Overclock.net › Forums › Industry News › Hardware News › [TI] AMD thinks most programmers will not use CUDA or OpenCL
New Posts  All Forums:Forum Nav:

[TI] AMD thinks most programmers will not use CUDA or OpenCL - Page 4

post #31 of 64
Well... The Folding@home project is planning to transition to OpenCL completely (even for their CPUs I think).
Dead Ryzen
(12 items)
 
Penguin Box
(9 items)
 
 
CPUMotherboardGraphicsRAM
AMD RYZEN 7 1800X  MSI B350M MORTAR ARCTIC AMD Radeon Vega Frontier Edition G.SKILL TridentZ Series 32GB DDR4 3200 (2 x 16G... 
Hard DriveHard DriveCoolingMonitor
SAMSUNG SSD 950 PRO 512GB Crucial MX300 2TB Noctua NH-U12S  BenQ XL2730Z - 27” 2560x1440 
KeyboardPowerCaseMouse
Das Keyboard 4 Professional Soft Tactile (MX Br... Seasonic X-850(SS-850KM3 Active PFC F3) Fractal Design Define Mini C BenQ ZOWIE EC1-A 
CPUMotherboardGraphicsRAM
Intel Core i7-2600K @4.6 GHz P8Z68-V AMD Radeon RX 480 (ref. design) Mushkin Enhanced Stealth 8GB (2 x 4GB) (PC3 17000) 
Hard DriveCoolingOSPower
SanDisk Extreme PRO 480GB Noctua NH-U12 S Arch Linux Seasonic SS-520FL 520W 
Case
Corsair Vengeance Series C70 
  hide details  
Reply
Dead Ryzen
(12 items)
 
Penguin Box
(9 items)
 
 
CPUMotherboardGraphicsRAM
AMD RYZEN 7 1800X  MSI B350M MORTAR ARCTIC AMD Radeon Vega Frontier Edition G.SKILL TridentZ Series 32GB DDR4 3200 (2 x 16G... 
Hard DriveHard DriveCoolingMonitor
SAMSUNG SSD 950 PRO 512GB Crucial MX300 2TB Noctua NH-U12S  BenQ XL2730Z - 27” 2560x1440 
KeyboardPowerCaseMouse
Das Keyboard 4 Professional Soft Tactile (MX Br... Seasonic X-850(SS-850KM3 Active PFC F3) Fractal Design Define Mini C BenQ ZOWIE EC1-A 
CPUMotherboardGraphicsRAM
Intel Core i7-2600K @4.6 GHz P8Z68-V AMD Radeon RX 480 (ref. design) Mushkin Enhanced Stealth 8GB (2 x 4GB) (PC3 17000) 
Hard DriveCoolingOSPower
SanDisk Extreme PRO 480GB Noctua NH-U12 S Arch Linux Seasonic SS-520FL 520W 
Case
Corsair Vengeance Series C70 
  hide details  
Reply
post #32 of 64
Quote:
Originally Posted by DrFPS View Post

Adobe is using CUDA,

Since CS6 Adobe has been using the mercury graphics engine that does not reqire CUDA to function optimally as it runs on both OpenGL and OpenCL without prejudice towards Nvidia or AMD thumb.gif

you can however use plugins that will utilise CUDA wink.gif
ULTIMA MKIII
(21 items)
 
  
CPUMotherboardGraphicsRAM
i7 4790k @4.6GHz ASUS Z97-A Gigabyte GTX 1070 G1 Gaming @ 2GHz (Boost) kingston Hyperx Fury 16GB 1866 
Hard DriveHard DriveHard DriveOptical Drive
Crucial MX200 250GB Crucial MX100 512GB ADATA Premier SP550 960GB TSSTcorp CDDVDW SH-S223F 
CoolingOSMonitorMonitor
Kraken x60 Windows 10 Pro 64bit  AOC I2477FWQ  AOC I2477FWQ 
MonitorKeyboardPowerCase
AOC I2477FWQ CM Storm Quickfire Ultimate (MX Brown) Corsair HX750i NZXT Switch 810 
MouseMouse PadAudioAudio
Corsair M65 Corsair MM200 Extended Asus Xonar U7 (External) Logitech Z906 5.1 
Audio
Steelseries Siberia V3 Headset 
  hide details  
Reply
ULTIMA MKIII
(21 items)
 
