diff --git a/onnxruntime/contrib_ops/cuda/quantization/qordered_ops/qordered_attention.cc b/onnxruntime/contrib_ops/cuda/quantization/qordered_ops/qordered_attention.cc index f14933c2de..813b3ea5ec 100644 --- a/onnxruntime/contrib_ops/cuda/quantization/qordered_ops/qordered_attention.cc +++ b/onnxruntime/contrib_ops/cuda/quantization/qordered_ops/qordered_attention.cc @@ -156,9 +156,9 @@ inline void debug_print([[maybe_unused]] const T* arr, #endif -QOrderedAttention::QOrderedAttention(const OpKernelInfo& info) : CudaKernel(info), AttentionBase(info), input_hidden_size_(0) { +QOrderedAttention::QOrderedAttention(const OpKernelInfo& info) : CudaKernel(info), AttentionBase(info) { #if defined(CUDA_VERSION) && CUDA_VERSION >= 11040 - + input_hidden_size_ = 0; int cuda_runtime_version = 0; CUDA_CALL_THROW(cudaRuntimeGetVersion(&cuda_runtime_version)); ORT_ENFORCE(cuda_runtime_version >= 11040, "QOrderedMatmul need cuda runtime higher than 11.4"); @@ -185,7 +185,7 @@ QOrderedAttention::QOrderedAttention(const OpKernelInfo& info) : CudaKernel(info #else - ORT_ENFORCE(false, "Compiling with CUDA_VERSION >= 11.4 is needed!") + ORT_ENFORCE(false, "Compiling with CUDA_VERSION >= 11.4 is needed!"); #endif } @@ -305,7 +305,7 @@ Status QOrderedAttention::ComputeInternal(OpKernelContext* context) const { #else ORT_UNUSED_PARAMETER(context); - ORT_ENFORCE(false, "Compiling with CUDA_VERSION >= 11.4 is needed!") + ORT_ENFORCE(false, "Compiling with CUDA_VERSION >= 11.4 is needed!"); #endif diff --git a/onnxruntime/contrib_ops/cuda/quantization/qordered_ops/qordered_longformer_attention.cc b/onnxruntime/contrib_ops/cuda/quantization/qordered_ops/qordered_longformer_attention.cc index 141f0d4312..263b914179 100644 --- a/onnxruntime/contrib_ops/cuda/quantization/qordered_ops/qordered_longformer_attention.cc +++ b/onnxruntime/contrib_ops/cuda/quantization/qordered_ops/qordered_longformer_attention.cc @@ -71,7 +71,7 @@ QOrderedLongformerAttention::QOrderedLongformerAttention(const OpKernelInfo& inf #else - ORT_ENFORCE(false, "Compiling with CUDA_VERSION >= 11.4 is needed!") + ORT_ENFORCE(false, "Compiling with CUDA_VERSION >= 11.4 is needed!"); #endif } @@ -267,7 +267,7 @@ QOrderedLongformerAttention::ComputeInternal(OpKernelContext* context) const { #else ORT_UNUSED_PARAMETER(context); - ORT_ENFORCE(false, "Compiling with CUDA_VERSION >= 11.4 is needed!") + ORT_ENFORCE(false, "Compiling with CUDA_VERSION >= 11.4 is needed!"); #endif diff --git a/onnxruntime/contrib_ops/cuda/quantization/qordered_ops/qordered_matmul.cc b/onnxruntime/contrib_ops/cuda/quantization/qordered_ops/qordered_matmul.cc index a64b5ac938..5bd465f440 100644 --- a/onnxruntime/contrib_ops/cuda/quantization/qordered_ops/qordered_matmul.cc +++ b/onnxruntime/contrib_ops/cuda/quantization/qordered_ops/qordered_matmul.cc @@ -52,7 +52,7 @@ QOrderedMatMul::QOrderedMatMul(const OpKernelInfo& info) : CudaKernel(info) { #else - ORT_ENFORCE(false, "Compiling with CUDA_VERSION >= 11.4 is needed!") + ORT_ENFORCE(false, "Compiling with CUDA_VERSION >= 11.4 is needed!"); #endif } @@ -110,7 +110,7 @@ Status QOrderedMatMul::PrePack(const Tensor& tensor, int input_idx, AllocatorPtr ORT_UNUSED_PARAMETER(input_idx); ORT_UNUSED_PARAMETER(alloc); ORT_UNUSED_PARAMETER(is_packed); - ORT_ENFORCE(false, "Compiling with CUDA_VERSION >= 11.4 is needed!") + ORT_ENFORCE(false, "Compiling with CUDA_VERSION >= 11.4 is needed!"); #endif @@ -191,7 +191,7 @@ Status QOrderedMatMul::ComputeInternal(OpKernelContext* context) const { #else ORT_UNUSED_PARAMETER(context); - ORT_ENFORCE(false, "Compiling with CUDA_VERSION >= 11.4 is needed!") + ORT_ENFORCE(false, "Compiling with CUDA_VERSION >= 11.4 is needed!"); #endif diff --git a/onnxruntime/contrib_ops/cuda/quantization/qordered_ops/qordered_qdq.cc b/onnxruntime/contrib_ops/cuda/quantization/qordered_ops/qordered_qdq.cc index b93c7c8f26..50d8cfcdb0 100644 --- a/onnxruntime/contrib_ops/cuda/quantization/qordered_ops/qordered_qdq.cc +++ b/onnxruntime/contrib_ops/cuda/quantization/qordered_ops/qordered_qdq.cc @@ -108,8 +108,8 @@ QuantizeWithOrder::QuantizeWithOrder(const OpKernelInfo& info) : CudaKernel(info #else ORT_UNUSED_PARAMETER(info); - ORT_ENFORCE(false, "Compiling with CUDA_VERSION >= 11.4 is needed!") - + ORT_ENFORCE(false, "Compiling with CUDA_VERSION >= 11.4 is needed!"); + #endif } @@ -133,7 +133,7 @@ DequantizeWithOrder::DequantizeWithOrder(const OpKernelInfo& info) : CudaKernel( #else ORT_UNUSED_PARAMETER(info); - ORT_ENFORCE(false, "Compiling with CUDA_VERSION >= 11.4 is needed!") + ORT_ENFORCE(false, "Compiling with CUDA_VERSION >= 11.4 is needed!"); #endif @@ -183,8 +183,8 @@ Status QuantizeWithOrder::ComputeInternal(OpKernelContext* context) const { #else - ORT_UNUSED_PARAMETER(info); - ORT_ENFORCE(false, "Compiling with CUDA_VERSION >= 11.4 is needed!") + ORT_UNUSED_PARAMETER(context); + ORT_ENFORCE(false, "Compiling with CUDA_VERSION >= 11.4 is needed!"); #endif } @@ -228,8 +228,8 @@ Status DequantizeWithOrder::ComputeInternal(OpKernelContext* context) const { #else - ORT_UNUSED_PARAMETER(info); - ORT_ENFORCE(false, "Compiling with CUDA_VERSION >= 11.4 is needed!") + ORT_UNUSED_PARAMETER(context); + ORT_ENFORCE(false, "Compiling with CUDA_VERSION >= 11.4 is needed!"); #endif }