Another scope of optimization which I feel is to use atomic operations. I
realized this while reading the rt.cl code. I feel there is a scope of
using atomic operations to boost.
What are your views regarding this?
Regards
Shailesh
Shailesh Tripathi
B.Tech. Part-IV
Electronics Engineering
IIT-BHU (Varanasi)
On Sun, Apr 2, 2017 at 1:58 AM, Shailesh Tripathi <
shailesh.tripathi.ec...@itbhu.ac.in> wrote:
>
>> - DSP
>> - HLF
>>
>> *HALF not included (sorry for typo).
>
>>
>> - HYP
>> - SUPERELL
>> - EBM
>> - EXTRUDE
>> - REVOLVE
>> - HF (I am not very sure)
>> - SUBMODEL (a bit difficult one)
>>
>>
>
> Shailesh Tripathi
> B.Tech. Part-IV
> Electronics Engineering
> IIT-BHU (Varanasi)
>
> On Sat, Apr 1, 2017 at 9:05 AM, Shailesh Tripathi <
> shailesh.tripathi.ec...@itbhu.ac.in> wrote:
>
>> Hello Vasco Sir,
>> Thanks a lot for your consideration.
>> So let me update the list of primitives which I am planning to port:
>>
>> - DSP
>> - HLF
>> - HYP
>> - SUPERELL
>> - EBM
>> - EXTRUDE
>> - REVOLVE
>> - HF (I am not very sure)
>> - SUBMODEL (a bit difficult one)
>>
>> As you described above, I feel these many primitives would be enough to
>> keep me engaged for the 3 months.
>>
>> Do let me know your suggestions.
>>
>> Also, I am active on irc.freenode with handle "shailesh". Please tell me
>> when you are free.
>>
>>
>> Regards
>>
>> Shailesh
>>
>> Shailesh Tripathi
>> B.Tech. Part-IV
>> Electronics Engineering
>> IIT-BHU (Varanasi)
>>
>> On Sat, Apr 1, 2017 at 8:27 AM, Vasco Alexandre da Silva Costa <
>> vasco.co...@gmail.com> wrote:
>>
>>> Hello,
>>> Like I said before, on our prior conversation, delete your first two
>>> pages. You don't need a title page or a contents page on a short proposal
>>> document. The written content seems a bit verbose and redundant but it's
>>> mostly ok.
>>>
>>> Don't use Comic Sans (it's an offense to font design!). Arial is already
>>> bad enough. Use justified mode as it makes following text on long
>>> paragraphs easier.
>>>
>>> RHC and RPC are already implemented, like you noted, and I've desk
>>> checked their patches and they're ok. I just haven't found the time to test
>>> and commit them yet. So you can remove them from your document.
>>>
>>> Some of the primitives are going to be harder to implement than
>>> considered in your plan. I'll give you an example. The HALF (Half Space)
>>> primitive has basically infinite volume so it doesn't play nicely with the
>>> BVH (Bounding Volume Hierarchy) object partitioning acceleration structure
>>> we currently have. For this reason some 3D solid modeling tools won't even
>>> allow that primitive.
>>>
>>> Some of the primitives are composite primitives (e.g. SUBMODEL) and this
>>> means their implementation will be more subtle.
>>>
>>> EXTRUDE and REVOLVE aren't included in your list of primitives to
>>> implement and those are typically fairly common solid modeling operations.
>>>
>>> Other primitives, like the METABALL, are experimental, possibly buggy,
>>> and I don't know if anyone even uses them. You would have to ask someone
>>> who's been along here longer to know it for sure.
>>>
>>> DSP (Displacement Map), HF (Heightfield), and EBM (Extruded Bitmap) can
>>> probably be merged together for the most part as their code is relatively
>>> similar. The problem is with backwards compatibility with existing code.
>>> AFAIK DSP and HF are already using the same algorithm on the default code
>>> path.
>>>
>>> VOL (3D Volume) is not similar to a quadric at all. If anything it's
>>> similar to the HF (Heightfield) code because it also uses grid traversal.