  
CPUMotherboardGraphicsRAM
i7 4790k @4.6GHz ASUS Z97-A Gigabyte GTX 1070 G1 Gaming @ 2GHz (Boost) kingston Hyperx Fury 16GB 1866 
Hard DriveHard DriveHard DriveOptical Drive
Crucial MX200 250GB Crucial MX100 512GB ADATA Premier SP550 960GB TSSTcorp CDDVDW SH-S223F 
CoolingOSMonitorMonitor
Kraken x60 Windows 10 Pro 64bit  AOC I2477FWQ  AOC I2477FWQ 
MonitorKeyboardPowerCase
AOC I2477FWQ CM Storm Quickfire Ultimate (MX Brown) Corsair HX750i NZXT Switch 810 
MouseMouse PadAudioAudio
Corsair M65 Corsair MM200 Extended Asus Xonar U7 (External) Logitech Z906 5.1 
Audio
Steelseries Siberia V3 Headset 
  hide details  
Reply
post #33 of 64
Quote:
Originally Posted by KyadCK View Post

You don't find it sad that programmers ignore OpenCL? I can understand ignoring something proprietary not being sad (CUDA), but OpenCL is an open standard everyone can run.
Everyone can technically run.

...the issue is how many people have the knowledge and skill to implement it? An developer with 10+ years experienced with GPGPU skills will cost easily $100K a year.
Quote:
Originally Posted by Eatfoodnow View Post

Any time when hardware goes unused because of programming shortcomings, it's sad. It's like modern programs that don't use more than two cores effectively.
Multi-threading is hard especially once you go on problems that are not "embarrassing parallel". A lot of applications just don't benefit enough from being heavily threaded.

Quote:
Originally Posted by gamer11200 View Post

Give it time. OpenCL is still maturing. After a few years, the use and development of OpenCL will become larger and many programmers will be jumping onto it.
Not until the hardware is abstracted more. Developers still have to understand GPU architecture to optimize code. Programmers of that caliber are expensive.
Quote:
Originally Posted by DrFPS View Post


Its not sadly. The joke is on AMD and of course the will try to distract you with their thoughts.

Thought is one thing. Facts on the hand is a different story. Adobe is using CUDA, Autocad/autodesk is using cuda, Harvard Medical School, The National Airspace System, Google, Microsoft,
Nvidia has put a lot in to this. Where else can you get a free programming language, that all you hvae to do is register and you can learn how to use it? Not to mention all the apps already developed, and or are being devloped.

That comes with the video card, free of charge.


CUDA is bringing forth the GPGPU revolution.

https://developer.nvidia.com/technologies/languages_apis
Actually, OpenCL is gaining ground. I saw a chart of scientific papers using CUDA vs OpenCL.... CUDA dominates but OpenCL is increasing.
Quote:
Originally Posted by SCollins View Post

In the lates 80's into the 90's, allot of people who aren't good enough at programming, got into programming. Anyone with a degree a pulse and 2 hands that could form a coherent sentence got a job as a programmer, in the late 90's it got even worse. Companies like microsoft would practicially pay your tution to get you to get a programming degree.

sadly, theres allot of really crapy sub optimal programmers out there, and becuase the bar is so low, its going to stay that way for a long time.
That's because programming is easy. Design and architecture is hard so sub-optimal programmers don't do advance design.... (except for quants/physicists/mathematicians who think they can but I digress).

