Sign in with
Sign up | Sign in

More Bing From Badaboom

CUDA-Enabled Apps: Measuing Mainstream GPU Performance
By

Nvidia’s headliner in the CUDA space is Badaboom, the idiot-proof tool from Elemental Technologies that served to counter AMD’s hullabaloo with the ATI Avivo Video Converter. In a nutshell, you can import just about any input video type you could ask (even .m2ts from Blu-ray rips), pick your target device (from YouTube to the BlackBerry Bold), and out comes an optimized .mp4 file. It’s about as streamlined and simple as you could possibly make a transcoder, although that doesn’t mean dumbed down. Props go to Elemental and Nvidia for enabling multi-GPU functionality. You can’t span CUDA acceleration for one job across multiple GPUs, but you can run one instance of Badaboom 1.1 for each CUDA-enabled GPU in the system—handy stuff if you have to bang out a bunch of files in a hurry before skipping town.

If all you care about is transcoding and couldn’t give a rip about editing, then Badaboom remains our favorite pick of the day. Many other reviews have looked at the comparative quality of Badaboom’s transcoded visuals, and this isn’t the place to rehash such material. Suffice it to say that the end results look great, and Badaboom puts far less strain than most transcoders on the CPU. After all, most of us don’t have dedicated video systems, especially in our budget-conscious context here. In general, the higher the quality of the transcode, the more Badaboom seems to stress the GPU and leave the CPU to itself. So beyond the world of benchmark testing, this says a lot about the application as a real world solution. Odds are good that you’re going to want to transcode while doing other things, and you don’t want one task crippling all of the others.

Ask a Category Expert

Create a new thread in the UK Article comments forum about this subject

Example: Notebook, Android, SSD hard drive

Display all 6 comments.
This thread is closed for comments
  • 0 Hide
    mi1ez , 18 May 2009 16:32
    Good article. Very interesting. Pity you can't spell measuRing! :p 
  • 0 Hide
    Hammeh , 18 May 2009 22:17
    Great article!! I ran Seti@home on my GPU alone for quite a while. Although it is not a powerfull card in the slightest, a 8500GT, it still completed workunits faster than my CPU could (even more using the optimized apps). However, it has damaged my system as there are a few problems with running it. The heat generated from running the card processors at full all the time is a real pain. The fans on the lower end cards simply are not good enough to cool the cards, even if manually upping the fanspeed to 100% with RivaTuner. Secondly, if you allow it to use your GPU when you are using the computer, total utter graphic lagg even when reading webpages.
    Overall, it is a good technology, still has a few bugs which need to be solved and also the lack of support for ATI cards fustrates me. I am an avid BOINC user, I crunch Rosetta@home on all of my CPU's however I now let the GPU's run idle because the problems it causes. I will probably re-enable GPU usage if I upgrade my card in the future.
  • 0 Hide
    lumpy , 19 May 2009 03:40
    Cuda will never be suported by ati,its a Nvidia thing.
    Also an 8500gt is just not powerfull enough for cuda+multitasking a pc.IMO. Hell thats a under $50.00 gpu.
  • 0 Hide
    wild9 , 19 May 2009 15:53
    Lumpy, even an 8500 would be advantageous considering the speed hike..what's so bad about the CPU getting extra help from something that would otherwise just be displaying raster graphics or waiting for Crysis to fire up?
  • 0 Hide
    wild9 , 19 May 2009 15:56
    Quote:
    AMD now has a grave challenge at its feet.


    But look at AMD's platform solutions: highly integrated, highly efficient..can't say the same for Intel. I'd rather wait a bit longer for the software to be stable rather than use the GPGPU equivalent of an Intel GMA chipset trying to run GTA IV..
  • 1 Hide
    cybot_x1024 , 20 May 2009 01:55
    wild9But look at AMD's platform solutions: highly integrated, highly efficient..can't say the same for Intel. I'd rather wait a bit longer for the software to be stable rather than use the GPGPU equivalent of an Intel GMA chipset trying to run GTA IV..

    OpenCL is AMD's answer