![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
How fast will PC's be in 2010?
- Thread starter mr. texas
- Start date
mr. texas said:10 Ghz processors?
3 Ghz Vid cards with 2 Gigs of Ram?
1 Terabyte HD's with 128 megs of cache?
....................................................................................
whats the deal?
So fast that the entire tonight show starring Conan O' Brian will be done from a lap-top in his living room starting in 2009
Moore's Law
think about it this way.... in the year 2000 what did you think computers in 2005 would be??
then take that figure how the actually turned out.... and then formulate a theory of where they will be then
I honestly believe SMP will become mainstream, or at least multi core tech, ram will probably double again, to where a 1Gb chip is considered "adaquate" . HDD's will be at at least SATA version 3 size I have no idea, I think 500 gig is not impossible for "average" use, vid cards are the one thing I can't call, who knows, SLI, super powerfull, modular maybe(upgradeable cores, memory, etc) but who knows that's what makes Tech fun
then take that figure how the actually turned out.... and then formulate a theory of where they will be then
I honestly believe SMP will become mainstream, or at least multi core tech, ram will probably double again, to where a 1Gb chip is considered "adaquate" . HDD's will be at at least SATA version 3 size I have no idea, I think 500 gig is not impossible for "average" use, vid cards are the one thing I can't call, who knows, SLI, super powerfull, modular maybe(upgradeable cores, memory, etc) but who knows that's what makes Tech fun
Well, assuming Microsoft can get Longhorn out by 2010, we should be seeing fully 64 bit native applications and maybe 6-7 GHz processors on the top end, and 4 GHz will probably be affordable for the average user.
I think more than anything we will probably see a revolution in motherboard design by then. Right now the choke point on most computers is the motherboard, and a new more efficient motherboard design could go a long way to improving performance.
I think more than anything we will probably see a revolution in motherboard design by then. Right now the choke point on most computers is the motherboard, and a new more efficient motherboard design could go a long way to improving performance.
10 Ghz processors?
probably not, they will go with more parralel type things now, like 1 - 4 CPU at 4-7? ghz with 2 or 4 cores plus hyperthreading on each core. by then everyone will be writing multi thread programs because the compiler will do it automatically.. and even intel will concentrate on doing more work per cycle than hyping up the speed. Note how they changed the names of their CPUs allready to de emphasise the speed.
3 Ghz Vid cards with 2 Gigs of Ram?
something like that, but they may go to multi-core as well. i hope ram gets that cheap, it probably will. someone may come out with some way where a seperate video card is not even needed with new CPUS.
1 Terabyte HD's with 128 megs of cache?
Easy! if we still have to use nasty clunky mechanical parts in our computers
probably not, they will go with more parralel type things now, like 1 - 4 CPU at 4-7? ghz with 2 or 4 cores plus hyperthreading on each core. by then everyone will be writing multi thread programs because the compiler will do it automatically.. and even intel will concentrate on doing more work per cycle than hyping up the speed. Note how they changed the names of their CPUs allready to de emphasise the speed.
3 Ghz Vid cards with 2 Gigs of Ram?
something like that, but they may go to multi-core as well. i hope ram gets that cheap, it probably will. someone may come out with some way where a seperate video card is not even needed with new CPUS.
1 Terabyte HD's with 128 megs of cache?
Easy! if we still have to use nasty clunky mechanical parts in our computers
Bugalaman
Really a Ugandachick
- Joined
- Apr 3, 2002
- Messages
- 3,275
lets hope that they come up with better names in the future, come on WTF is a Radeon? heh
TheMostWantedPolishTwin
Supreme [H]ardness
- Joined
- Sep 16, 2000
- Messages
- 6,206
well it might be that dual core CPUs will make it to the mainstream..
64 bit Dual/Multi core will be mainstream, RAM will continue to expand to 1+GB in size as standard. Dell will continue to offer 256MB in it's low end configs
Compiler technology will become increasingly important,and will be a driving source of innovation/performance.
GPU's will parallelize their architectures, and on board RAM will increase to perhaps 1GB on high end consumer cards. PCIe will enable high end workstations to have personal render farms of 2-4 cards. Due to their increasingly programmable nature - GPU's will be used as supplemental processors by certain classes of applications.
