I have experience with D3D11 and want to learn D3D12. I am reading the official D3D12 multithread example and don't understand why the shadow map (generated in the first pass as a DSV, consumed in the second pass as SRV) is created for each frame (actually only 2 copies, as the FrameResource is reused every 2 frames).
The code that creates the shadow map resource is here, in the FrameResource class, instances of which is created here.
There is actually another resource that is created for each frame, the constant buffer. I kind of understand the constant buffer. Because it is written by CPU (D3D11 dynamic usage) and need to remain unchanged until the GPU finish using it, so there need to be 2 copies. However, I don't understand why the shadow map needs to do the same, because it is only modified by GPU (D3D11 default usage), and there are fence commands to separate reading and writing to that texture anyway. As long as the GPU follows the fence, a single texture should be enough for the GPU to work correctly. Where am I wrong?
Thanks in advance.
EDIT
According to the comment below, the "fence" I mentioned above should more accurately be called "resource barrier".
The key issue is that you don't want to stall the GPU for best performance. Double-buffering is a minimal requirement, but typically triple-buffering is better for smoothing out frame-to-frame rendering spikes, etc.
FWIW, the default behavior of DXGI Present is to stall only after you have submitted THREE frames of work, not two.
Of course, there's a trade-off between triple-buffering and input responsiveness, but if you are maintaining 60 Hz or better than it's likely not noticeable.
With all that said, typically you don't need to double-buffered depth/stencil buffers for rendering, although if you wanted to make the initial write of the depth-buffer overlap with the read of the previous depth-buffer passes then you would want distinct buffers per frame for performance and correctness.
The 'writes' are all complete before the 'reads' in DX12 because of the injection of the 'Resource Barrier' into the command-list:
void FrameResource::SwapBarriers()
{
// Transition the shadow map from writeable to readable.
m_commandLists[CommandListMid]->ResourceBarrier(1, &CD3DX12_RESOURCE_BARRIER::Transition(m_shadowTexture.Get(), D3D12_RESOURCE_STATE_DEPTH_WRITE, D3D12_RESOURCE_STATE_PIXEL_SHADER_RESOURCE));
}
void FrameResource::Finish()
{
m_commandLists[CommandListPost]->ResourceBarrier(1, &CD3DX12_RESOURCE_BARRIER::Transition(m_shadowTexture.Get(), D3D12_RESOURCE_STATE_PIXEL_SHADER_RESOURCE, D3D12_RESOURCE_STATE_DEPTH_WRITE));
}
Note that this sample is a port/rewrite of the older legacy DirectX SDK sample MultithreadedRendering11, so it may be just an artifact of convenience to have two shadow buffers instead of just one.
Related
I'm having an issue with atomics in wgpu / WGSL but I'm not sure if it's due to a fundamental misunderstanding or a bug in my code.
I have a input array declared in WGSL as
struct FourTileUpdate {
// (u32 = 4 bytes)
data: array<u32, 9>
};
#group(0) #binding(0) var<storage, read> tile_updates : array<FourTileUpdate>;
I'm limiting the size of this array to around 5MB, but sometimes I need to transfer more than that for a single frame and so use multiple command encoders & compute passes.
Each "tile update" has an associated position (x & y) and a ms_since_epoch property that indicates when the tile update was created. Tile updates get written to a texture.
I don't want to overwrite newer tile updates with older tile updates, so in my shader I have a guard:
storageBarrier();
let previous_timestamp_value = atomicMax(&last_timestamp_for_tile[x + y * r_locals.width], ms_since_epoch);
if (previous_timestamp_value > ms_since_epoch) {
return;
}
However, something is going wrong and older tile updates are overwriting newer tile updates. I can't reproduce this on Windows / Vulkan but it consistently happens on macOS / Metal. Here's an image of the rendered texture--it should be completely green instead of the occasional red and black pixel:
rendered texture
A few questions:
is execution order guaranteed to be the same as the order of the command encoder constructions?
do storageBarrier() and atomics work across all invocations in a single frame or just the compute pass?
