Chip Davis

I want to get to 0% failure by the time Mesa releases their driver. I think if the three of us make a concerted push, we can make that happen.

> [@billhollings](https://github.com/billhollings) with texture cube atomics support on Apple10 family, was going to open issue requesting toggling on support in the portability extension for this family but seems no bit...

> * Is there a big reason why we bind buffers from largest index in Metal to 0? If there's no actual reason, would it be good to bind them...

It looks like there's an object lifetime issue here, where an object has been destroyed, but Metal still has an outstanding command buffer out for it, so when the completion...

The problem is that we don't support `VkPipelineShaderStageRequiredSubgroupSizeCreateInfo`. _At all._ This change will indeed silence the validation failure, but any extension struct you pass will continue to be ignored. The...

I was actually disappointed that I couldn't implement actual subgroup size control--especially after I learned that Direct3D 12 has it now as well. I begged Apple to support it, but...

The comment there says it's important for input attachments. Maybe we should gate this only on the `VK_IMAGE_USAGE_INPUT_ATTACHMENT_BIT`.

The MSL that was generated, which you provided in #1496, looks correct. This suggests that this is a bug in the Intel Metal driver. Intel drivers have been known to...

So much for that theory. I wonder if the driver is choking on the `float3` type. Does using a `vec4` for the color help?

> it seems that the huge amount of memory consumed comes from the intermediary buffers used for the tessellation shader output. Those buffers should not actually be a problem. We...