Heat, rather than pure speed will become the limiting factor in PC's. Watercooling will become more mainstream - but will not supplant air cooling. Spray cooling may come into mainstream use, refrigerators may replace case fans in some situations. Liquid helium will replace LN2
Hard drives will begin to be supplemented with some sort of solid state device for high speed applications (think cheap SRAM without wearout). Disk will still be used to dump data to.
Disk and CPU caches willl not increase dramatically in size due to rapidly diminishing returns - but disk density will continue to outpace Moore's law - leading to TB+ drives with 300+GB per platter.
Nuclear Fusion will be a by-product of Hyper-super-duper-mark 2010.
The focus will move from bigger faster to smaller quieter - as performance will become good enough. Wearable and portable computing will become more relevant, a Doom 3 server will be hosted on a PDA.
F@H and UD will attain sentience and battle for control of the earth.
Compiler technology will become increasingly important,and will be a driving source of innovation/performance.
GPU's will parallelize their architectures, and on board RAM will increase to perhaps 1GB on high end consumer cards. PCIe will enable high end workstations to have personal render farms of 2-4 cards. Due to their increasingly programmable nature - GPU's will be used as supplemental processors by certain classes of applications.
Heat, rather than pure speed will become the limiting factor in PC's. Watercooling will become more mainstream - but will not supplant air cooling. Spray cooling may come into mainstream use, refrigerators may replace case fans in some situations. Liquid helium will replace LN2
Hard drives will begin to be supplemented with some sort of solid state device for high speed applications (think cheap SRAM without wearout). Disk will still be used to dump data to.
Disk and CPU caches willl not increase dramatically in size due to rapidly diminishing returns - but disk density will continue to outpace Moore's law - leading to TB+ drives with 300+GB per platter.
Nuclear Fusion will be a by-product of Hyper-super-duper-mark 2010.
The focus will move from bigger faster to smaller quieter - as performance will become good enough. Wearable and portable computing will become more relevant, a Doom 3 server will be hosted on a PDA.
F@H and UD will attain sentience and battle for control of the earth.
64bit will be the thing... i dont think we'll see 10GHz, at least not with today's designs for processors... that's why they are working on new technology to increase speeds instead of the approach they've been going at for the past x amount of years of just making crap smaller and jamming more onto that one wafer.
video cards will improve significantly probably, computers will have alot more ram, bigger harddrives problem, and if we are lucky... someone will start creating alot faster harddrives (to give raptor some competition).
really though... how much improvement do we need? outside of games and video/graphic type of apps, hardware is currently outpacing software by a long shot... whereas 5 years ago hardware had alot of catching up to do.
video cards will improve significantly probably, computers will have alot more ram, bigger harddrives problem, and if we are lucky... someone will start creating alot faster harddrives (to give raptor some competition).
really though... how much improvement do we need? outside of games and video/graphic type of apps, hardware is currently outpacing software by a long shot... whereas 5 years ago hardware had alot of catching up to do.
so_cal_forever
[H]ard|Gawd
- Joined
- Jan 5, 2004
- Messages
- 1,125
Once synthetic diamonds become cheap to produce, they'll be used to make processors. This is good stuff, because diamon doesn't heat up with voltage present. So basically, once that happens, the skys the limit.
Master [H]
2[H]4U
- Joined
- Mar 17, 2003
- Messages
- 2,629
In 2010, we'll all be laughing about how 4GHz is so slow.
Let's see. Ten years ago was 1994. Back then we had the 486DX4, and early Pentiums. And we had mainboards, HDDs, CD-ROM drives, keyboards, mice, CRTs and FDDs.
The only really new thing has been the introduction of the GPU as a '3D-accelerator' (or decelerator in the case of S3's initial offerings). TFTs are kind of new, but hideously inferior to CRTs. MSFT gave us a shiny new key on our keyboards, mice got more buttons, and ISA still refuses to die despite PCI being faster and easier to configure.
Mainboards got easier to work on, and have assimilated countless features previously offered by I/O-cards and similar. Mainstream CPUs are catching up with the server technology of ten years ago. OSs as well. HDDs have grown in size, making even those new DVDs look tiny in terms of storage capacity (in 1994/1995 I could have backed up my entire HDD on a single CD-R).
