Overclock.net › Forums › Industry News › Hardware News › [CPC] Intel: CUDA will be just a 'footnote' in computing history
New Posts  All Forums:Forum Nav:

[CPC] Intel: CUDA will be just a 'footnote' in computing history

post #1 of 17
Thread Starter 
Quote:
Intel has revealed that it sees no place in the future of computing for general purpose GPU (GPGPU) programming models such as Nvidia’s CUDA, which has enabled Stanford's Nvidia GPU folding client, saying that programmers don’t have the time to learn how to program for radical new architectures.

In a Q&A session after announcing Intel’s 40th birthday, we asked Intel’s senior vice president and co-general manager of Intel Corporation's Digital Enterprise Group, Pat Gelsinger, where he saw GPGPU languages such as CUDA in the future. He said that they would be nothing more than ‘interesting footnotes in the history of computing annals.’

‘The problem that we’ve seen over and over and over again in the computing industry is that there’s a cool new idea, and it promises a 10x or 20x performance improvements, but you’ve just got to go through this little orifice called a new programming model,’ Gelsinger explained to Custom PC. Those orifices, says Gelsinger, have always been ‘insurmountable as long as the general purpose computing models evolve into the future.’

Gelsinger used the Cell architecture used in the PlayStation 3’s CPU as an example to prove his point. ‘It [Cell] promised to be this radical new computing architecture,’ said Gelsinger, ‘and basically years later the application programmers have barely been able to comprehend how to write applications for it.’

This, according to Gelsinger, is one of the major reasons why Intel’s forthcoming Larrabee graphics chip will be entirely based on IA (Intel Architecture) x86 cores. ‘Our approach to this has been to not force programmers to make a radical architectural shift,’ explained Gelsinger, ‘but to take what they already know – this IA-compatible architecture – and extend the programming model to comprehend new visual computing data-parallel-throughput workloads, and that’s the strategy that we’re taking with Larrabee.’

Larrabee, according to Gelsinger, will simply expand on a standard programming model. ‘It’s an IA-compatible core,’ explained Gelsinger, ‘and we’re extending it with a graphics vector visualisation instruction set that has full support for native programming models, full support for the graphics APIs like DX and OpenGL, and then this broad set of new programming models to go with it.’

Gelsinger claims that the ISVs (independent software vendors) that are currently dealing with Larrabee have responded with ‘nothing but sheer passion and enthusiasm for that direction.’ As such, he added that ‘we expect things like CUDA and CTN will end up in the same interesting footnotes in the history of computing annals – they had great promise and there were a few applications that were able to take advantage of them, but generally an evolutionary compatible computing model, such as we’re proposing with Larrabee, we expect will be the right answer long term.’
http://www.custompc.co.uk/news/60286...g-history.html

The INQ had an article on this earlier, but given that it was in the usual unprofessional INQ tone I didnt even bother posting it. This one is also a fair bit longer.

I agree with Gelsinger, but not in the way he might want me to. CUDA and CAL are just stepping stones, vendor-specific languages can only go so far. Once the Khronos group decides on an open parallel computing standard it will really take off, and the programs will run on both Larrabee, GeForce and Radeon.

So yeah, if Larrabee was out now itd be much more interesting as a parallel coprocessor, but itll be out in 2010 and DX11/compute shaders/open languages will have caught up to them by then.

I posted this in the Hardware section as its Larrabee related of sorts, move it if you deem it more correct in Software.
Edited by bowman - 7/1/08 at 5:43pm
Small box
(13 items)
 
  
CPUMotherboardGraphicsRAM
Intel 7700K ASUS Z270F Strix ASUS 1080 Ti FE Corsair Vengeance 16 GB 3200 MHz 
Hard DriveCoolingOSMonitor
Samsung 960 Evo Corsair H115i Windows 10 ASUS PG279Q 
KeyboardPowerCaseMouse
Unicomp Corsair CX850 Fractal Design Define C Logitech G PRO 
Mouse Pad
Steelseries QCK XXL 
  hide details  
Reply
Small box
(13 items)
 
  
CPUMotherboardGraphicsRAM
Intel 7700K ASUS Z270F Strix ASUS 1080 Ti FE Corsair Vengeance 16 GB 3200 MHz 
Hard DriveCoolingOSMonitor
Samsung 960 Evo Corsair H115i Windows 10 ASUS PG279Q 
KeyboardPowerCaseMouse
Unicomp Corsair CX850 Fractal Design Define C Logitech G PRO 
Mouse Pad
Steelseries QCK XXL 
  hide details  
Reply
post #2 of 17
This is all PR. Intel has donated millions to (and I think is leading) a large body of research into unlocking the power of massively parallel computing. Sequential programming is good at a lot of things, but Moore's law is running out for single threaded applications. CUDA and similar tech is the way of the future, and Intel knows it. CUDA itself is probably just the first step in a field that will mature and generalize, but the tech is here to stay.
It goes to eleven
(13 items)
 