I tried submitting each encoder with queue.submit(Some(encoder.finish())) before creating the next encoder for the frame, and even waiting for the queue to finish processing for each submitted encoder with
let (tx, rx) = mpsc::channel();
queue.on_submitted_work_done(move || {
tx.send().unwrap();
});
device.poll(wgpu::Maintain::Wait);
rx.rev().unwrap()
// ... loop back and create & submit next encoder for current frame
but that didn't work either.
Good questions!
is execution order guaranteed to be the same as the order of the command encoder constructions?
I believe that is the case. But I checked and the spec is actually unclear about this. I filed https://github.com/gpuweb/gpuweb/issues/3809 to fix this.
Further, I believe the intent is that all memory accesses (e.g. to storage buffers) from one GPU command will complete before the next GPU command begins. So the effect of any writes in one command will be visible in the next command (read-after-write hazard). Also, a write in a later command will not be visible in an earlier command (write-after-read hazard).
do storageBarrier() and atomics work across all invocations in a single frame or just the compute pass?
Another good question. storageBarrier() only works within a single workgroup. This may be surprising, but is due to a limitation in some platforms.
For further details, see https://github.com/gpuweb/gpuweb/issues/3774
This will be a FAQ because it is surprising, and subtle!
Update: I suspect the bad behaviour you're seeing is that storageBarrier() does not work across workgroups. It's a limitation in Metal.
I have a QQuickImageProvider,
The frequency of the requestPixmap is not always stable. Sometimes the delta between 2 calls exceed 20 ms.
And a visual dropping effect can be observed on the screen.
Someone have an idea ? It's the good way to do that ?
Can I monitor or debug this ?
Thanks
If you want to render frames at a high frequency, consider using another approach: create a custom QQuickItem and reimplement the updatePaintNode method: https://doc.qt.io/qt-6/qquickitem.html#updatePaintNode. As an alternative, you can also use a QQuickPaintedItem, but performance is slower: https://doc.qt.io/qt-6/qquickpainteditem.html.
In any case, note that it takes time to decode images (you don't say what is the source format) and upload to the GPU (you did not say the size). On some embedded systems, 20ms may be challenging.
As my question title says, I want update texture for every frame.
I got an idea :
create a VkImage as a texture buffer with bellow configurations :
initialLayout = VK_IMAGE_LAYOUT_PREINITIALIZED
usage= VK_IMAGE_USAGE_SAMPLED_BIT
and it's memory type is VK_MEMORY_PROPERTY_HOST_VISIBLE_BIT | VK_MEMORY_PROPERTY_HOST_COHERENT_BIT
In draw loop :
first frame :
map texure data to VkImage(use vkMapMemory).
change VkImage layout from VK_IMAGE_LAYOUT_PREINITIALIZED to VK_IMAGE_LAYOUT_SHADER_READ_ONLY_OPTIMAL.
use this VkImage as texture buffer.
second frame:
The layout will be VK_IMAGE_LAYOUT_SHADER_READ_ONLY_OPTIMAL after the first frame , can I map next texure data to this VkImage directly without change it's layout ? if I can not do that, which layout can I change this VkImage to ?
In vkspec 11.4 it says :
The new layout used in a
transition must not be VK_IMAGE_LAYOUT_UNDEFINED or VK_IMAGE_LAYOUT_PREINITIALIZED
So , I can not change layout back to _PREINITIALIZED .
Any help will be appreciated.
For your case you do not need LAYOUT_PREINITIALIZED. That would only complicate your code (forcing you to provide separate code for the first frame).
LAYOUT_PREINITIALIZED is indeed a very special layout intended only for the start of the life of the Image. It is more useful for static textures.
Start with LAYOUT_UNDEFINED and use LAYOUT_GENERAL when you need to write the Image from CPU side.