So, things did get faster, bigger and easier, but aside from being able to render 3D-scenes on our 20+" displays in realtime, current tech would not make a '94 geek fall over backwards in amazement.
What does the future hold?
We've got SED- and OLED-displays to finally get rid of those pesky CRTs and TFTs, MRAM and other non-volatile RAM technologies to replace some of the last remaining mechanical parts of our systems (HDD and CD/DVD-ROM), PCIe to ease the pain of PCI, which is obviously on its last legs, SATA to bring the advantages of SCSI (finally) to IDE. And to get rid of those airflow-restricting ribbon-cables at the same time.
Mainboards will continue to assimilate more features, just like CPUs are assimilating features previously restricted to the chipset. The SB will still remain a mess of ancient buses, which nonetheless still work perfectly fine.
Instead of a single or dual CPU, the emphasis will be placed on parallel processing. And not just SMP, but possibly assisted by different PUs (Processing Units), each with a specific function, possibly dynamic (think FPGA).
Meanwhile, we'll still be typing on our Qwerty-keyboards, using the same kind of mice/trackball (but with 10+ buttons and a 2000 DPI resolution) to click widgets in the same failed attempt at cramming a 3D interface on a 2D surface.
The x86-64/128 architecture will still have fewer GP registers than the very first RISC CPU, there'll be too few IRQs to keep a low-level developer sane, and 16-bit applications will still work in the, also still supported and default, Real Mode.
And we'll probably still be arguing about when COM/LPT-ports, the PS/2 bus and the trusty FDD will finally die.
Faster, bigger, easier. And more prone to crashing or breaking down, if the last ten years is anything to judge by.
The only really new thing has been the introduction of the GPU as a '3D-accelerator' (or decelerator in the case of S3's initial offerings). TFTs are kind of new, but hideously inferior to CRTs. MSFT gave us a shiny new key on our keyboards, mice got more buttons, and ISA still refuses to die despite PCI being faster and easier to configure.
Mainboards got easier to work on, and have assimilated countless features previously offered by I/O-cards and similar. Mainstream CPUs are catching up with the server technology of ten years ago. OSs as well. HDDs have grown in size, making even those new DVDs look tiny in terms of storage capacity (in 1994/1995 I could have backed up my entire HDD on a single CD-R).
So, things did get faster, bigger and easier, but aside from being able to render 3D-scenes on our 20+" displays in realtime, current tech would not make a '94 geek fall over backwards in amazement.
What does the future hold?
We've got SED- and OLED-displays to finally get rid of those pesky CRTs and TFTs, MRAM and other non-volatile RAM technologies to replace some of the last remaining mechanical parts of our systems (HDD and CD/DVD-ROM), PCIe to ease the pain of PCI, which is obviously on its last legs, SATA to bring the advantages of SCSI (finally) to IDE. And to get rid of those airflow-restricting ribbon-cables at the same time.
Mainboards will continue to assimilate more features, just like CPUs are assimilating features previously restricted to the chipset. The SB will still remain a mess of ancient buses, which nonetheless still work perfectly fine.
Instead of a single or dual CPU, the emphasis will be placed on parallel processing. And not just SMP, but possibly assisted by different PUs (Processing Units), each with a specific function, possibly dynamic (think FPGA).
Meanwhile, we'll still be typing on our Qwerty-keyboards, using the same kind of mice/trackball (but with 10+ buttons and a 2000 DPI resolution) to click widgets in the same failed attempt at cramming a 3D interface on a 2D surface.
The x86-64/128 architecture will still have fewer GP registers than the very first RISC CPU, there'll be too few IRQs to keep a low-level developer sane, and 16-bit applications will still work in the, also still supported and default, Real Mode.
And we'll probably still be arguing about when COM/LPT-ports, the PS/2 bus and the trusty FDD will finally die.
Faster, bigger, easier. And more prone to crashing or breaking down, if the last ten years is anything to judge by.
Master [H] said:In 2010, we'll all be laughing about how 4GHz is so slow.
