backtop


Print 38 comment(s) - last by R0B0Ninja.. on Aug 8 at 7:42 AM

NVIDIA says its GPUs are in fact programmable in C language

Yesterday, DailyTech ran a story about details on Intel's upcoming Larrabee architecture for the graphics market. One of Intel's most important talking points when it plays up the benefits of Larrabee over NVIDIA's GPUs is the fact that NVIDIA's GPUs require developers to learn a new programming language called CUDA.

Intel says that with its Larrabee architecture developers can simply program in C or C++ languages for just as they would for any other x86 processor. According to Intel, the ability to program Larrabee with C or C++ makes it much easier for developers to port applications from other platforms to the Larrabee architecture.

After DailyTech ran the story, NVIDIA wanted to address what it considers to be misinformation when it comes to CUDA. NVIDIA says:

CUDA is a C-language compiler that is based on the PathScale C compiler. This open source compiler was originally developed for the x86 architecture. The NVIDIA computing architecture was specifically designed to support the C language - like any other processor architecture. Competitive comments that the GPU is only partially programmable are incorrect - all the processors in the NVIDIA GPU are programmable in the C language.

NVIDIA's approach to parallel computing has already proven to scale from 8 to 240 GPU cores. Also, NVIDIA is just about to release a multi-core CPU version of the CUDA compiler. This allows the developer to write an application once and run across multiple platforms. Larrabee's development environment is proprietary to Intel and, at least disclosed in marketing materials to date, is different than a multi-core CPU software environment.

Andrew Humber from NVIDIA distilled things a bit further saying, "CUDA is just our brand name for the C-compiler. They aren't two different things."

Humber also pointed out that at NVIDIA's financial analyst day in April it showed an astrophysics simulation running on integrated graphics with an eight-core GPU, a GeForce 8 series GPU with 128 cores and a quad-core CPU. NVIDIA says that the demonstration used exactly the same binary program across the range of GPUs and the exact same source code for the CPU and GPU.



Comments     Threshold


This article is over a month old, voting and posting comments is disabled

To quote myself from yesterday....
By DanoruX on 8/5/2008 1:41:39 PM , Rating: 2
http://www.dailytech.com/article.aspx?newsid=12567...

quote:
1. Cuda is based on C.
2. Computer graphics (today) is mostly done in HLSL and GLSL which are both also based on C.
3. Above languages are really easy to learn.


:-)




RE: To quote myself from yesterday....
By Master Kenobi (blog) on 8/5/2008 4:12:58 PM , Rating: 3
I would like to point out that "Based on C" and "C/C++" are not the same thing. I can say C# is "Baced on C" but anyone who has ever programmed in C# would know that my statement is half baked at best.


RE: To quote myself from yesterday....
By Master Kenobi (blog) on 8/5/2008 4:14:25 PM , Rating: 2
That should be Based* wow I can't spell today. Ready to go home and relax eh?


By chmilz on 8/5/2008 4:43:10 PM , Rating: 2
Quick, someone program Tuesday to be the new Friday!


By JakLee on 8/5/2008 4:55:00 PM , Rating: 5
All your BASED are belong to us.


RE: To quote myself from yesterday....
By DanoruX on 8/5/2008 4:39:29 PM , Rating: 2
Let me correct the semantics then - "closely related" to C, C-like, etc.

Better? :P

Oh and yes, C# is very different (though very awesome in its own way).


By jvillaro on 8/6/2008 12:31:49 AM , Rating: 1
C# is more "closely related" to JAVA actually :) with a little bit of C++

But it's the most "awesomest" of them all ;)


By DeepBlue1975 on 8/5/2008 5:18:10 PM , Rating: 2
I'm sure I wouldn't wanna touch one of those half baked arguments even with a 10 meter pole, have enough digestive disorders by the moment :D


By MisterAnderson42 on 8/5/2008 5:23:45 PM , Rating: 5
CUDA is MUCH closer to C than you are assuming. The only additions NVIDIA made to C were some built in variables and a new syntax to call a function specifying the grid of threads it executes on the GPU.

Basically, any C code goes in the kernel: enums, structs, pointer arithmetic, switch statments, etc... everything in the C syntax just works. NVIDIA wasn't joking when they said as quoted in the article above that nvcc is just a modified C compiler :) C# definately is not that.

If you don't believe me, check out the programming guide: http://www.nvidia.com/object/cuda_get.html

As for Larabee: I'll wait and see when they release their programming guide before I make a judgment. If they just provide a C language with big 16-wide SSE instructions for math operations I will be very disappointed because that will be an absolute pain to program compared to CUDA which has one thread working on each SP and automatically handles divergences.


RE: To quote myself from yesterday....
By Flunk on 8/5/2008 7:53:09 PM , Rating: 5
I don't think Microsoft ever claimed C# was based on C. I think the term "C-like syntax" is what they were marketing.


By drmo on 8/6/2008 7:37:56 AM , Rating: 2
Intel's compiler will also contain the much loved feature of: if not GenuineIntel then suckass.


By TomZ on 8/7/2008 9:03:50 AM , Rating: 2
quote:
I don't think Microsoft ever claimed C# was based on C. I think the term "C-like syntax" is what they were marketing.

C# is clearly based on Java, C++, and C before it. To deny that is like denying the existance of gravity.


You sure about that?
By marsbound2024 on 8/5/2008 1:38:05 PM , Rating: 5
quote:
and a quad-core CPU with four cores...


