From 67ca54997d445e15891965b8d77561b9d10bb18c Mon Sep 17 00:00:00 2001
From: jsmall-nvidia <jsmall@nvidia.com>
Date: Mon, 24 Aug 2020 14:26:27 -0400
Subject: NVAPI improvements (#1512)

* 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.

* Fixed NVAPI working on D3D12.

* Test for specific NVAPI features.

* Remove requiredFeatures from Renderer::Desc as was ignored. Tried to document more around nvapiExtnSlot.

* Readded requiredFeatures to Renderer::Desc

* Improve comments in the tests.
---
 tools/gfx/d3d12/render-d3d12.cpp | 57 +++++++++++++++++++++++++++-------------
 1 file changed, 39 insertions(+), 18 deletions(-)

(limited to 'tools/gfx/d3d12/render-d3d12.cpp')

diff --git a/tools/gfx/d3d12/render-d3d12.cpp b/tools/gfx/d3d12/render-d3d12.cpp
index ca44aa04d..0f23d8dd4 100644
--- a/tools/gfx/d3d12/render-d3d12.cpp
+++ b/tools/gfx/d3d12/render-d3d12.cpp
@@ -1523,6 +1523,19 @@ Result D3D12Renderer::_createDevice(DeviceCheckFlags deviceCheckFlags, const Uno
     return SLANG_OK;
 }
 
+static bool _isSupportedNVAPIOp(ID3D12Device* dev, uint32_t op)
+{
+#ifdef GFX_NVAPI
+    {
+        bool isSupported;
+        NvAPI_Status status = NvAPI_D3D12_IsNvShaderExtnOpCodeSupported(dev, NvU32(op), &isSupported);
+        return status == NVAPI_OK && isSupported;
+    }
+#else
+    return false;
+#endif
+}
+
 Result D3D12Renderer::initialize(const Desc& desc, void* inWindowHandle)
 {
     m_hwnd = (HWND)inWindowHandle;
@@ -1602,32 +1615,40 @@ Result D3D12Renderer::initialize(const Desc& desc, void* inWindowHandle)
         return SLANG_FAIL;
     }
 
+    // Set the device
+    m_device = m_deviceInfo.m_device;
+
     // NVAPI
+    if (desc.nvapiExtnSlot >= 0)
     {
-        const char* features[] = { "nvapi", "atomic-float", "atomic-int64" };
-        bool needsNvapi = false;
-        for (Index i = 0; i < SLANG_COUNT_OF(features); ++i)
+        if (SLANG_FAILED(NVAPIUtil::initialize()))
         {
-            if (desc.requiredFeatures.indexOf(features[i]) >= 0)
-            {
-                needsNvapi = true;
-                break;
-            }
+            return SLANG_E_NOT_AVAILABLE;
         }
 
-        if (needsNvapi && SLANG_SUCCEEDED(NVAPIUtil::initialize()))
+#ifdef GFX_NVAPI
+        // From DOCS: Applications are expected to bind null UAV to this slot.
+        // NOTE! We don't currently do this, but doesn't seem to be a problem.
+
+        const NvAPI_Status status = NvAPI_D3D12_SetNvShaderExtnSlotSpace(m_device, NvU32(desc.nvapiExtnSlot), NvU32(0));
+        
+        if (status != NVAPI_OK)
         {
-            // TODO(JS): We should test for specific features here.
-            for (Index i = 0; i < SLANG_COUNT_OF(features); ++i)
-            {
-                m_features.add(features[i]);
-            }
-            m_nvapi = true;
+            return SLANG_E_NOT_AVAILABLE;
         }
-    }
 
-    // Set the device
-    m_device = m_deviceInfo.m_device;
+        if (_isSupportedNVAPIOp(m_device, NV_EXTN_OP_UINT64_ATOMIC))
+        {
+            m_features.add("atomic-int64");
+        }
+        if (_isSupportedNVAPIOp(m_device, NV_EXTN_OP_FP32_ATOMIC))
+        {
+            m_features.add("atomic-float");
+        }
+
+        m_nvapi = true;
+#endif
+    }
 
     // Find what features are supported
     {
-- 
cgit v1.2.3