I don't know that we will neccessarily. Many of the most powerful modern processors have clock speeds of approximately 1-2 GHz. Dual core takes a lot of the pressure to ramp single pipeling execution way up via clockspeed off. I have trouble evnisioning an 8GHZ processor, because other than marketing reasons - you can do a lot more with many slower processors working in parrallel.
Take the human brain for example: big, massively parrallel - stupidly slow computer
I can see intel pushing speed into the 5-6 GHz range, but I have a really hard time seeing AMD move to an architecture that pushes it's procs much past the 4 GHz range. (for sure over the next 3 years) 2010 is a long way off, but remember that Moore's law says the transistors per unit area will double every 16-18 months - NOT clockspeed. in 5 years I would think you could build a HELLA fast 4GHZ cpu.
1). Definately 64 bit multi-core processors running at 4GHz a piece
2). MRAM will be mainstream
3). 1 TB HD will be standard
4). Windows Longhorn 64 SP4 will be the standard OS
5). You will need to have at least 2GB of RAM to turn your computer on
6). Half Life 2 and Duke Nukem Forever will be out just in time for Christmas
7). Your video card will also have a multi-core, 64 bit processor with at least a gig of on-board ram
8). chips will need to use thermoaccustic refrigeration to keep them at sub-fusion temps
2). MRAM will be mainstream
3). 1 TB HD will be standard
4). Windows Longhorn 64 SP4 will be the standard OS
5). You will need to have at least 2GB of RAM to turn your computer on
6). Half Life 2 and Duke Nukem Forever will be out just in time for Christmas
7). Your video card will also have a multi-core, 64 bit processor with at least a gig of on-board ram
8). chips will need to use thermoaccustic refrigeration to keep them at sub-fusion temps
Just thought of something...
we won't ever reach 10GHz with the current way of doing things at all. Wire's can't take 10GHz frequencies.. well, they can, but not easily.
at 3GHz a wires characteristic impedence starts to be a nuisance... at 18GHz wire simple cannot be used becuase the characteristic impedence is waaay to high.... 10GHz falls in the middle, which probably also is unusable for processors.
in other words... becuase of the wiring, we might be nearing the top of the CPU spectrum. Now, there are other ways to make faster computers of course... dual procs, etc... but we'll have to wait for fabricated diamonds and what not before we really see a jump in things.
we won't ever reach 10GHz with the current way of doing things at all. Wire's can't take 10GHz frequencies.. well, they can, but not easily.
at 3GHz a wires characteristic impedence starts to be a nuisance... at 18GHz wire simple cannot be used becuase the characteristic impedence is waaay to high.... 10GHz falls in the middle, which probably also is unusable for processors.
in other words... becuase of the wiring, we might be nearing the top of the CPU spectrum. Now, there are other ways to make faster computers of course... dual procs, etc... but we'll have to wait for fabricated diamonds and what not before we really see a jump in things.
JulesBravo
Weaksauce
- Joined
- Nov 28, 2003
- Messages
- 68
With the falling costs on synthetic diamonds in 5 years I believe we will begin seeing the first chips using diamonds, and a rapid increase in speeds, quite possibly in the 50 Ghz range. Diamonds run so cool that you can do practically anything with them.
so_cal_forever
[H]ard|Gawd
- Joined
- Jan 5, 2004
- Messages
- 1,125
JulesBravo said:
Exactly. However, we will hit a wall, and not with how far we can take the speed, but how far we actually need it. If we don't need over 15ghz of processing power, then there's no reason to even try. However, synthetic diamonds will be great because the CPU could work on a sort of overdrive, only clocking as high as it's needed.
D
Deleted member 89137
Guest
havnt u heard? amd is supposedly working on a mid 1ghz proc that will out perform any 3ghz + processor...
I wish they would make hard drives or some kind of storage that could keep up with everything else at a reasonable price.
