A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 14, 23:45 UTC
[Bugfix][Apple Silicon] fix missing symbols when build from source on Mac with Apple Silicon (#21380)
Signed-off-by: Yeju Zhou <yejuzhou@outlook.com>
Yeju Zhou committed
Jul 26, 2025 at 22:09 UTC
9094d11c5d74db904dff9e9b1c63da5d73707eeb
1 file changed
+1
-1
csrc/cpu/torch_bindings.cpp
+1
-1
@@ -151,7 +151,7 @@ TORCH_LIBRARY_EXPAND(TORCH_EXTENSION_NAME, ops) {
151
ops.impl("rotary_embedding", torch::kCPU, &rotary_embedding);
152
153
// Quantization
154
-#if defined(__AVX512F__) || defined(__aarch64__)
154
+#if defined(__AVX512F__) || (defined(__aarch64__) && !defined(__APPLE__))
155
at::Tag stride_tag = at::Tag::needs_fixed_stride_order;
156
157
// Compute int8 quantized tensor for given scaling factor.