Desktop CPU upgrades have now shifted to a 20 year cycle.

[IntelUser2000]

Will I be using my 2600K rig in 2021? No my mother probably will be by 2018 or 2019.

20 years is a stretch.

However, using a well taken care of 2600K system until 2021 that is basically being used for internet surfing and youtube browsing is quite likely. Its quite likely that it'll be used until some critical functions don't work anymore, like the CPU. That may well make it close to 20 years. I am excited for the new technology that's coming out, but I look at my usage, and the hardware that's coming out, and its not enough for me. Even if its a good advancement, do I really need it?

My current case.

Your mother will probably buy one in 2018-2019 if its using HDDs and don't know much about computers, and go into a regular upgrade cycle when it slows down so much. That still means its "use it until it breaks".

 

[Fjodor2001]

* Small embedded CPUs. You get more than 2x performance for 2x price.

Nope, as quoted here, you said that you get free performance, by buying some unnamed embedded CPU.

What I'm saying is that if you buy an embedded CPU with performance X for Y, then you can often buy another similar embedded CPU with performance 2*X or greater for 2*Y. For anyone that has worked with embedded CPUs this is well known.

Also, another example:

RPi A+ vs RPi2. The former cost $20, the latter $35. The latter is more than twice as fast in MT performance, and in addition has lots of other HW that is better as a bonus.

 

[readers]

You are wrong according to steam stats CS 1.6 has over 20,000 players daily.

And the only reason it lost its 60,000 daily players is because many play CS GO for its skins because you can sell it back and get cash in your steam wallet to buy games and stuff.

Prior to CS GO and its Skin feature, CS Source which was much newer than 1.6 actually had WAY LESS players than CS 1.6

So you may want to fact check

PS CS 1.6 has more players than the latest COD on PC which shows I am right in that CS 1.6 is one of the most active FPS on the planet.

No, you are wrong and I am right and you just admitted in your own post.

CS:GO has waaaaay more players than 1.6 now.

Did I say anything about CS: Source? or COD? NO!

Get a clue.

This is what you said


"CS 1.6 and Starcraft 1, 15 and 17 years old respectively are 2 of the most active online games currently."
133,068 652,739 Counter-Strike: Global Offensive

6,421 23,188 Counter-Strike

Just pathetic.

 

[tenks]

who makes a thread about their own opinion, and then responds to people's comments saying they are wrong? were you craving acknowledgement or something?



Def troll/attention whore thread.

 

[TeknoBug]

I've read tales about telegraph wires acting as long antennas and starting fires in telegraph offices.

Yes there was a solar storm back in 1859 and it hit a Utah telegraph office, but it was the most powerful solar storm, more powerful than that one that hit Quebec's power grid in 1989.

 

[skipsneeky2]

For $21,i went from a E5200 to a Q6600 G0 recently and pretty much my game entire library went from borderline playable or unplayable to suddenly playable if not silky smooth.Games as old as COD BO1 to games as new as GTA V are in my library.

Desktop usage of course is flawless,typical Netflix/youtuber/facebook usage doesn't really stress out this cpu.Unless someone has a junker of a motherboard,anyone with a C2D can certainly gain much for $20.

 

[kimmel]

For $21,i went from a E5200 to a Q6600 G0 recently and pretty much my game entire library went from borderline playable or unplayable to suddenly playable if not silky smooth.Games as old as COD BO1 to games as new as GTA V are in my library.

Desktop usage of course is flawless,typical Netflix/youtuber/facebook usage doesn't really stress out this cpu.Unless someone has a junker of a motherboard,anyone with a C2D can certainly gain much for $20.

From someone who went from a Q6600 to a 4770k (both with ssds and same video card) in the last year you apparently have a high tolerance for slowness. I immediately noticed a marked speedup in just about everything from boot, to game load times, to transcode times. Q6600 is not in any way shape or form "silky smooth" even for web browsing compared to something modern. It regularly chunked on things that I fed it. Of course, when used the 6600 daily I thought it was smooth right up to the point where I switched.