Having the fast cpu and ram etc makes hard drives seem even slower. what is a raptor anyways, like 1.5 - 2x the fastest drive avaliable 5 years ago meanwhile everything else has increased by a lot more.
it doesnt help that as new games come out they have bigger and bigger levels to take advantage of better CPU and video cards. so they even take longer to load.
same with VSTi plugins for music, image/video editing and even everyday programs that have many added "features"
if they could figure that one out then even a p3 600 would "feel" a LOT faster. than the best new machines for browsing and basic office tasks.
when all these wonderful "diamond CPUs" and everything are out i hope we dont have to use hard drives that are only 2 times faster than the ones now.
just for fun if you have photoshop 4 saved somewhere install it and see how it seems to start as fast as notepad with your new machine but the latest and greatest 6 or "CS" one takes as long or even longer than ps4 did when it was new on your shiny p2 300!
/rant
oh yeah i am not complaining because i have particularly slow drives, when i play bf1942 or DC i am usually the first or second on the map when it loads i just want it MUCH faster !!!!111
Having the fast cpu and ram etc makes hard drives seem even slower. what is a raptor anyways, like 1.5 - 2x the fastest drive avaliable 5 years ago meanwhile everything else has increased by a lot more.
it doesnt help that as new games come out they have bigger and bigger levels to take advantage of better CPU and video cards. so they even take longer to load.
same with VSTi plugins for music, image/video editing and even everyday programs that have many added "features"
if they could figure that one out then even a p3 600 would "feel" a LOT faster. than the best new machines for browsing and basic office tasks.
when all these wonderful "diamond CPUs" and everything are out i hope we dont have to use hard drives that are only 2 times faster than the ones now.
just for fun if you have photoshop 4 saved somewhere install it and see how it seems to start as fast as notepad with your new machine
but the latest and greatest 6 or "CS" one takes as long or even longer than ps4 did when it was new on your shiny p2 300!/rant
oh yeah i am not complaining because i have particularly slow drives, when i play bf1942 or DC i am usually the first or second on the map when it loads
i just want it MUCH faster !!!!111RancidWAnnaRIot
EspantaPajaros
- Joined
- Feb 19, 2004
- Messages
- 11,033
Hopefully my computer will be able to do all my work without it even asking..
Non-volatile RAM technologies which don't wear out easily (like Flash) have already reached where prototypes are produced. MRAM is already being used in a limited number of applications.SB22 said:
Because scaling down the structures on a RAM chip is relatively easy (it are the same elements repeated over and over again) it would not be too outlandish to expect such technologies to begin to replace HDDs in the coming five years or so.
Also, this'll mean that we'll go back to cartridges for removable data storage, rendering FDs (finally), CDs, DVDs, etc. obsolete.
I assume by 2010 manufacturers will have had to make some radical shifts in design to keep pace, and they will because that is where their income is coming from. The real interesting thing to me is how big of a player the PRC is going to become, they are building fabs and absorbing tech at an incredible rate and could be a major player in 6 years. What they bring with could have a huge impact on what we see today, another CPU war so to speak like Intel/AMD's rush to 1GHz. But either way, multicore CPU's I believe will become a standard part, 4+ GB's of RAM will be mainstream, and I hope to God that SSD's finally replace magnetic sotrage, and hopefully in a universal 2.5" format. Honestly those things will probably happen in two years rather than six, so I guess my magic 8 Ball is busted again Oh, and it would be nice if the average user could get a WUXGA caliber display for a decent price, just think what that could do for video, HD adoption and general productivity...more than all the fast CPU's in a lot of ways.
Oh, and it would be nice if the average user could get a WUXGA caliber display for a decent price, just think what that could do for video, HD adoption and general productivity...more than all the fast CPU's in a lot of ways.Solid state harddrives would be nice.
Also, a more reliable manufacturing process for LCD's to drive down the price would be nice... like 40% of all LCDs come off the line flawed, so they have to scrap them. This is why LCDs cost so damn much.
Also, a more reliable manufacturing process for LCD's to drive down the price would be nice... like 40% of all LCDs come off the line flawed, so they have to scrap them. This is why LCDs cost so damn much.
how bout LCD's that are dirt cheap, have no dead pixels, and are as good, or better than professional CRT's in terms of refresh rate and color quality/response.
add in some HDR contrast ratios, perfect blackness with no "mirror coat", and 1600x1200 as a standard for 15-19" inch models. (as some laptops have)
add in some HDR contrast ratios, perfect blackness with no "mirror coat", and 1600x1200 as a standard for 15-19" inch models. (as some laptops have)