Go to advanced search

by Andrew Holme
Sun Jan 25, 2015 11:02 pm
Forum: C/C++
Topic: Feed ALSA capture into GPU_FFT
Replies: 5
Views: 2285

Re: Feed ALSA capture into GPU_FFT

Shouldn't this line:
fft_compute_forward(buffer, NLOG2, OutData, 1);
be:
fft_compute_forward(float_buffer, NLOG2, OutData, 1);

?
by Andrew Holme
Fri Jan 02, 2015 10:13 pm
Forum: Bare metal, Assembly language
Topic: V3D and the L2 Cache
Replies: 1
Views: 1246

Re: V3D and the L2 Cache

AXI memory requests from the V3D are routed either through or around the system L2 cache, according to which of the 4 address aliases is used. OpenGL uses the uncached alias. By default, GPU_FFT uses the non-allocating cache coherent ailias, which means requests pass through the cache. There is a #d...
by Andrew Holme
Sun Dec 21, 2014 4:49 pm
Forum: Advanced users
Topic: GPU_FFT troubleshooting (moved from beginners)
Replies: 17
Views: 4295

Re: GPU_FFT troubleshooting (moved from beginners)

That return value 2147483648 (0x80000000) is a timeout error. You need to increase the timeout in gpu_fft_base.c:

#define GPU_FFT_TIMEOUT 1000 // ms

Computing six 1M-point transforms takes longer than one second. The timeout could be chosen more intelligently and is not adequately documented.
by Andrew Holme
Sun Dec 21, 2014 10:54 am
Forum: Bare metal, Assembly language
Topic: Host interupt on qpu (ra38)
Replies: 2
Views: 1388

Re: Host interupt on qpu (ra38)

Andy,

The QPU interrupt is handled by an interrupt service routine in start.elf which clears the interrupt flags.

Regards,
Andrew.

Go to advanced search