From fcac02e405661de311b5ceebbd6d3e2c78bf8aea Mon Sep 17 00:00:00 2001
From: jsmall-nvidia <jsmall@nvidia.com>
Date: Fri, 21 Aug 2020 16:04:42 -0400
Subject: Vulkan update/NVAPI support (#1511)

* First pass at incorporating nvapi into test harness.

* D3d12 Atomic Float Add via NVAPI working

* Dx12 atomic float appears to work.

* Atomic float add on Dx12.

* Added atomic64 feature addition to vk.
Fix correct output for atomic-float-byte-address.slang

* Disable atomic float failing tests.

* Upgraded VK headers.

* Detect atomic float availability on VK.

* Try to get test working for in64 atomic.

* Made HLSL prelude controlled via the render-test requirements.

* Added -enable-nvapi to premake.

* Fix D3D12Renderer when NVAPI is not available.

* Small improvements to VKRenderer.

* Improve atomic documentation in target-compatibility.md.
---
 .../atomic-float-byte-address-buffer.slang         | 48 ++++++++++++++--------
 1 file changed, 32 insertions(+), 16 deletions(-)

(limited to 'tests/slang-extension/atomic-float-byte-address-buffer.slang')

diff --git a/tests/slang-extension/atomic-float-byte-address-buffer.slang b/tests/slang-extension/atomic-float-byte-address-buffer.slang
index d89ee8bd1..a332ddeac 100644
--- a/tests/slang-extension/atomic-float-byte-address-buffer.slang
+++ b/tests/slang-extension/atomic-float-byte-address-buffer.slang
@@ -1,18 +1,28 @@
-// byte-address-buffer.slang
+// atomic-float-byte-address-buffer.slang
 
-//DISABLE_TEST(compute):COMPARE_COMPUTE_EX:-cpu -compute
-//DISABLE_TEST(compute):COMPARE_COMPUTE_EX:-slang -compute
+//DISABLE_TEST(compute):COMPARE_COMPUTE_EX:-cpu -compute -output-using-type
+// Produces a different result
+//TEST(compute):COMPARE_COMPUTE_EX:-dx11 -slang -compute -render-features atomic-float -output-using-type -nvapi-register u0
 // Disabled because crashes currently on vulkan 
-//DISABLE_TEST(compute, vulkan):COMPARE_COMPUTE_EX:-vk -compute
-// Disabled because requires nvapi to work
-// Note for this feature we require dxc and we can force that with -use-dxil
-//DISABLE_TEST(compute):COMPARE_COMPUTE_EX:-d3d12 -compute -use-dxil
-//TEST(compute):COMPARE_COMPUTE_EX:-cuda -compute
+// https://vulkan.gpuinfo.org/listdevicescoverage.php?extension=VK_EXT_shader_atomic_float
+//TEST(compute, vulkan):COMPARE_COMPUTE_EX:-vk -compute -render-features atomic-float -output-using-type
+// Doesn't work on D3D12 for unknown reasons
+//DISABLE_TEST(compute):COMPARE_COMPUTE_EX:-d3d12 -compute -render-features atomic-float -output-using-type -compile-arg -O2 -nvapi-register u0
+//DISABLE_TEST(compute):COMPARE_COMPUTE_EX:-d3d12 -compute -use-dxil -render-features atomic-float -output-using-type -compile-arg -O2 -nvapi-register u0
+//TEST(compute):COMPARE_COMPUTE_EX:-cuda -compute -output-using-type 
 
-//TEST_INPUT:ubuffer(data=[0.1 0.2 0.3 0.4]):out,name=outputBuffer
-RWByteAddressBuffer outputBuffer;
+// We don't use this, but having this defined makes the 0 slot available if nvapi is going to be used
+// Only strictly necessary on the dx12 path
+//TEST_INPUT:ubuffer(data=[0 0 0 0 ], stride=4):name=nvapiBuffer
+RWStructuredBuffer<int> nvapiBuffer;
 
-//TEST_INPUT:ubuffer(data=[0.7 0.5 0.2 0.6]):name=anotherBuffer
+//TEST_INPUT:ubuffer(data=[1.0 2.0 3.0 4.0], stride=4):out,name=outputBuffer
+RWStructuredBuffer<float> outputBuffer;
+
+//TEST_INPUT:ubuffer(data=[1.0 2.0 3.0 4.0], stride=4):name=workBuffer
+RWByteAddressBuffer workBuffer;
+
+//TEST_INPUT:ubuffer(data=[0.7 0.5 0.2 0.6], stride=4):name=anotherBuffer
 RWStructuredBuffer<float> anotherBuffer;
 
 [numthreads(16, 1, 1)]
@@ -22,13 +32,19 @@ void computeMain(uint3 dispatchThreadID : SV_DispatchThreadID)
     int idx = (tid & 3) ^ (tid >> 2); 
 
     //const float delta = anotherBuffer[idx & 3];
-    const float delta = 0.0f;
-
+    
     float previousValue = 0;
-    outputBuffer.InterlockedAddFp32((idx << 2), 1.0f, previousValue);
-    //outputBuffer.InterlockedAddFp32((idx ^ 2) << 2, 2.0f + delta);
+    workBuffer.InterlockedAddFp32((idx << 2), 1.0f, previousValue);
+    //workBuffer.InterlockedAddFp32((idx ^ 2) << 2, 2.0f + delta);
     
     // The sum of values in anotherBuffer should also be added
     //int anotherIdx = tid >> 2;
-    //outputBuffer.InterlockedAddFp32(anotherIdx << 2, delta);
+    //workBuffer.InterlockedAddFp32(anotherIdx << 2, delta);
+    
+    GroupMemoryBarrierWithGroupSync();
+    
+    if (tid < 4)
+    {
+        outputBuffer[tid] = asfloat(workBuffer.Load(tid << 2));
+    }
 }
\ No newline at end of file
-- 
cgit v1.2.3