I propose this scheme:
berfore render loop
Create your VkImage with UNDEFINED
1st to Nth frame (aka render loop)
Transition image to GENERAL
Synchronize (likely with VkFence)
Map the image, write it, unmap it (weell, the mapping and unmaping can perhaps be outside render loop)
Synchronize (potentially done implicitly)
Transition image to whatever layout you need next
Do your rendering and whatnot
start over at 1.
It is a naive implementation but should suffice for ordinary hobbyist uses.
Double buffered access can be implemented — that is e.g. VkBuffer for CPU access and VkImage of the same for GPU access. And VkCmdCopy* must be done for the data hand-off.
It is not that much more complicated than the above approach and there can be some performance benefits (if you need those at your stage of your project). You usually want your resources in device local memory, which often is not also host visible.
It would go something like:
berfore render loop
Create your VkBuffer b with UNDEFINED backed by HOST_VISIBLE memory and map it
Create your VkImage i with UNDEFINED backed by DEVICE_LOCAL memory
Prepare your synchronization primitives between i and b: E.g. two Semaphores, or Events could be used or Barriers if the transfer is in the same queue
1st to Nth frame (aka render loop)
Operations on b and i can be pretty detached (even can be on different queues) so:
For b:
Transition b to GENERAL
Synchronize to CPU (likely waiting on VkFence or vkQueueIdle)
invalidate(if non-coherent), write it, flush(if non-coherent)
Synchronize to GPU (done implicitly if 3. before queue submission)
Transition b to TRANSFER
Synchronize to make sure i is not in use (likely waiting on a VkSemaphore)
Transition i to TRANSFER
Do vkCmdCopy* from b to i
Synchronize to make known I am finished with i (likely signalling a VkSemaphore)
start over at 1.
(The fence at 2. and semaphore at 6. have to be pre-signalled or skipped for first frame to work)
For i:
Synchronize to make sure i is free to use (likely waiting on a VkSemaphore)
Transition i to whatever needed
Do your rendering
Synchronize to make known I am finished with i (likely signalling a VkSemaphore)
start over at 1.
You have a number of problems here.
First:
create a VkImage as a texture buffer
There's no such thing. The equivalent of an OpenGL buffer texture is a Vulkan buffer view. This does not use a VkImage of any sort. VkBufferViews do not have an image layout.
Second, assuming that you are working with a VkImage of some sort, you have recognized the layout problem. You cannot modify the memory behind the texture unless the texture is in the GENERAL layout (among other things). So you have to force a transition to that, wait until the transition command has actually completed execution, then do your modifications.
Third, Vulkan is asynchronous in its execution, and unlike OpenGL, it will not hide this from you. The image in question may still be accessed by the shader when you want to change it. So usually, you need to double buffer these things.
On frame 1, you set the data for image 1, then render with it. On frame 2, you set the data for image 2, then render with it. On frame 3, you overwrite the data for image 1 (using events to ensure that the GPU has actually finished frame 1).
Alternatively, you can use double-buffering without possible CPU waiting, by using staging buffers. That is, instead of writing to images directly, you write to host-visible memory. Then you use a vkCmdCopyBufferToImage command to copy that data into the image. This way, the CPU doesn't have to wait on events or fences to make sure that the image is in the GENERAL layout before sending data.
And BTW, Vulkan is not OpenGL. Mapping of memory is always persistent; there's no reason to unmap a piece of memory if you're going to map it every frame.
As part of an app that allows auditors to create findings and associate photos to them (Saved as Base64 strings due to a limitation on the web service) I have to loop through all findings and their photos within an audit and set their sync value to true.
Whilst I perform this loop I see a memory spike jumping from around 40MB up to 500MB (for roughly 350 photos and 255 findings) and this number never goes down. On average our users are creating around 1000 findings and 500-700 photos before attempting to use this feature. I have attempted to use #autorelease pools to keep the memory down but it never seems to get released.
for (Finding * __autoreleasing f in self.audit.findings){
#autoreleasepool {
[f setToSync:#YES];
NSLog(#"%#", f.idFinding);
for (FindingPhoto * __autoreleasing p in f.photos){
#autoreleasepool {
[p setToSync:#YES];
p = nil;
}
}
f = nil;
}
}
The relationships and retain cycles look like this
Audit has a strong reference to Finding
Finding has a weak reference to Audit and a strong reference to FindingPhoto
FindingPhoto has a weak reference to Finding
What am I missing in terms of being able to effectively loop through these objects and set their properties without causing such a huge spike in memory. I'm assuming it's got something to do with so many Base64 strings being loaded into memory when looping through but never being released.