  
CPUMotherboardGraphicsRAM
E6300 DS3 EVGA 8600GTS 2GB XMS2 DDR2-800 
Hard DriveOSMonitorKeyboard
1.294 TB Arch Linux/XP Samsung 226bw Eclipse II 
PowerCaseMouse
Corsair 520HX Lian-Li v1000B Plus G7 
  hide details  
Reply
It goes to eleven
(13 items)
 
  
CPUMotherboardGraphicsRAM
E6300 DS3 EVGA 8600GTS 2GB XMS2 DDR2-800 
Hard DriveOSMonitorKeyboard
1.294 TB Arch Linux/XP Samsung 226bw Eclipse II 
PowerCaseMouse
Corsair 520HX Lian-Li v1000B Plus G7 
  hide details  
Reply
post #3 of 17
We need to get of x86 IMO, its a really old tech, they should just slowly adapt and change it, for example with how they're doing x64, but more forcing us to upgrade.
My system
(23 items)
 
  
CPUMotherboardGraphicsRAM
Intel Core i5 3570k @ 4.5Ghz ASRock Z77 Pro3 Powercolor Radeon HD7950 3GB @ 1150/1350 4x4GB G.Skill Ares 2000Mhz CL9 
Hard DriveHard DriveHard DriveHard Drive
Samsung 840 250GB Western Digital Black 1TB WD1002FAEX Seagate Barracuda 3TB ST3000DM001 Samsung Spinpoint EcoGreen 2TB 
Optical DriveCoolingCoolingCooling
Pioneer DVR-220LBKS Noctua NH-D14 Scythe Gentle Typhoon 1850rpm Corsair AF140 Quiet Edition 
CoolingOSMonitorMonitor
Arcitc Cooling Acclero Twin Turbo II Arch Linux x86-64, amdgpu BenQ G2220HD BenQ G2020HD 
KeyboardPowerCaseMouse
Ducky Shine III Year of the Snake, Cherry Blue Silverstone Strider Plus 600w CoolerMaster CM690 II Black and White SteelSeries Sensei Professional 
Mouse PadAudioOther
Artisan Hien Mid Japan Black Large ASUS Xonar DX NZXT Sentry Mesh 30w Fan Controller 
  hide details  
Reply
My system
(23 items)
 
  
CPUMotherboardGraphicsRAM
Intel Core i5 3570k @ 4.5Ghz ASRock Z77 Pro3 Powercolor Radeon HD7950 3GB @ 1150/1350 4x4GB G.Skill Ares 2000Mhz CL9 
Hard DriveHard DriveHard DriveHard Drive
Samsung 840 250GB Western Digital Black 1TB WD1002FAEX Seagate Barracuda 3TB ST3000DM001 Samsung Spinpoint EcoGreen 2TB 
Optical DriveCoolingCoolingCooling
Pioneer DVR-220LBKS Noctua NH-D14 Scythe Gentle Typhoon 1850rpm Corsair AF140 Quiet Edition 
CoolingOSMonitorMonitor
Arcitc Cooling Acclero Twin Turbo II Arch Linux x86-64, amdgpu BenQ G2220HD BenQ G2020HD 
KeyboardPowerCaseMouse
Ducky Shine III Year of the Snake, Cherry Blue Silverstone Strider Plus 600w CoolerMaster CM690 II Black and White SteelSeries Sensei Professional 
Mouse PadAudioOther
Artisan Hien Mid Japan Black Large ASUS Xonar DX NZXT Sentry Mesh 30w Fan Controller 
  hide details  
Reply
post #4 of 17
Thread Starter 
Quote:
Originally Posted by rabidgnome229 View Post
This is all PR. Intel has donated millions to (and I think is leading) a large body of research into unlocking the power of massively parallel computing. Sequential programming is good at a lot of things, but Moore's law is running out for single threaded applications. CUDA and similar tech is the way of the future, and Intel knows it. CUDA itself is probably just the first step in a field that will mature and generalize, but the tech is here to stay.
Hes not putting down parallel processing, hes putting down the vendor-locked programming language model. He says its not going to grow big, and I think anyone can see that.

