Hello, running the compute example: View full code <d

Update: removing the <div class="highlight highlight-source-rust notranslate posit

Ok, here's the problem: <a class="commit-link" href="https://github.

<a class="user-mention notranslate" data-hovercard-type="user" data-hovercard-url="/us

Example panics at runtime (`COPY_DST` flag) about emu HOT 4 OPEN

calebwin commented on May 26, 2024

Example panics at runtime (`COPY_DST` flag)

from emu.

Comments (4)

wbrickner commented on May 26, 2024

Weird! I see that you are setting the copy destination flag correctly. Following the execution from .as_device_boxed_mut() above:

emu/emu_core/src/device.rs

Lines 260 to 302 in 9fe3db3

 fn create_from_as<T, B: Borrow<T>>( 

 &mut self, 

 host_obj: B, 

 mutability: Mutability, 

 ) -> DeviceBox<T> 

 where 

 T: AsBytes + ?Sized, 

 { 

 // serialize the data into bytes 

 // these bytes can later be deserialized back into T 

 let host_obj_bytes = host_obj.borrow().as_bytes(); 

 // create a staging buffer with host_obj copied over 

 let staging_buffer = self.device.create_buffer(&wgpu::BufferDescriptor { 

 label: None, 

 size: host_obj_bytes.len() as u64, 

 usage: wgpu::BufferUsage::MAP_READ | wgpu::BufferUsage::COPY_DST, 

 mapped_at_creation: false, 

 }); 

 // then create an initialized storage buffer of appropriate size 

 let storage_buffer = self 

 .device 

 .create_buffer_init(&wgpu::util::BufferInitDescriptor { 

 label: None, 

 usage: match mutability { 

 Mutability::Mut => wgpu::BufferUsage::STORAGE, 

 Mutability::Const => wgpu::BufferUsage::STORAGE, 

 } | wgpu::BufferUsage::COPY_SRC 

 | wgpu::BufferUsage::COPY_DST, 

 contents: host_obj_bytes, 

 }); 

 // return the final DeviceBox 

 // note that we keep both the storage buffer and the staging buffer 

 // we will re-use the staging buffer for reads (but not for writes, for writes we just create a new staging buffer) 

 DeviceBox { 

 staging_buffer, 

 storage_buffer, 

 size: host_obj_bytes.len() as u64, 

 phantom: PhantomData, 

 mutability: Some(mutability), 

 } 

 }

Any idea what the issue is?

from emu.

wbrickner commented on May 26, 2024

Update: removing the

shapes.set(vec![
  Shape {
    x: 0,
    y: 0,
    w: 100,
    h: 100,
    r: [2, 9]
  };
  1024
])?;

eliminates the error, and things always complete successfully.
Alternatively, leaving the .set call and removing the final .get call also prevents the panic.

Are usage flags being corrupted by .set?

from emu.

wbrickner commented on May 26, 2024

Ok, here's the problem:

80a28e7#diff-c42b8503ace5f383ffbc7c9f91463d31dc54fc36b2d3974c49a3e2fc434024e9R331-R337

when setting, the existing staging buffer is discarded and replaced with a new staging buffer, which is created with the usage flag COPY_SRC.

Adding to the end of the function the construction of a new staging buffer with MAP_READ | COPY_DST (as it usually has, from the DeviceBox construction pathway) solves the issue, when it comes time to read the correct usage is there.

I would like to not do it this way, it seems quite wasteful to prepare a new allocation and immediately discard it etc. Mixing
COPY_SRC and COPY_DST and MAP_READ permissions is not allowed evidently. The usage flag paradigm is the problem here, unsure how fundamentally important it really is. Perhaps two staging buffers could be lazily prepared, one for copies to/from the GPU device.

from emu.

calebwin commented on May 26, 2024

@wbrickner Thanks for doing this investigation in this issue. I see how the staging buffer creation is problematic...

If anyone has a PR that fixes this, I can review/edit/merge.

from emu.

Example panics at runtime (`COPY_DST` flag) about emu HOT 4 OPEN

Comments (4)

Related Issues (20)

Recommend Projects

React

Vue.js

Typescript

TensorFlow

Django

Laravel

D3

Recommend Topics

javascript

web

server

Machine learning

Visualization

Game

Recommend Org

Facebook

Microsoft

Google

Alibaba

D3

Tencent

	fn create_from_as<T, B: Borrow<T>>(
	&mut self,
	host_obj: B,
	mutability: Mutability,
	) -> DeviceBox<T>
	where
	T: AsBytes + ?Sized,
	{
	// serialize the data into bytes
	// these bytes can later be deserialized back into T
	let host_obj_bytes = host_obj.borrow().as_bytes();

	// create a staging buffer with host_obj copied over
	let staging_buffer = self.device.create_buffer(&wgpu::BufferDescriptor {
	label: None,
	size: host_obj_bytes.len() as u64,
	usage: wgpu::BufferUsage::MAP_READ \| wgpu::BufferUsage::COPY_DST,
	mapped_at_creation: false,
	});
	// then create an initialized storage buffer of appropriate size
	let storage_buffer = self
	.device
	.create_buffer_init(&wgpu::util::BufferInitDescriptor {
	label: None,
	usage: match mutability {
	Mutability::Mut => wgpu::BufferUsage::STORAGE,
	Mutability::Const => wgpu::BufferUsage::STORAGE,
	} \| wgpu::BufferUsage::COPY_SRC
	\| wgpu::BufferUsage::COPY_DST,
	contents: host_obj_bytes,
	});

	// return the final DeviceBox
	// note that we keep both the storage buffer and the staging buffer
	// we will re-use the staging buffer for reads (but not for writes, for writes we just create a new staging buffer)
	DeviceBox {
	staging_buffer,
	storage_buffer,
	size: host_obj_bytes.len() as u64,
	phantom: PhantomData,
	mutability: Some(mutability),
	}
	}