chore: cleanups, typos, simplifying

CHANGELOG.md

@@ -110,7 +110,7 @@ By @bradwerth [#6216](https://github.com/gfx-rs/wgpu/pull/6216).
 
 #### General
 
-- Added an example that splits up large data across multiple buffers and accesses them as a single contiguous array in the shader, demonstrating the naga/wgsl only `binding_array`, and some of the, less documented `wgpu::Features`. By @alphastrata in [#6130](https://github.com/gfx-rs/wgpu/pull/6138)
+- Added an example that splits up large data across multiple buffers and accesses them as a single contiguous array in the shader, demonstrating the naga/wgsl only `binding_array`, and some of the, less documented `wgpu::Features`. By @alphastrata in [#6138](https://github.com/gfx-rs/wgpu/pull/6138)
 
 - If GL context creation fails retry with GLES. By @Rapdorian in [#5996](https://github.com/gfx-rs/wgpu/pull/5996)
 - Fix profiling with `tracy`. By @waywardmonkeys in [#5988](https://github.com/gfx-rs/wgpu/pull/5988)

examples/src/big_compute_buffers/README.md

@@ -1,11 +1,8 @@
-# WIP
-
-
 # big-compute-buffers
 
 This example assumes you're familiar with the other GP-GPU compute examples in this repository, if you're not you should go look at those first.
 
-Showcases how to split larger datasets (things too big to fit into a single buffer), across multiple buffers whilst treating them as a single, contiguous buffer on the GPU.
+Demonstrates how to split larger datasets (things too big to fit into a single buffer), across multiple buffers whilst treating them as a single, contiguous buffer on the GPU.
 
 - Creates a large set of buffers totalling `1GB`, full of `0.0f32`.
 - Moves those buffers to the DEVICE.
@@ -19,6 +16,10 @@ const MAX_BUFFER_SIZE: u64 = 1 << 27; // 134_217_728 // 134MB
 const MAX_DISPATCH_SIZE: u32 = (1 << 16) - 1; // 65_535
 ```
 
+Note:
+- Large buffers can fail to allocate due to fragmentation issues.
+- Working with large buffers is not always ideal, you should also see the pagination example to see if that approach suits your needs better.
+
 ## To Run
 ```sh
 RUST_LOG=wgpu_examples::big_compute_buffers=info cargo run -r --bin wgpu-examples -- big_compute_buffers

examples/src/big_compute_buffers/mod.rs

@@ -1,6 +1,7 @@
 use std::{borrow::Cow, num::NonZeroU32, sync::Arc};
 use wgpu::{util::DeviceExt, Features};
 
+// These are set by the minimum required defaults for webgpu.
 const MAX_BUFFER_SIZE: u64 = 1 << 27; // 134_217_728 // 134MB
 const MAX_DISPATCH_SIZE: u32 = (1 << 16) - 1;
 
@@ -15,10 +16,11 @@ pub async fn execute_gpu(numbers: &[f32]) -> Option<Vec<f32>> {
         .request_device(
             &wgpu::DeviceDescriptor {
                 label: None,
-                required_features: Features::STORAGE_RESOURCE_BINDING_ARRAY |
-                        // These features are required to use `binding_array` in your wgsl.
-                    Features::BUFFER_BINDING_ARRAY |
-                    Features::UNIFORM_BUFFER_AND_STORAGE_TEXTURE_ARRAY_NON_UNIFORM_INDEXING,
+                // These features are required to use `binding_array` in your wgsl.
+                // Without them your shader may fail to compile.
+                required_features: Features::STORAGE_RESOURCE_BINDING_ARRAY
+                    | Features::BUFFER_BINDING_ARRAY
+                    | Features::UNIFORM_BUFFER_AND_STORAGE_TEXTURE_ARRAY_NON_UNIFORM_INDEXING,
 
                 memory_hints: wgpu::MemoryHints::Performance,
                 required_limits: wgpu::Limits {

examples/src/big_compute_buffers/tests.rs

@@ -2,7 +2,7 @@ use super::*;
 use wgpu_test::{gpu_test, GpuTestConfiguration, TestParameters};
 
 #[gpu_test]
-static COMPUTE_1: GpuTestConfiguration = GpuTestConfiguration::new()
+static TWO_BUFFERS: GpuTestConfiguration = GpuTestConfiguration::new()
     .parameters(
         TestParameters::default()
             .features(