NVIDIA, AMD and Intel are all in the Khronos group working on the open parallel computing standard which theyll all support, itll take over where CUDA and CAL leave the stage and go into the footnote of the history books as the little engine that could bring parallel computing to the front pages.

Also, x86 is hardly going anywhere, its everywhere now. Next year there will be x86 Atom processors running Linux in mobile phones, the fastest supercomputer cluster runs on x86 Opterons and Linux. They solved the scalability problem when they designed RISC processors and added a translator to it. Its not a problem, rather a reality and not one that is going to change. Upgrade paths and backwards compatibility is important.
Small box
(13 items)
 
  
CPUMotherboardGraphicsRAM
Intel 7700K ASUS Z270F Strix ASUS 1080 Ti FE Corsair Vengeance 16 GB 3200 MHz 
Hard DriveCoolingOSMonitor
Samsung 960 Evo Corsair H115i Windows 10 ASUS PG279Q 
KeyboardPowerCaseMouse
Unicomp Corsair CX850 Fractal Design Define C Logitech G PRO 
Mouse Pad
Steelseries QCK XXL 
  hide details  
Reply
Small box
(13 items)
 
  
CPUMotherboardGraphicsRAM
Intel 7700K ASUS Z270F Strix ASUS 1080 Ti FE Corsair Vengeance 16 GB 3200 MHz 
Hard DriveCoolingOSMonitor
Samsung 960 Evo Corsair H115i Windows 10 ASUS PG279Q 
KeyboardPowerCaseMouse
Unicomp Corsair CX850 Fractal Design Define C Logitech G PRO 
Mouse Pad
Steelseries QCK XXL 
  hide details  
Reply
post #5 of 17
Quote:
Originally Posted by Brutuz View Post
We need to get of x86 IMO, its a really old tech, they should just slowly adapt and change it, for example with how they're doing x64, but more forcing us to upgrade.
Extremely off topic but I was just wondering, you claim to be a graphics card aficionado yet your running a 6800? Forgive me If I'm missing something as I do not mean to offend, I simply don't understand as a lot of people claim the same.
    
CPUMotherboardGraphicsRAM
Q6600 @ 3.8Ghz / E8600 @ 4.8Ghz 24/7! DFI X48 DK HD 4870 + (4870x2) 4gb Crucial Ballistix Tracer @1100 4.4.4.12 
Hard DriveOptical DriveOSMonitor
3 x 250GB Sata II in raid 0 LG Dual Layer Burner XP pro sp2 / Vista Ultimate 22 Inch BenQ / 24 Inch Samsung 
KeyboardPowerCaseMouse
Logitech G 11 Silverstone Strider 850w Modular Antec 900 Razer Copperhead 
Mouse Pad
Nova Microoptic 
  hide details  
Reply
    
CPUMotherboardGraphicsRAM
Q6600 @ 3.8Ghz / E8600 @ 4.8Ghz 24/7! DFI X48 DK HD 4870 + (4870x2) 4gb Crucial Ballistix Tracer @1100 4.4.4.12 
Hard DriveOptical DriveOSMonitor
3 x 250GB Sata II in raid 0 LG Dual Layer Burner XP pro sp2 / Vista Ultimate 22 Inch BenQ / 24 Inch Samsung 
KeyboardPowerCaseMouse
Logitech G 11 Silverstone Strider 850w Modular Antec 900 Razer Copperhead 
Mouse Pad
Nova Microoptic 
  hide details  
Reply
post #6 of 17
When all GPU can run code similar to that of regular code then we will enter the point where the CPU is only used for light work, any 'heavy lifting' that can be done by GPU will be done by GPU. If necessary we can split the work between the GPU and CPU and have them work together with the CPU carrying out x86 exclusive operations and the GPU tackling the rest.
Lee XT
(17 items)
 
  
CPUMotherboardGraphicsRAM
AMD FX-6300 Asus M5A97 SAPPHIRE Radeon HD 7850 AMD 4GB DDR3 1333MHZ 
RAMRAMRAMHard Drive
AMD 4GB DDR3 1333MHZ AMD 4GB DDR3 1333MHZ AMD 4GB DDR3 1333MHZ OCZ Vertex 4 256GB 
CoolingOSMonitorKeyboard
Corsair H80 Windows 8.1 Pro MCE Dell P2414H WHXV7  Microsoft Generic 
PowerCaseMouseMouse Pad
Ultra 600W Limited Edition NZXT Black Steel Razer Deathadder Razer Goliath 
Audio
Realtek HD Audio 
  hide details  
Reply
Lee XT
(17 items)
 
