GTX 470 wont work with Badaboom please help

i have a pair of gtx 470 but badaboon keep showing a error message when ever i try to import a movie for convert but when i switch back to 275 it work ok so is it 470 too new for badaboom?

I think Elemental updated Badaboom for Fermi, have you tried their latest version, 1.2.1?

REALLY? i dont see it

1.2.1.74 still dont work…

Badaboom 1.2.1.74 doesn’t work for me also. I have GTX 480 and drivers 197.41 also tried older 197.17 and 197.44 , makes no difference.

Here is badaboom.log .

[codebox]GPU 0: NVIDIA GeForce GTX 480

GPU 1: NVIDIA GeForce GTX 480

Compute Capability 2.0.

GeForce GTX 480 GPU selected

Num Multi Proc : 15

Clock rate : 810000

Video Codec: M

Audio Codec: M

psInputInfo =

eCodec = 7

dwFourCC = 0 (	sResolution = 352 x 272

eColorFmt = 1

iCmdQP = 0

fFrameRate = 23.976

fSAR = 1.000

fDAR = 0.000

bTelecine = 0

sBitRate = ( 0 0 0 0 0 )

uAvgBitRate = 0

sGOPControl = (0 0 0)

uLevel = 0

uProfile = 0

bInterlace = 0

eFrameStructure = 0

psOutputInfo =

eCodec = 0

dwFourCC = 0 (	sResolution = 280 x 194

eColorFmt = 3

iCmdQP = 0

fFrameRate = 23.976

fSAR = 0.000

fDAR = 0.000

bTelecine = 0

sBitRate = ( 0 0 0 0 -1 )

uAvgBitRate = 0

sGOPControl = (0 0 0)

uLevel = 0

uProfile = 0

bInterlace = 0

eFrameStructure = 0

Codec extensions not recognized!

CRITICAL:gpuCtrlLoadModule: CUDA_ERROR_INVALID_SOURCE

FATAL:Could not load image proc kernel(s)

CPU Logical Processors Detected: 8

CPU Cores per processor: 1

CPU Processors Detected: 4

FATAL:[/codebox]
Badaboom_Log.txt (1001 Bytes)

hey i got the same message too
i was hoping the 470 can out run the gtx 295 but now it aint do crap on the badaboom

It should be recompiled for 4xx most likely, wait a bit.

Hopefully someone in this thread is emailing the badaboom developers directly…

This is a known issue. The developers of Badaboom will be releasing a newer version of the software which will allow it to work with Geforce GTX 480/470 series. The only other CUDA app which requires an update to work with the Geforce GTX 480/470 is TMPGEnc which will also get an update in the future.

I found this quote from another forum maybe its helpful, It helped me to decide not to buy a GTX4xx card until Q4 2010…

“Some of you are probably wondering whether Badaboom will support the latest NVIDIA GTX 480 and GTX 470 cards, code named “GF100”. Well, the short answer is yes, but it’s not ready yet. Badaboom uses an earlier version of Elemental’s video engine code base. It is likely that this engine is using some older constructs that are no longer supported with more recent CUDA releases. In addition to this, there are likely changes in the GF100’s error handling relative to the GT200 line of cards. In terms of addressing these issues, Elemental is currently focused on our enterprise product line, and this is where our GPU development resources are completely committed right now. We’ve made major updates to our underlying video processing engine, which improve on features and functionality along with optimizations for GF100. We are first integrating that into our enterprise products. Support for this engine within Badaboom is currently expected to be available in Q4 of 2010.”

[url=“http://forums.amd.com/forum/messageview.cfm?catid=12&threadid=132776”]http://forums.amd.com/forum/messageview.cf...threadid=132776[/url]

How did tweaktown get it to work???

[url=“http://www.tweaktown.com/articles/3204/nvidia_geforce_gtx_470_video_card_overclocked/index15.html”]http://www.tweaktown.com/articles/3204/nvi...ed/index15.html[/url]

Technically, this sounds like Badaboom’s code has a bug which segfaults in shared memory. That’s silently ignored on G200 but (properly) crashes on GF100. There aren’t any “older constructs” you need to worry about with GF100 in CUDA anyway. (Perhaps there is something if they’re using some of the non-CUDA features like PureVideo?)

Yeah, frankly, I’m amazed that they say it will take them until Q4 to support Fermi. Unless they have some crazy and unsupported hacks (or previously silent bugs) in their code, I would expect their CUDA kernels to run out of the box on the new architecture. Tuning things to run well and take advantage of the Fermi cache would obviously take longer, but instant crashing suggests problems that should be fixed sooner rather than later.

People at overclock.net suspect tweaktown cooked this article. They think a normal 470 reference card can also overclock to the same performance as Galaxy 470 GC. So they are not sure why the glowing review at tweaktown. :rolleyes: They do agree that fan added by Galaxy is a decent addition to keep the temp low.