Loading...
Searching...
No Matches
Implementation of nv_dvp_bridge.h. Wraps NVIDIA's "GPUDirect for Video" (DVP) SDK so score's AJA GPU-direct output path can DMA encoder output from GPU memory to AJA-DMA-locked sysmem on Windows. More...
Detailed Description
Implementation of nv_dvp_bridge.h. Wraps NVIDIA's "GPUDirect for Video" (DVP) SDK so score's AJA GPU-direct output path can DMA encoder output from GPU memory to AJA-DMA-locked sysmem on Windows.
Reference implementation: AJA's ntv2glTextureTransferNV.cpp in 3rdparty/libajantv2/demos/NVIDIA/common/. This bridge collapses the demo's chunked-async transfer into one synchronous call per frame; the AJA strategies in score-plugin-gfx provide pipelining via their slot ring + the AJAConsumerThread instead of via DVP-side chunking.
Classes | |
| struct | NvDvpContext_t |
Macros | |
| #define | DVP_CHECK(call, ctx, retval) |
Macro Definition Documentation
◆ DVP_CHECK
| #define DVP_CHECK | ( | call, | |
| ctx, | |||
| retval | |||
| ) |
Value:
do \
{ \
DVPStatus _st = (call); \
if(_st != DVP_STATUS_OK) \
{ \
setError((ctx), #call, _st); \
return (retval); \
} \
} while(0)