  
CPUMotherboardGraphicsRAM
AMD FX-6300 Asus M5A97 SAPPHIRE Radeon HD 7850 AMD 4GB DDR3 1333MHZ 
RAMRAMRAMHard Drive
AMD 4GB DDR3 1333MHZ AMD 4GB DDR3 1333MHZ AMD 4GB DDR3 1333MHZ OCZ Vertex 4 256GB 
CoolingOSMonitorKeyboard
Corsair H80 Windows 8.1 Pro MCE Dell P2414H WHXV7  Microsoft Generic 
PowerCaseMouseMouse Pad
Ultra 600W Limited Edition NZXT Black Steel Razer Deathadder Razer Goliath 
Audio
Realtek HD Audio 
  hide details  
Reply
post #7 of 17
Thread Starter 
Quote:
Originally Posted by Licht View Post
When all GPU can run code similar to that of regular code then we will enter the point where the CPU is only used for light work, any 'heavy lifting' that can be done by GPU will be done by GPU. If necessary we can split the work between the GPU and CPU and have them work together with the CPU carrying out x86 exclusive operations and the GPU tackling the rest.
Every PC will be a 'mini-Roadrunner', CPU takes care of file management, I/O, parallel coprocessors do heavy computational tasks.
Small box
(13 items)
 
  
CPUMotherboardGraphicsRAM
Intel 7700K ASUS Z270F Strix ASUS 1080 Ti FE Corsair Vengeance 16 GB 3200 MHz 
Hard DriveCoolingOSMonitor
Samsung 960 Evo Corsair H115i Windows 10 ASUS PG279Q 
KeyboardPowerCaseMouse
Unicomp Corsair CX850 Fractal Design Define C Logitech G PRO 
Mouse Pad
Steelseries QCK XXL 
  hide details  
Reply
Small box
(13 items)
 
  
CPUMotherboardGraphicsRAM
Intel 7700K ASUS Z270F Strix ASUS 1080 Ti FE Corsair Vengeance 16 GB 3200 MHz 
Hard DriveCoolingOSMonitor
Samsung 960 Evo Corsair H115i Windows 10 ASUS PG279Q 
KeyboardPowerCaseMouse
Unicomp Corsair CX850 Fractal Design Define C Logitech G PRO 
Mouse Pad
Steelseries QCK XXL 
  hide details  
Reply
post #8 of 17
Quote:
Originally Posted by zaccy001 View Post
Extremely off topic but I was just wondering, you claim to be a graphics card aficionado yet your running a 6800? Forgive me If I'm missing something as I do not mean to offend, I simply don't understand as a lot of people claim the same.
You don't have to own the best before you can be an aficionado. Hell, I know a guy with an X800XT that knows more about graphics cards than most people with 9800GX2 cards in SLI. Being an aficionado is more about knowing what's up that spending your money on every minor performance increase that comes along.
The Replacement
(16 items)
 
  
CPUMotherboardGraphicsRAM
Core I5 6600K Gigabyte GA-Z170X-UD3 Ultra  EVGA GTX 750 Ti Crucial Ballistix Sport 2x16GB DDR4 2400 
Hard DriveOptical DriveCoolingOS
Mushkin Reactor 1TB LG Blu-Ray Burner Cryorig H7 Window 10 
MonitorKeyboardPowerCase
BenQ XL2420T Code 87 Clears Rosewill Photon 650 Fractal Design Core 3500W 
MouseMouse PadAudioAudio
CM Storm Xornet II Steelseries QCK+ Creative Sound Blaster Zx Sennheiser PC-360 
  hide details  
Reply
The Replacement
(16 items)
 
  
CPUMotherboardGraphicsRAM
Core I5 6600K Gigabyte GA-Z170X-UD3 Ultra  EVGA GTX 750 Ti Crucial Ballistix Sport 2x16GB DDR4 2400 
Hard DriveOptical DriveCoolingOS
Mushkin Reactor 1TB LG Blu-Ray Burner Cryorig H7 Window 10 
MonitorKeyboardPowerCase
BenQ XL2420T Code 87 Clears Rosewill Photon 650 Fractal Design Core 3500W 
MouseMouse PadAudioAudio
CM Storm Xornet II Steelseries QCK+ Creative Sound Blaster Zx Sennheiser PC-360 
  hide details  
Reply
post #9 of 17
Quote:
Originally Posted by Licht View Post
When all GPU can run code similar to that of regular code then we will enter the point where the CPU is only used for light work, any 'heavy lifting' that can be done by GPU will be done by GPU. If necessary we can split the work between the GPU and CPU and have them work together with the CPU carrying out x86 exclusive operations and the GPU tackling the rest.
I don't see that happening. GPU's are good at parallelism, and floating point code in particular. CPU's are better at logic and integer calculation. That's why you only ever see GPU's rated by FLOPS (floating point ops per second). For good old sequential, integer code the CPU still reigns and will continue to do so.

