Epic Games Tim Sweeny Claims GPGPU Too Costly To Develop
Posted 08/16/09 at 05:25:31 PM by Justin Kerr
Epic Games’ CEO Tim Sweeney is on record as being stanchly against GPGPU computing in the past, but in a recent keynote delivered at the High Performance Graphics conference he further downplayed its future. From a developer standpoint he claims that GPGPU based applications can cost nearly 10x as much as a single threaded versions, with multi-core based software being the current sweet spot.
This isn’t the first time Sweeney has predicted the demise of GPGPU based computing technologies, but he has now further expanded his list of endangered technologies to include DirectX and OpenGL. In his speech last year Sweeny claimed that “In the next generation we’ll write 100-percent of our rendering code in a real programming language--not DirectX, not OpenGL, but a language like C++ or CUDA. Whether that runs on Nvidia hardware, Intel hardware or ATI hardware is really an independent question. You could potentially run it on any hardware that's capable of running general-purpose code efficiently."
Some might consider Sweeny’s comments a bit misguided considering that both Apple and Microsoft are strongly backing OpenCL, and ultimately if it turns out to be a more efficient way of doing certain tasks, couldn’t the development costs be justified? Clearly the GPU has future potential in the transcoding market, but do you think Sweeney has a point here?
Not that it really matters...
Submitted by LVmonkey on Mon, 08/17/2009 - 6:37am
but as a point of interest, what is with all the rainbow flags and kites in his office? hiding something while making a political statement? Or is he just not aware of what the rainbow is suppossed to proudly exclaim? lol
Is the sky hiding something
Submitted by robotsneedhugs2 on Mon, 08/17/2009 - 6:55am
Is the sky hiding something when the sun is shining on water droplets?
OMG!
Submitted by mesiah on Mon, 08/17/2009 - 11:44pm
Oh...My.....Gawd! I never realised... The sky is so totally gay!
Also, I don't think its a bunch of flags/kite, It looks like to me its one large kite or possibly a small sail. Kites like that are almost always brightly colored so they are easier to see in the sunlight.
Nice office he has there
Submitted by lunchbox73 on Mon, 08/17/2009 - 5:26am
Nice office he has there though.
proactive blamegame
Submitted by darkcactus1957 on Mon, 08/17/2009 - 4:50am
time to get working on new solutions.
That's interesting... I
Submitted by vistageek on Sun, 08/16/2009 - 8:21pm
That's interesting... I wonder if the cost of production is too high, or if the R n D is to expensive. BTW, could you have meant sweet spot, rather than sweat spot?
SWEET spot
Submitted by vulchan on Sun, 08/16/2009 - 10:49pm
Darn, you beat me to it :(
Why is it that of late
Submitted by I Jedi on Sun, 08/16/2009 - 7:29pm
Why is it that of late everytime I look there is some doom-sayer sprouting the end of the PC gaming desktop?
I say
Submitted by MeTo on Sun, 08/16/2009 - 4:17pm
I say do what you do to cut cost and deliver a good product. I don't need movie quality as long as it is a good game with a story.
He seems to rely more on conjecture than evidence each day
Submitted by QUINTIX256 on Sun, 08/16/2009 - 3:03pm
Really, which is more cost effect to develop: Going through a loop to do 100 dot products with 200 different operands, or telling a throughput device to just do them all in parallel? Data is rarely arranged that simply you say? Ok, how about a linked list with a variable number of pointers (links) at each node. Would it be easier with a serial architecture, where you have to go back and forth, verify left and right, and do complex navigation because the local register file is “short sighted”, and can only deal with so many items at once, or would you rather work with virtually as many items you want “on the table” as it where, instead of a few items at a time on a small plate?
You can have your recession. I'm not participating.
My two cents
Submitted by 1337Goose on Sun, 08/16/2009 - 4:24pm
I think we should distinguish application time from programmer time.
For a computer, dealing with data in parallel may be faster, but from a programmer's point of view, working in a more abstracted and higher level language saves hours of coding, debugging, and maintenance time.
I understand your point, and I may just be playing the devil's advocate; but I think GPU based parallel applications will be more of a niche market... at least until we revolutionize our understanding of the topic.
Feature
Review
Feature
Feature
Feature






