| phyar |
2007-12-04 10:03:44 |
32 |
|
I'm very interested in SPE due to it's thread supporting nature out of the box. I'm curious whether it supports batch shape casts and ray casts though. In Havok at work, we often create a big queue of raycasts or shapecasts and let it resolve on a seperate thread. This is used a ton for AI especially. Does SPE have similar functionality? In addition, batched queries are significantly cheaper, as they only do 1 broad phase for the entire batch(batches are most often used for close proximity tests). I'd be very interested in SPE if batched parallel queries were available. Is it planned?
------------------------------------------
Currently SPE use single thread to do ray cast, but these works are natural parallel task and easy to make it multi-threads. I'll add this functionality in the next release, remember your feedback or request is often a part of my plan. And could you tell me more about shape cast? |
|
| Jeremy |
2007-12-03 23:05:08 |
31 |
|
I'm very interested in SPE due to it's thread supporting nature out of the box. I'm curious whether it supports batch shape casts and ray casts though. In Havok at work, we often create a big queue of raycasts or shapecasts and let it resolve on a seperate thread. This is used a ton for AI especially. Does SPE have similar functionality? In addition, batched queries are significantly cheaper, as they only do 1 broad phase for the entire batch(batches are most often used for close proximity tests). I'd be very interested in SPE if batched parallel queries were available. Is it planned? |
|
| Steven |
2007-11-21 19:30:36 |
30 |
|
Hellow, phyar.
I guess, 'easy of use' is a first key-feature of SPE. And it's really easy to learn and use!
What are you think about 'destroyable physical environment' as a second key-feature of your engine? I have some experience and interest in this area.
Mail me, if you are interested:
dev <at> lss-dev.ru |
|
| Matt |
2007-11-14 21:52:07 |
29 |
|
Thanks, it's now working. I really appreciate the physics engine and tutorial; helps amateurs like me get a grasp on this stuff. |
|
| phyar |
2007-11-14 15:23:24 |
28 |
|
I'm using Dev-C/C++ from Bloodshed.
I only tried to compile the include statement for the SPE.h to see if it worked.
-------------------------------------------
OK, to avoid warning, you can simplely press \"Enter\" several times at the end of the each file, to fix the other problem, you can delete two lines as follow:
//if(size==0)
// error(0);
Thanks. |
|
| Matt |
2007-11-14 12:18:44 |
27 |
|
I'm using Dev-C/C++ from Bloodshed.
I only tried to compile the include statement for the SPE.h to see if it worked. |
|
| phyar |
2007-11-14 11:28:58 |
26 |
|
When I try to include the \"SPE.h\" header, I receive about 10 errors, all except 1 are \"[Warning] no newline at end of file\" regarding the other header files.
The other one is SPE_SDK_2.00[1]\\\\include\\\\SPEArray.h In member function `T& SPEArray<T>::pop()':
Help would be greatly appreciated. Thank you.
-----------------------------------
What IDE are you using? Can you compile the tutorials in the SDK successfully? I afraid SPE does not support VS 6.0. |
|
| Matt |
2007-11-14 02:04:29 |
25 |
|
When I try to include the \"SPE.h\" header, I receive about 10 errors, all except 1 are \"[Warning] no newline at end of file\" regarding the other header files.
The other one is SPE_SDK_2.00[1]\\include\\SPEArray.h In member function `T& SPEArray<T>::pop()':
Help would be greatly appreciated. Thank you. |
|
| phyar |
2007-11-12 13:40:03 |
24 |
|
It'd be really handy if we could query an object's momentum at a point, like the inverse of RigidBody::AddWorldImpulse.
Something like:
SPEVector RigidBody::GetWorldImpulse(SPEVector &pos), or ::GetTensor(void *pmat).
This could be used for calculating collision damage, or buoyancy forces against a boat, or friction against a ray-cast car wheel. Would this be possible in the next release?
-------------------------------------------------
OK, your suggestion is valuable and I'll add some interfaces related to these use in the next release.
Thanks. |
|
| Alex McLeod |
2007-11-12 12:47:45 |
23 |
|
It'd be really handy if we could query an object's momentum at a point, like the inverse of RigidBody::AddWorldImpulse.
Something like:
SPEVector RigidBody::GetWorldImpulse(SPEVector &pos), or ::GetTensor(void *pmat).
This could be used for calculating collision damage, or buoyancy forces against a boat, or friction against a ray-cast car wheel. Would this be possible in the next release? |
|
|