 

[moonbogg]

If someone buys a sandy hex core, or newer version hex core or better, I agree that they could use their rig and play relevant games for probably 20 years. For average, every day tasks, that CPU will physically break before it can no longer run programs due to speed/cores/features. You won't get amazing performance after about a decade or so, but nowadays I can easily see a CPU lasting for about a quarter century for the average gamer/user. Its sad, but I'm serious.
Another way to put this is the following:

If you buy a new CPU and have a baby around the same time, by the time you become a grand parent is the time you should upgrade your CPU if you are an average user.

 

[TheELF]

If you buy a new CPU and have a baby around the same time, by the time you become a grand parent is the time you should upgrade your CPU if you are an average user.

If you buy a new CPU....in the $400-$500 range that is,and even then it's only a maybe,although things seem slow these days(years) nobody can know what the future holds.
Not that I generally disagree, but one should never be sure about something one can not be sure about.
I mean the way the IGPUs go they could be quantum computers in 15 years (Yup you "only" have to make every stream processor be able to be in both states at the same time) and cortana could be real artificial intelligence.

(Well, science does do huge steps these days and one can always dream,right? )

 

[skipsneeky2]

Q6600 is not in any way shape or form "silky smooth" even for web browsing compared to something modern.

Maybe a power browser user with 20 tabs open is going to notice slowdowns but for someone like me and my wife who open up Facebook on one tab and maybe have music playing on Youtube in another one certainly have no issues.It simply works.

Modern games show how bad off it is but some of my games like CS:GO and COD BO2 pretty much doubled their frame rate the second the Q6600 fell in here.80+fps in CS:GO and well over 100fps in COD BO2.:thumbsup:

 

[Shehriazad]

20 years? No.

5 years? Maaaaaaybe. Even though we have to see what Vulkan and DX12 etc cause to the gaming market first.


But then again...5-6 year old CPUs can still be kind of solid today(Phenom II started in 2009, right?)...so a 5 year cycle is already in the realm of the acceptable. But I wouldn't dare to say 20 years...especially not when breakthroughs can happen at any given time.

 

[myocardia]

I don't know how things work in the dark lands of the south, but up here in the dignified north, when we pay for something, we give currency in exchange. If the thing we're buying is an improved model that has a better priceerformance curve, that's called a better deal than that which was offered with the older model.

Now, if we got a complimentary bar of chocolate when we buy the new model from a particular store, that bar of chocolate would be considered free. But the new model? Still gotta pay for it.

So, you're admitting that all of the excess received, in excess of what you've paid currency to acquire, equals "free". Gotcha. That was what I originally said, to which you then seemed to be attempting to play that ole "dumb card" that you were accusing me of attempting to play. See, receiving a 300% performance gain, when "only" paying a 200% higher price, equals receiving performance for free, since anything for which you did not pay equals free.

 

[myocardia]

What I'm saying is that if you buy an embedded CPU with performance X for Y, then you can often buy another similar embedded CPU with performance 2*X or greater for 2*Y.

I'm just going to have to assume, after 4 or 5 replies that have contained exactly zero examples of this, that it has been sufficiently "MythBusted".

RPi A+ vs RPi2. The former cost $20, the latter $35.

This. This was the only example you ever needed. It's a great example, btw. Exactly the type I would have given, actually. Although we must all assume, since it took you this long to come up with your first actual example, that it is not something that happens anywhere near as often as the two of you were implying it happened. The Raspberry Pi example nevertheless proves your point, that it does in fact happen occasionally.

 

[Makaveli]

You are one of the strangest trolls I have had the pleasure of encountering.

Are you by any chance that autistic chap that got banned after throwing a tantrum?

You mean tweakboy?

lol this is 100% a troll thread.

There have been a couple good post from some members the rest of this is just bait.

 

[MrTeal]

Yes there was a solar storm back in 1859 and it hit a Utah telegraph office, but it was the most powerful solar storm, more powerful than that one that hit Quebec's power grid in 1989.