Quote:
Originally Posted by bowman View Post
Hes not putting down parallel processing, hes putting down the vendor-locked programming language model. He says its not going to grow big, and I think anyone can see that.

NVIDIA, AMD and Intel are all in the Khronos group working on the open parallel computing standard which theyll all support, itll take over where CUDA and CAL leave the stage and go into the footnote of the history books as the little engine that could bring parallel computing to the front pages.

Also, x86 is hardly going anywhere, its everywhere now. Next year there will be x86 Atom processors running Linux in mobile phones, the fastest supercomputer cluster runs on x86 Opterons and Linux. They solved the scalability problem when they designed RISC processors and added a translator to it. Its not a problem, rather a reality and not one that is going to change. Upgrade paths and backwards compatibility is important.
Good point - I was reading that wrong
It goes to eleven
(13 items)
 
  
CPUMotherboardGraphicsRAM
E6300 DS3 EVGA 8600GTS 2GB XMS2 DDR2-800 
Hard DriveOSMonitorKeyboard
1.294 TB Arch Linux/XP Samsung 226bw Eclipse II 
PowerCaseMouse
Corsair 520HX Lian-Li v1000B Plus G7 
  hide details  
Reply
It goes to eleven
(13 items)
 
  
CPUMotherboardGraphicsRAM
E6300 DS3 EVGA 8600GTS 2GB XMS2 DDR2-800 
Hard DriveOSMonitorKeyboard
1.294 TB Arch Linux/XP Samsung 226bw Eclipse II 
PowerCaseMouse
Corsair 520HX Lian-Li v1000B Plus G7 
  hide details  
Reply
post #10 of 17
Quote:
Originally Posted by rabidgnome229 View Post
I don't see that happening. GPU's are good at parallelism, and floating point code in particular. CPU's are better at logic and integer calculation. That's why you only ever see GPU's rated by FLOPS (floating point ops per second). For good old sequential, integer code the CPU still reigns and will continue to do so.
I was talking about splitting the work between them. For example if it would run faster on the GPU get it done there, if it would on CPU do it there. I think that kind of system would be the true future of computing. Like now we have sound, graphics, physics, central, and occasionally some other processors. If we put these basics in all PC, and tailer coded every application that has large demands to utilize this system : we could have PC with current technology that are far beyond the capabilities of today's machines. Even multi-GPU has outstanding promise in this line of thinking. Of course implementation is ridiculously hard, but still i think it is the future of how we compute.
Lee XT
(17 items)
 
  
CPUMotherboardGraphicsRAM
AMD FX-6300 Asus M5A97 SAPPHIRE Radeon HD 7850 AMD 4GB DDR3 1333MHZ 
RAMRAMRAMHard Drive
AMD 4GB DDR3 1333MHZ AMD 4GB DDR3 1333MHZ AMD 4GB DDR3 1333MHZ OCZ Vertex 4 256GB 
CoolingOSMonitorKeyboard
Corsair H80 Windows 8.1 Pro MCE Dell P2414H WHXV7  Microsoft Generic 
PowerCaseMouseMouse Pad
Ultra 600W Limited Edition NZXT Black Steel Razer Deathadder Razer Goliath 
Audio
Realtek HD Audio 
  hide details  
Reply
Lee XT
(17 items)
 
  
CPUMotherboardGraphicsRAM
AMD FX-6300 Asus M5A97 SAPPHIRE Radeon HD 7850 AMD 4GB DDR3 1333MHZ 
RAMRAMRAMHard Drive
AMD 4GB DDR3 1333MHZ AMD 4GB DDR3 1333MHZ AMD 4GB DDR3 1333MHZ OCZ Vertex 4 256GB 
CoolingOSMonitorKeyboard
Corsair H80 Windows 8.1 Pro MCE Dell P2414H WHXV7  Microsoft Generic 
PowerCaseMouseMouse Pad
Ultra 600W Limited Edition NZXT Black Steel Razer Deathadder Razer Goliath 
Audio
Realtek HD Audio 
  hide details  
Reply
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Hardware News
Overclock.net › Forums › Industry News › Hardware News › [CPC] Intel: CUDA will be just a 'footnote' in computing history