Furthermore, multi-threading can be very hard for non-trivial applications. Testing multi-threading is even harder. It comes down to cost-benefit.
Once again...
(13 items)
 
  
CPUMotherboardGraphicsRAM
i7 920 [4.28GHz, HT] Asus P6T + Broadcom NetXtreme II VisionTek HD5850 [900/1200] + Galaxy GT240 2x4GB G.Skill Ripjaw X [1632 MHz] 
Hard DriveOSMonitorKeyboard
Intel X25-M 160GB + 3xRAID0 500GB 7200.12 Window 7 Pro 64 Acer H243H + Samsung 226BW XARMOR-U9BL  
PowerCaseMouseMouse Pad
Antec Truepower New 750W Li Lian PC-V2100 [10x120mm fans] Logitech G9 X-Trac Pro 
  hide details  
Reply
Once again...
(13 items)
 
  
CPUMotherboardGraphicsRAM
i7 920 [4.28GHz, HT] Asus P6T + Broadcom NetXtreme II VisionTek HD5850 [900/1200] + Galaxy GT240 2x4GB G.Skill Ripjaw X [1632 MHz] 
Hard DriveOSMonitorKeyboard
Intel X25-M 160GB + 3xRAID0 500GB 7200.12 Window 7 Pro 64 Acer H243H + Samsung 226BW XARMOR-U9BL  
PowerCaseMouseMouse Pad
Antec Truepower New 750W Li Lian PC-V2100 [10x120mm fans] Logitech G9 X-Trac Pro 
  hide details  
Reply
post #34 of 64
Quote:
Originally Posted by DuckieHo View Post


Actually, OpenCL is gaining ground. I saw a chart of scientific papers using CUDA vs OpenCL.... CUDA dominates but OpenCL is increasing.

 

That's encouraging. Mind posting the link?

post #35 of 64
Quote:
Originally Posted by TFL Replica View Post

That's encouraging. Mind posting the link?

http://hgpu.org/?page_id=3529
(Note: 2013 isn't done yet so the chart's last data point is skewed.)
Once again...
(13 items)
 
  
CPUMotherboardGraphicsRAM
i7 920 [4.28GHz, HT] Asus P6T + Broadcom NetXtreme II VisionTek HD5850 [900/1200] + Galaxy GT240 2x4GB G.Skill Ripjaw X [1632 MHz] 
Hard DriveOSMonitorKeyboard
Intel X25-M 160GB + 3xRAID0 500GB 7200.12 Window 7 Pro 64 Acer H243H + Samsung 226BW XARMOR-U9BL  
PowerCaseMouseMouse Pad
Antec Truepower New 750W Li Lian PC-V2100 [10x120mm fans] Logitech G9 X-Trac Pro 
  hide details  
Reply
Once again...
(13 items)
 
  
CPUMotherboardGraphicsRAM
i7 920 [4.28GHz, HT] Asus P6T + Broadcom NetXtreme II VisionTek HD5850 [900/1200] + Galaxy GT240 2x4GB G.Skill Ripjaw X [1632 MHz] 
Hard DriveOSMonitorKeyboard
Intel X25-M 160GB + 3xRAID0 500GB 7200.12 Window 7 Pro 64 Acer H243H + Samsung 226BW XARMOR-U9BL  
PowerCaseMouseMouse Pad
Antec Truepower New 750W Li Lian PC-V2100 [10x120mm fans] Logitech G9 X-Trac Pro 
  hide details  
Reply
post #36 of 64
Quote:
Originally Posted by DuckieHo View Post


http://hgpu.org/?page_id=3529
(Note: 2013 isn't done yet so the chart's last data point is skewed.)