Pretty revolutionary, that.




RE: You sure about that?
By Sulphademus on 8/5/2008 1:42:18 PM , Rating: 5
Department of Redundancy Department.


RE: You sure about that?
By Brandon Hill (blog) on 8/5/2008 1:42:29 PM , Rating: 5
Not as revolutionary as a dozen of Krispy Kreme doughnuts which come 12 to a box.


RE: You sure about that?
By Oregonian2 on 8/5/2008 6:04:01 PM , Rating: 2
144 doughnuts are a lot of doughnuts!

:-)

Although I know what you meant to say...

:-)


RE: You sure about that?
By killerroach on 8/5/2008 10:16:27 PM , Rating: 5
quote:
144 doughnuts are a lot of doughnuts!

:-)

Although I know what you meant to say...

:-)


That's just [a] gross. :)


RE: You sure about that?
By JonnyDough on 8/7/2008 12:02:29 AM , Rating: 2
Not to a cop! Nothing against cops, unless they're here on behalf of the RIAA. Then I might not be too happy with them.


RE: You sure about that?
By tmouse on 8/6/2008 3:49:32 PM , Rating: 2
Thats why I always prefer a bakers dozen. ;)


RE: You sure about that?
By Amiga500 on 8/5/2008 1:49:45 PM , Rating: 2
Could have been saved with an emergency switch to:

and a quad-core CPU with four threads

But alas, it was too late - the sentence was confined to the grammatical underworld where half of my posts belong.


I'd be more happy
By FITCamaro on 8/5/2008 2:09:00 PM , Rating: 3
If they put up the Physx drivers for the GeForce 8 series that they promised us for today.




RE: I'd be more happy
By voodooboy on 8/5/2008 2:13:00 PM , Rating: 2
Your wish has been granted! Head over to http://www.laptopvideo2go.com/ for some yummy CUDA drivers! :)


RE: I'd be more happy
By voodooboy on 8/5/2008 2:46:01 PM , Rating: 3
Or better yet...a direct link to the drivers from the nVidia website:

http://www.nvidia.com/object/thankyou.html?url=/co...


RE: I'd be more happy
By FITCamaro on 8/5/2008 3:48:46 PM , Rating: 2
Oh. I was expecting this to be a ForceWare driver from their main download site. Do you just need the 177.35 driver or do you need the toolkit as well?


RE: I'd be more happy
By voodooboy on 8/5/2008 4:28:12 PM , Rating: 2
I'd say you just need the driver and install the latest version of the PhysX driver (also on the www.laptopvideo2go.com website) and that's about it...UNLESS you intend developing CUDA programs.. :)


Interesting limitations
By Khato on 8/5/2008 5:10:59 PM , Rating: 5
http://en.wikipedia.org/wiki/CUDA

I'm too lazy to confirm them, but if true the lack of recursive functions and full IEEE 754 support are somewhat disappointing. Then, of course, there's the fact that if you're wanting to do graphics in C, there's the slight issue of all that fixed function hardware being unavailable and perhaps not able to be bypassed - believe that's what the note on texture rendering being unsupported is indicating.




ATi's card do the same
By Rodney McNaggerton on 8/5/2008 6:29:24 PM , Rating: 5
Ati's cards are just as programmable through CAL, except they don't lock out certain parts of their programming unless you pay, like Nvidia does.




cheapshot
By Samus on 8/6/2008 4:22:10 AM , Rating: 2
come on intel, you can do better than that.




folding@home
By R0B0Ninja on 8/8/2008 7:42:17 AM , Rating: 2
When will we be able to utilise our 8800s for folding@home?




Obviously You Got...
By mmatis on 8/5/08, Rating: -1
RE: Obviously You Got...
By Oregonian2 on 8/5/2008 6:05:15 PM , Rating: 4
What about Intel's architecture is a lie?

I missed your point.


RE: Obviously You Got...
By mmatis on 8/5/08, Rating: -1
RE: Obviously You Got...
By Yossarian22 on 8/5/08, Rating: 0
RE: Obviously You Got...
By 4wardtristan on 8/6/2008 2:20:11 AM , Rating: 2
and you think you comming on here and flaming the dude above you is any better?

wake up already.


RE: Obviously You Got...
By vandalizmo on 8/6/2008 5:52:59 AM , Rating: 2
I think `Liarbee' is funny :)


RE: Obviously You Got...
By Oregonian2 on 8/6/2008 1:54:20 PM , Rating: 2
Yes, that was the total value of his assertion -- even though it had no validity whatsoever.


RE: Obviously You Got...
By mmatis on 8/6/08, Rating: 0
RE: Obviously You Got...
By Oregonian2 on 8/6/2008 1:53:00 PM , Rating: 2
For your characterization of the Intel quote to be false, that means that Intel providing support for languages OTHER than C and C++ would have made it easier "for developers to port applications from other platforms to the Larrabee architecture".

I think your assertion is completely wrong and that C/C++ do in fact make it easier than any other language for developers as Intel asserts.


"Nowadays you can buy a CPU cheaper than the CPU fan." -- Unnamed AMD executive

Related Articles
Intel Talks Details on Larrabee
August 4, 2008, 12:46 PM













botimage
Copyright 2014 DailyTech LLC. - RSS Feed | Advertise | About Us | Ethics | FAQ | Terms, Conditions & Privacy Information | Kristopher Kubicki