Hi Vasco,
I tried everything and still no luck. I'm pretty sure just copying inhits
and outhits is not enough. Going by Sean's explanation of segments and
partitions, it looks like I have no option but to implement partitions on
the GPU.
Unless you know what is wrong or have an alternative approach, I feel I
should go ahead with GPU partition structure. Let me know what you feel.
Thank you,
Param
On Mon, Jun 27, 2016 at 6:54 PM Vasco Alexandre da Silva Costa <
vasco.co...@gmail.com> wrote:
> On Mon, Jun 27, 2016 at 11:33 AM, Param Hanji <param.catchch...@gmail.com>
> wrote:
>
>>
>> Also the GPU side seg_sti doesn't have the same meaning at all. It's a
>>> unique solid id. In the soltabs list. So '0' would be the first valid solid
>>> in the soltabs list. So your code should work like this:
>>>
>>> unsigned int i=0;
>>> RT_VISIT_ALL_SOLTABS_START(stp, a->a_rt_i) {
>>> /* Ignore "dead" solids in the list. (They failed prep) */
>>> if (stp->st_aradius <= 0) continue;
>>> /* Infinite solids make the BVH construction explode. */
>>> if (stp->st_aradius >= INFINITY) continue;
>>>
>>> if (i == gpu_seg->seg_sti) {
>>> cpu_seg->seg_stp = stp;
>>> break;
>>> }
>>> i++;
>>> } RT_VISIT_ALL_SOLTABS_END
>>>
>>>
>>> Yup, this worked! Just one more thing, how do I perform the reverse of
>> this i.e. how do I set seg_sti of the GPU segment given the CPU storage's
>> stp pointer?
>>
>> I believe this is the last missing piece. If even this doesn't work, I'll
>> have define and use partitions on the GPU, only segments just won't cut it.
>>
>
> That should be pretty easy. You just do the same loop and check if the
> iterator stp == the pointer to the soltab you want to find. You can just
> compare the pointers to see if they point to the same memory region.
>
> If we actually stored the dead solids on the GPU you would be able to do
> the same thing by checking the 'stp->st_bit value'. And you should be able
> to get the pointer to a solid like this: 'struct soltab *stp =
> rtip->rti_Solids[seg_sti]'. As it is now the indexes may mismatch...
> So to make the code faster without these loops we can either store the
> dead solids, and mark them somehow to ignore them on traversal, or we can
> create an array that converts from one index (i.e. stp->st_bit) to the
> other (i.e. gpu_seg->seg_sti).
>
> Regards,
>
> --
> Vasco Alexandre da Silva Costa
> PhD in Computer Engineering (Computer Graphics)
> Instituto Superior Técnico/University of Lisbon, Portugal
>
> ------------------------------------------------------------------------------
> Attend Shape: An AT&T Tech Expo July 15-16. Meet us at AT&T Park in San
> Francisco, CA to explore cutting-edge tech and listen to tech luminaries
> present their vision of the future. This family event has something for
> everyone, including kids. Get more information and register today.
> http://sdm.link/attshape_______________________________________________
> BRL-CAD Developer mailing list
> brlcad-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/brlcad-devel
>
------------------------------------------------------------------------------
Attend Shape: An AT&T Tech Expo July 15-16. Meet us at AT&T Park in San
Francisco, CA to explore cutting-edge tech and listen to tech luminaries
present their vision of the future. This family event has something for
everyone, including kids. Get more information and register today.
http://sdm.link/attshape
_______________________________________________
BRL-CAD Developer mailing list
brlcad-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/brlcad-devel