Storms as big as 1859 aren't common, but they aren't that rare. There was a massive CME a few years ago that was as large but happened to be pointed away from Earth. It might have actually been a good thing for computer gaming; devs could assume everyone that has a computer is running Ivy Bridge or newer.

 

[Fjodor2001]

I'm just going to have to assume, after 4 or 5 replies that have contained exactly zero examples of this, that it has been sufficiently "MythBusted".

Yes, your Myth has been busted. And the other replies contained examples busing it too, although you were reluctant to admit it. The reason for the additional replies was to clarify that to you. All the actual examples were given in a single post, and updated in another one with the RPi example.

Finally there are lots of similar examples as RPi and embedded CPUs in general if you start thinking about it.

Anyway, I think this discussion is not leading any further so I'm out of it now.

 

[MajinCry]

So, you're admitting that all of the excess received, in excess of what you've paid currency to acquire, equals "free". Gotcha. That was what I originally said, to which you then seemed to be attempting to play that ole "dumb card" that you were accusing me of attempting to play. See, receiving a 300% performance gain, when "only" paying a 200% higher price, equals receiving performance for free, since anything for which you did not pay equals free.

Admitting that getting excess in what you paid for is free? Uh. Yeah. That's what free stuff is. Getting a complimentary chocolate bar after buying a new CPU, means you got a free chocolate bar.

300% performance @ 200% price == 1.5x a better deal.

So, let me get this straight.

The intel 8086 launched for $360. The Core i5 6600k launched for $187. If you bought the i5 6600k, but if it only performed around 20% faster than 8086, you'd be fine with that? You describe that as "free performance" after all.

And before you begin, yes, I know the i5 6600k is insurmountably faster than the 8086. The above example is for the sake of showing you how outlandish and backwards your proposition of "free performance" is.

You paid for a CPU. A CPU better than it's predecessor. It so happens to perform vastly better, to the point that it is superior in every way, including it's priceerformance ratio.

That doesn't mean you got free performance. That means you got exponentially more performance for the increase in cost, yet you think priceerformance is linear?

 

[CHADBOGA]

You mean tweakboy?

lol this is 100% a troll thread.

There have been a couple good post from some members the rest of this is just bait.

I can't remember the name of the poster now, but a week or two ago I accidentally stumbled across a post by a guy who seemed reasonable at first, but got increasingly agitated when anyone disagreed with him, and then he mentioned he was autistic.

Soon after I noticed he had been banned.

 

[SOFTengCOMPelec]

Wrong thread!

 

[sm625]

Nobody on this planet plays League of Legends or DOTA 2 at the same time while watching a video.

I doubt it. I almost always have a video playing, usually a video podcast like the joe rogan experience, or similar content where you're not really watching just listening.

 

[selni]

20 years? No.

5 years? Maaaaaaybe. Even though we have to see what Vulkan and DX12 etc cause to the gaming market first.


But then again...5-6 year old CPUs can still be kind of solid today(Phenom II started in 2009, right?)...so a 5 year cycle is already in the realm of the acceptable. But I wouldn't dare to say 20 years...especially not when breakthroughs can happen at any given time.

You never really know what's going to happen - predicting technology is hard of course, but it's not completely unfounded in home/office use cases (gaming - which I know the OP mentions - is right out the window though) given year on year advances currently. For another example the original i7 is 2008 or so.

A 10% year to year increase for 20 years is a bit under 7x faster, which is obviously huge but also a gap we'd see now between high and low end stuff. 5% is 2.6x, and so on.

If we see increases much larger than that or something unexpected happens then all bets are off, but at that sort of rate it'll be chipset/peripheral connectivity etc that will become an issue for a lot of people before raw CPU performance.

 

[escrow4]

Maybe a power browser user with 20 tabs open is going to notice slowdowns but for someone like me and my wife who open up Facebook on one tab and maybe have music playing on Youtube in another one certainly have no issues.It simply works.

