VL.Fuse - Public Room Timeline

	VL.Fuse	208 Members
		3 Servers

Load older messages

Sender	Message	Time
6 Aug 2024
	bj-rn changed their profile picture.	09:42:20
	prt changed their display name from philipp_prt to prt.	14:10:33
bj-rn	Hey texone just watched the second Fuse workshop on noises. Nice stuff. Was wondering if you could somehow share the noise overview miro board as pdf (maybe add it as resource to the Fuse nuget)?	15:00:36
7 Aug 2024
Raphael Dupont	Download pleaideTofuse.PNG	18:08:17
Raphael Dupont	texone: texone Hello Everyone, its my foot here! I followed all the fuse workshops, Thanks again texone :) During your workshop I've noticed the pleiades (Augmenta) logo on your desktop so I Think you will probably my answer with you. So I'm using the last tracking system by augmenta and I received the point cloud of each objet via websocket .... I want to used those cluster of point as source of emmision in Fuse ... But I dont Know To input A spread of spread into a bufferIn of fuse... How do I keep the ID of each cluster and not only each point...	18:08:18
Raphael Dupont	Tnaks you :)	18:08:33
8 Aug 2024
tim_ru	Download BufferGroups.vl	13:02:21
tim_ru	You can do it like this. Use a `Flatten` node to get a buffer of Vector3 and then you can use the count of the individual groups to determine which group you are addressing	13:03:57
tim_ru	you could also add a custom attribute "GroupID" to the particle system and write the result of the `<` as an int to it, then you can use that for all lerps or switches you use later on	13:09:19
Maxime	Hello ! I am digging into Fuse now, and have a question regarding the For loop. I have a calculation which is currently using a ForEach loop and obviously it is not really optimized. So, I would like to use a For (Fuse) loop instead, but I don't understand how to get the same results because of the lack of splicer output.	14:54:10
Maxime	Download image.png	14:54:21
Maxime	Here is what I am doing with the ForEach loop	14:54:34
Maxime	Download image.png	14:54:48
Maxime	And this is my current attempt	14:54:53
Maxime	Because there is no splicer output, the AppendBuffer and AppendItem are accumulated to just one ComputeStage. With the ForEach loop, I have a ComputeStage per Buffer. I don't see how it change that. Also, because I can't output the result of AppendBuffer to a Spread, I was thinking of using another Buffer to store the values. But it seems that I can't use Index witch SetItem. I don't understand that.	14:59:05
Maxime	* Because there is no splicer output, the `AppendBuffer` and `AppendItem` are accumulated to just one `ComputeStage`. With the ForEach loop, I have a `ComputeStage` per Buffer. I don't see how it change that. Also, because I can't output the result of `AppendBuffer` to a Spread, I was thinking of using another Buffer to store the values. But it seems that I can't use `Index` witch `SetItem`. I don't understand that.	14:59:53
Maxime	Any clue?	15:00:45
Maxime	* Because there is no splicer output, the `AppendBuffer` and `AppendItem` are accumulated to just one `ComputeStage`. With the ForEach loop, I have a `ComputeStage` per Buffer. I don't see how to change that. Also, because I can't output the result of `AppendBuffer` to a Spread, I was thinking of using another Buffer to store the values. But it seems that I can't use `Index` witch `SetItem`. I don't understand that.	15:03:48
Maxime	* Because there is no splicer output, the `AppendBuffer` and `AppendItem` are accumulated to just one `ComputeStage`. With the ForEach loop, I have a `ComputeStage` per Buffer. I don't see how to change that. Also, because I can't output the result of `AppendBuffer` to a Spread, I was thinking of using another Buffer to store the values. But it seems that I can't use `Index` with `SetItem`. I don't understand that.	15:04:11
tebjan	I think you should unpack the loop one level. A compute shader itself is already a for each loop. Check out the intro of my shader workshop to get a general understanding of the concepts involved in a compute shader...	15:07:00
tebjan	As for the splicer, you would need 8 buffers, same as in your first screenshot	15:11:23
tebjan	As you are using append, you need to separate the state if each of the append buffers	15:12:11
tebjan	You can't use one append buffer for all	15:12:40
tebjan	* As you are using append, you need to separate the state of each of the append buffers	15:13:03
Maxime	In reply to @tebjan:matrix.org I think you should unpack the loop one level. A compute shader itself is already a for each loop. Check out the intro of my shader workshop to get a general understanding of the concepts involved in a compute shader... Thank, I will check it. Just saw the one available on YouTube on vvvvTv, but not the one on TNI	15:19:00
Maxime	In reply to @tebjan:matrix.org As for the splicer, you would need 8 buffers, same as in your first screenshot So I can't use one buffer with 8 entries? I will need 8 buffers with just 1 entry?	15:19:32
tebjan	In reply to @maximesouvestre:matrix.org Because there is no splicer output, the `AppendBuffer` and `AppendItem` are accumulated to just one `ComputeStage`. With the ForEach loop, I have a `ComputeStage` per Buffer. I don't see how to change that. Also, because I can't output the result of `AppendBuffer` to a Spread, I was thinking of using another Buffer to store the values. But it seems that I can't use `Index` with `SetItem`. I don't understand that. Also check out the compute workshop from @texone:matrix.org: there is an example of atomic increments. You can getItem and Set item of 8 values in a compute buffer. Instead of using append	15:20:56
Maxime	In reply to @tebjan:matrix.org You can't use one append buffer for all Yes I would like to separate each AppendBuffer like I did in the first screenshot. I just couldn't find a way to do so for now, that's one of the reason why I am blocked	15:22:03
tebjan	I think it's called InterlockedAdd but it can slow down the shader quite a bit	15:22:09
tebjan	As all threads need to be synced	15:22:32

Show newer messages

Back to Room ListRoom Version: 6