 

Much appreciated! thumb.gif

post #37 of 64
This is where specialized/closed-source software can really make a difference. Most developers don't use CUDA or OpenCL, but at least Nvidia is profiting from those that do use CUDA. Not so with AMD and OCL.
Po' Pimpin'
(11 items)
 
  
CPUMotherboardRAMHard Drive
i5 2500k @ stock Biostar TZ68K+ [A3] 4GB  Sandforce 1222 64GB SSD 
Optical DriveCoolingOSMonitor
LG 22x DVD-+RW  Stock Windows 7 x64 Acer S211HL 1080p 
PowerCaseMouse
600w Diablotek Linkworld Electronic Inland 
  hide details  
Reply
Po' Pimpin'
(11 items)
 
  
CPUMotherboardRAMHard Drive
i5 2500k @ stock Biostar TZ68K+ [A3] 4GB  Sandforce 1222 64GB SSD 
Optical DriveCoolingOSMonitor
LG 22x DVD-+RW  Stock Windows 7 x64 Acer S211HL 1080p 
PowerCaseMouse
600w Diablotek Linkworld Electronic Inland 
  hide details  
Reply
post #38 of 64
Quote:
Originally Posted by DuckieHo View Post

http://hgpu.org/?page_id=3529
(Note: 2013 isn't done yet so the chart's last data point is skewed.)

nice find smile.gif

Given the compute monster GCN is, it makes sense... see the very popular application of VLIW GPUs (HD 5000 series, HD6000 series) for password cracking but with more efficiency
Workstation stuff
(407 photos)
SpecViewperf 12.0.1
(158 photos)
 
Reply
Workstation stuff
(407 photos)
SpecViewperf 12.0.1
(158 photos)
 
Reply
post #39 of 64
Quote:
Originally Posted by DuckieHo View Post

Everyone can technically run.

...the issue is how many people have the knowledge and skill to implement it? An developer with 10+ years experienced with GPGPU skills will cost easily $100K a year.

Multi-threading is hard especially once you go on problems that are not "embarrassing parallel". A lot of applications just don't benefit enough from being heavily threaded.

Not until the hardware is abstracted more. Developers still have to understand GPU architecture to optimize code. Programmers of that caliber are expensive.

That's because programming is easy. Design and architecture is hard so sub-optimal programmers don't do advance design.... (except for quants/physicists/mathematicians who think they can but I digress).

Furthermore, multi-threading can be very hard for non-trivial applications. Testing multi-threading is even harder. It comes down to cost-benefit.

I came across this interesting programming language: http://en.wikipedia.org/wiki/ParaSail_%28programming_language%29

http://www.technologyreview.com/news/424836/new-language-for-programming-in-parallel/

Essentially the compiler is designed to do the multi-threading task instead of the programmer.
post #40 of 64
Quote:
Originally Posted by DrFPS View Post


Its not sadly. The joke is on AMD and of course the will try to distract you with their thoughts.

Thought is one thing. Facts on the hand is a different story. Adobe is using CUDA, Autocad/autodesk is using cuda, Harvard Medical School, The National Airspace System, Google, Microsoft,
Nvidia has put a lot in to this. Where else can you get a free programming language, that all you hvae to do is register and you can learn how to use it? Not to mention all the apps already developed, and or are being devloped.

That comes with the video card, free of charge.


CUDA is bringing forth the GPGPU revolution.

https://developer.nvidia.com/technologies/languages_apis

evidently there is a little life left in C++ amp. Not sure how much but.....
http://blogs.msdn.com/b/nativeconcurrency/
Stumpy
(13 items)
 
  
CPUMotherboardGraphicsRAM
A10 5800k gigabyte  7970 mushkin ddr3  
Hard DriveOSPower
intel ssd Win7 64 Corsair TX850W 
  hide details  
Reply
Stumpy
(13 items)
 
  
CPUMotherboardGraphicsRAM
A10 5800k gigabyte  7970 mushkin ddr3  
Hard DriveOSPower
intel ssd Win7 64 Corsair TX850W 
  hide details  
Reply
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Hardware News
Overclock.net › Forums › Industry News › Hardware News › [TI] AMD thinks most programmers will not use CUDA or OpenCL