Modern games show how bad off it is but some of my games like CS:GO and COD BO2 pretty much doubled their frame rate the second the Q6600 fell in here.80+fps in CS:GO and well over 100fps in COD BO2.:thumbsup:

720p60 and 1080p60 would murder that CPU. On top of that its a power hog, not only would an i3 be substantially faster in single threaded tasks it would dump half the power and heat. Plus as someone has mentioned there is a difference in a modern quad (or i3) and an old Core 2. Its noticeable. As in, oooh, fast.

 

[IntelUser2000]

If we see increases much larger than that or something unexpected happens then all bets are off, but at that sort of rate it'll be chipset/peripheral connectivity etc that will become an issue for a lot of people before raw CPU performance.

7x faster for people that don't need that level of performance? They still won't care.

Give us something really different. I've been really excited by the prospects of 3D XPoint because of the potential. If they can advance that tech in the future, this is what should happen.

Super fast DRAM with small capacity acting as sort of a cache + 3D XPoint Gen xx acting as general RAM and Storage replacement.

Software taking FULL advantage of the new memory hierarchy.

-Instant boot computers. No more worrying about sleep modes, because full boot is near instant
-Instant loading of applications and files
-Instant management applications(like virus scanning)
-Games with massive worlds and textures with no loading times

No more constant moving data between RAM and storage would probably save significant power as well.

In fact, this is what the Intel patent that was related to XPoint and future talks about. Now THAT is a computer worth replacing. Not some minutely faster CPUs that we don't need.

http://www.kitguru.net/components/m...nt-ssds-will-feature-up-to-6gbs-of-bandwidth/

XPoint DIMMs seem already a revolutionary technology considering they are expecting 60% of DRAM bandwidth per channel.

Here it shows that XPoint DIMMs to be having 6GB/s bandwidth per DIMM, and DRAM to be 10GB/s bandwidth per DIMM: http://www.kitguru.net/components/m...nt-ssds-will-feature-up-to-6gbs-of-bandwidth/

While the author misunderstands the XPoint 6GB/s bandwidth to be theoretical bandwidth, Intel's slide shows it's probably based on Stream. On Stream the 8 DDR4 2133 memory channels of the 2x Xeon E5 v3 achieves about 95GB/s, or about 12GB/s per channel. XPoint being half that is quite impressive, with dual socket Xeon E5 Skylake that would come out to be 72GB/s(6 channel/CPU x 2 x 6GB/s).

http://www.anandtech.com/show/8423/intel-xeon-e5-version-3-up-to-18-haswell-ep-cores-/11

Regards to latency the DDR4 2133 DIMM is at 80-90ns, which is very close to the ~100ns figure Intel quotes. And think of in that same slide in KitGuru article where XPoint is at 250ns. That's quite amazing. Consider how Skylake is at 80ns.

 

[skipsneeky2]

720p60 and 1080p60 would murder that CPU. On top of that its a power hog, not only would an i3 be substantially faster in single threaded tasks it would dump half the power and heat. Plus as someone has mentioned there is a difference in a modern quad (or i3) and an old Core 2. Its noticeable. As in, oooh, fast.

No issues for me with 720p60 or 1080p60,usage is higher at upper 60's but no stuttering or anything.I use Chrome if that matters.

Only issues i honest to god have with this pc is new games,its a outdated cpu sure but having had chips as good as a i7 3770 i don't miss much of anything till i launch a 2013+game.With a ssd,the rig holds up well enough for the $20 i invested into a free computer.:biggrin:

My recommendation of a C2Q simply is more mileage on the cheap,$20 vs $150+ for 4 modern cores,a new motherboard+memory is a big difference especially if someone doesn't game.

 

[eton975]

I can't remember the name of the poster now, but a week or two ago I accidentally stumbled across a post by a guy who seemed reasonable at first, but got increasingly agitated when anyone disagreed with him, and then he mentioned he was autistic.

Soon after I noticed he had been banned.

Nec_V20? He was going on about how eeeeeevil Backblaze was for publishing their results and explaining their methodology.

Almost looked like a Seagate shill.