, the only exception being game engine programmers).

 

[MrTeal]

Smartphones should continue to have pretty good turnover relative to desktops, regardless of a slowdown in performance increase. Other than HDD or fan failure, most desktops are relatively bulletproof and will usually last decades. Smartphones on the other hand are small pieces of glass we casually hold and manipulate in one hand while fumbling for keys and holding groceries in an asphalt parking lot. Attrition will keep smartphones selling if nothing else, both from physical damage as well as the increase difficulty and cost for sourcing replacement batteries every 2-3 years.

 

[escrow4]

Smartphones should continue to have pretty good turnover relative to desktops, regardless of a slowdown in performance increase. Other than HDD or fan failure, most desktops are relatively bulletproof and will usually last decades. Smartphones on the other hand are small pieces of glass we casually hold and manipulate in one hand while fumbling for keys and holding groceries in an asphalt parking lot. Attrition will keep smartphones selling if nothing else, both from physical damage as well as the increase difficulty and cost for sourcing replacement batteries every 2-3 years.

I've tried around 6 smartphones including a pricier one and I've given up on them. A $10 candybar that costs (over here not in the US) $60 a year to keep the number active (that I barely use) saves a fortune.

 

[Systems analyst]

The majority of the desktop and mobile market gains no performance going above 4 cores, and it actually hurts power efficiency. Only niche applications benefit from it, and it's going to cost you niche prices. Back in 2006, Dadi Perlmutter, Executive Vice President and General Manager of Intel Corporation's Intel Architecture Group, said this:

https://slashdot.org/story/06/07/27/1753255/intel---market-doesnt-need-eight-cores


Ten years later, it is still true. There are rumors that Cannonlake in 2017 might go above 4 cores, but they're only rumors. There's just no widespread need for it, as you can see from AMD offering 8-core mainstream chips for years and years now, without hurting Intel's business one bit.

And looking into the foreseeable future, people who know what they're talking about believe there will never be a need for more than four cores.

http://www.realworldtech.com/forum/?threadid=146066&curpostid=146227
http://www.realworldtech.com/forum/?threadid=146066&curpostid=146238


Sorry to say, but if you're looking for more cores, there's just no justifiable reason to make a mass-market chip with that.

Linus Torvalds is talking nonsense, but then, he is good at that; he does not have the experience to see that there are many and diverse solutions to many engineering problems, not just one solution to each. He is also not considering the cost of solutions; it is true that you could buy an expensive dual-core CPU from Intel, to execute an algorithm, which could otherwise be run on a 'wimpy' and cheap ARM based multi-core SoC, but most of us are cost-constrained. As a case in point, I am developing a system which runs in Mathematica, and uses all four processors in a RaspberryPi 2, at times; it would run fasterstill with eight wimpy cores, so I look forward to a future Pi offering this; the Pi is an order of magnitude cheaper than Intel's Core CPUs; it is good engineering to use the cheapest solution to this; as I leave my Pi on 24 hours a day, the wimpy solution is cheaper on electricity too.

 

[Atreidin]

When one actually reads and comprehends what Linus said, he says that "end users" are fine with on the order of four cores. He wrote this in 2014. This wasn't a statement of " this is true now and forever", either. Unless you are trying to purposely misrepresent what he wrote in order to set up a strawman, clearly he is only talking about the typical PC user having workloads that work better with fewer, beefier cores; the magic number 4 was an order of magnitude statement, not a specific hard value. He is saying that the people who evangelize extreme parallelism as a solution to everything are wrong, not that nobody anywhere, ever would ever want more than 4 cores. I realize one might not like his sometimes brash writing style, but he makes good points if you actually have good reading comprehension.