>>>
>>> Regards,
>>>
>>>
>>> On Sat, Apr 1, 2017 at 1:13 AM, Shailesh Tripathi <
>>> shailesh.tripathi.ec...@itbhu.ac.in> wrote:
>>>
>>>> My proposal: https://docs.google.com/docume
>>>> nt/d/1cmIVxU22cFENhS4OvE2JCpIJy_g4ZRYsNIPyklTC1i8/edit?usp=sharing
>>>>
>>>>
>>>> Shailesh Tripathi
>>>> B.Tech. Part-IV
>>>> Electronics Engineering
>>>> IIT-BHU (Varanasi)
>>>>
>>>> On Sat, Apr 1, 2017 at 5:33 AM, Shailesh Tripathi <
>>>> shailesh.tripathi.ec...@itbhu.ac.in> wrote:
>>>>
>>>>> Hello,
>>>>> Please review my proposal for project "OpenCL GPGPU raytracing" in
>>>>> GSOC,2017.
>>>>> In the proposal, the code work flow chart is left. I will complete it
>>>>> by today.
>>>>>
>>>>> Thanks in advance!
>>>>>
>>>>> Regards
>>>>> Shailesh
>>>>>
>>>>> Shailesh Tripathi
>>>>> B.Tech. Part-IV
>>>>> Electronics Engineering
>>>>> IIT-BHU (Varanasi)
>>>>>
>>>>> On Tue, Mar 21, 2017 at 9:36 AM, Christopher Sean Morrison <
>>>>> brl...@mac.com> wrote:
>>>>>
>>>>>>
>>>>>> > I am Shailesh Tripathi pursuing B.Tech. (Part-IV) in Electronics
>>>>>> Engineering from IIT-BHU, Varanasi, India. I am looking forward to apply
>>>>>> for GSOC,2017 for the project "OpenCL GPGPU raytracing”.
>>>>>>
>>>>>> Thanks for the introduction Shailesh.
>>>>>>
>>>>>> > Currently I am working on porting the primitive "hrt". Other than
>>>>>> selecting as much primitives as possible to port in the time frame, I am
>>>>>> looking forward to add a feature of automatic GPU device selection in
>>>>>> case
>>>>>> of multi-GPU system on the basis of a few parameters which would choose
>>>>>> the
>>>>>> most powerful GPU of all (the parameter selection needs more thinking,
>>>>>> probably selecting the one with maximum global memory would be helpful).
>>>>>> I
>>>>>> am looking forward for other optimizations too.
>>>>>>
>>>>>> Note that because we’re using OpenCL means that we’re not just
>>>>>> limited to selecting or relying on GPU devices. We want to be able to
>>>>>> take
>>>>>> advantages of all available computational resources.
>>>>>>
>>>>>> > I have some experience in GPGPU computing, CUDA and OpenCL, and
>>>>>> OpenGL. I am trying to learn CAD and ray-tracing techniques.
>>>>>>
>>>>>> Great! Look forward to seeing a code modification and details of
>>>>>> your proposal.
>>>>>>
>>>>>> Cheers!
>>>>>> Sean
>>>>>>
>>>>>>
>>>>>
>>>>
>>>> ------------------------------------------------------------
>>>> ------------------
>>>> Check out the vibrant tech community on one of the world's most
>>>> engaging tech sites, Slashdot.org! http://sdm.link/slashdot
>>>> _______________________________________________
>>>> BRL-CAD Developer mailing list
>>>> brlcad-devel@lists.sourceforge.net
>>>> https://lists.sourceforge.net/lists/listinfo/brlcad-devel
>>>>
>>>>
>>>
>>>
>>> --
>>> Vasco Alexandre da Silva Costa
>>> PhD in Computer Engineering (Computer Graphics)
>>> Instituto Superior Técnico/University of Lisbon, Portugal
>>>
>>> ------------------------------------------------------------
>>> ------------------
>>> Check out the vibrant tech community on one of the world's most
>>> engaging tech sites, Slashdot.org! http://sdm.link/slashdot
>>> _______________________________________________
>>> BRL-CAD Developer mailing list
>>> brlcad-devel@lists.sourceforge.net
>>> https://lists.sourceforge.net/lists/listinfo/brlcad-devel
>>>
>>>
>>
>
------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
BRL-CAD Developer mailing list
brlcad-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/brlcad-devel
