From: Samuel Pitoiset Date: Thu, 20 Sep 2018 20:17:03 +0000 (+0200) Subject: radv: only enable shaderInt16 on GFX9+ and LLVM7+ X-Git-Url: https://git.libre-soc.org/?a=commitdiff_plain;h=674fcfaecca760b6d1befcf78c39dad184ad452d;p=mesa.git radv: only enable shaderInt16 on GFX9+ and LLVM7+ The throughput is similar to 32-bit integers on GFX8 and AMDVLK does not expose 16-bit integers on pre Vega as well. On GFX9+, only LLVM 7+ has support. This fixes a bunch of CTS crashes on GFX9/LLVM 6. Signed-off-by: Samuel Pitoiset Reviewed-by: Bas Nieuwenhuizen --- diff --git a/src/amd/vulkan/radv_device.c b/src/amd/vulkan/radv_device.c index 31d9bb59637..f7752eac83b 100644 --- a/src/amd/vulkan/radv_device.c +++ b/src/amd/vulkan/radv_device.c @@ -763,7 +763,7 @@ void radv_GetPhysicalDeviceFeatures( .shaderCullDistance = true, .shaderFloat64 = true, .shaderInt64 = true, - .shaderInt16 = true, + .shaderInt16 = pdevice->rad_info.chip_class >= GFX9 && HAVE_LLVM >= 0x700, .sparseBinding = true, .variableMultisampleRate = true, .inheritedQueries = true,