So, first, make sure you have a batch size set on the fetch request. Choose a relatively small number, but not too small because this isn't for UI processing. You want to batch a reasonable number of objects into memory to reduce loading overhead while keeping memory usage down. Try 50 or 100 and see how it goes, then consider upping the batch size a little.
If all of the objects you're loading are managed objects then the correct way to evict them during processing is to turn them into faults. That's done by calling refreshObject:mergeChanges: on the context. BUT - that discards any changes, and your loop is specifically there to make changes.
So, what you should really be doing is batch saving the objects you've modified and then turning those objects back into faults to remove the data from memory.
So, in your loop, keep a counter of how many you've modified and save the context each time you hit that count and refresh all the objects that were processed so far. The batch on the fetch and the batch size to save should be the same number.
There's probably a big difference in size between your "Finding" objects and the associated images. So your primary aim should be to redesign your database in a way so that unfaulting (loading) a Finding object does not automatically load the base64 encoded image.
That's actually one of the major strengths of Code Data: Loading part of an object hierarchy. Just try to move the base64 encoded data to an own (managed) object so that Core Data does not load it. It will still be loaded as needed when the reference is touched.
I'm developing an app (XNA Game) for the XBOX, which is a pretty simple app. The startpage contains tiles with moving gif images. Those gif images are actually all png images, which gets loaded once by every tile, and put in an array. Then, using a defined delay, these images are played (using a counter which increases every time a delay passes).
This all works well, however, I noticed some small lag every x seconds in the movement of the GIF images. I then started to add some benchmarking stuff:
http://gyazo.com/f5fe0da3ff81bd45c0c52d963feb91d8
As you can see, the FPS is pretty low for such a simple program (This is in debug, when running the app from the Xbox itself, I get an avg of 62fps).
2 important settings:
Graphics.SynchronizeWithVerticalRetrace = false;
IsFixedTimeStep = false;
Changing isFixedTimeStep to true increases the lag. The settings tile has wheels which rotate, and you can see the wheels go back a little every x seconds. The same counts for SynchronizeWVR, also increases lag.
I noticed a connection between the lag and the moment the garbage collector kicks in, every time it kicks in, there is a lag...
Don't mind the MAX HMU(Heap memory usage), as this is takes the amount of the start, the avg is more realistic.
Here is another screen from the performance monitor, however I don't understand much from this tool, first time I'm using it... Hope it helps:
http://gyazo.com/f70a3d400657ac61e6e9f2caaaf17587
After a little research I found the culprit.
I have custom components that all derive from GameComponent, and who get added to the Component list of the main Game class.
This was one (of a total of 2) major problem, causing to update everything that wasn't needing an update. (The draw method was the only one who kept the page state in mind, and only drew if needed).
I fixed this by using different "screens" (or pages as I called them), wich are the only components who derive from GameComponent.
Then I only update the page wich is active, and the custom components on that page also get updated. Problem fixed.
The second big problem, is the following;
I made a class which helps me on positioning stuff on the screen, relative that is, with percentages and stuff like that. Parent containers, aligns & v-aligns etc etc.
That class had properties, for size & vectors, but instead of saving the calculated value in a backing field, I recalculated them everytime I accessed a property. But calculating complex stuff like that uses references (to parent & child containers for example) wich made it very hard for the CLR, because it had alot of work to do.
I now rebuilt the whole positioning class to a fully functional optimized class, with different flags for recalculating when necessairy, and instead of drops of 20fps, I now get an average of 170+fps!