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 9, 12:30 UTC
[Docs] Simplify CPU x86 Docker build documentation (#33071)
Signed-off-by: Maryam Tahhan <mtahhan@redhat.com>
Maryam Tahhan committed
Jan 28, 2026 at 06:37 UTC
2dd359f953789b1a2d950f096480039a252f29d7
1 file changed
+9
-23
docs/getting_started/installation/cpu.x86.inc.md
+9
-23
@@ -171,10 +171,6 @@ uv pip install dist/*.whl
171
172
## Building for your target CPU
173
174
-vLLM supports building Docker images for x86 CPU platforms with automatic instruction set detection.
175
-
176
-### Basic build command
177
-
174
```bash
175
docker build -f docker/Dockerfile.cpu \
176
--build-arg VLLM_CPU_DISABLE_AVX512=<false (default)|true> \
@@ -187,29 +183,21 @@ docker build -f docker/Dockerfile.cpu \
183
--target vllm-openai .
184
```
185
190
-!!! note "Instruction set auto-detection"
191
- By default, vLLM will auto-detect CPU instruction sets (AVX512, AVX2, etc.) from the build system's CPU flags. Build arguments like `VLLM_CPU_AVX2`, `VLLM_CPU_AVX512`, `VLLM_CPU_AVX512BF16`, `VLLM_CPU_AVX512VNNI`, and `VLLM_CPU_AMXBF16` are primarily used for **cross-compilation** or for building container images on systems that don't have the target platforms ISA:
186
+!!! note "Auto-detection by default"
187
+ By default, CPU instruction sets (AVX512, AVX2, etc.) are automatically detected from the build system's CPU flags. Build arguments like `VLLM_CPU_AVX2`, `VLLM_CPU_AVX512`, `VLLM_CPU_AVX512BF16`, `VLLM_CPU_AVX512VNNI`, and `VLLM_CPU_AMXBF16` are used for cross-compilation:
188
193
- - Set `VLLM_CPU_{ISA}=true` to force-enable an instruction set (for cross-compilation to target platforms with that ISA)
194
- - Set `VLLM_CPU_{ISA}=false` to rely on auto-detection
195
- - When an ISA build arg is set to `true`, vLLM will build with that instruction set regardless of the build system's CPU capabilities
189
+ - `VLLM_CPU_{ISA}=true` - Force-enable the instruction set (build with ISA regardless of build system capabilities)
190
+ - `VLLM_CPU_{ISA}=false` - Rely on auto-detection (default)
191
197
-### Build examples
192
+### Examples
193
199
-**Example 1: Auto-detection (native build)**
200
-
201
-Build on a machine with the same CPU as your target deployment:
194
+**Auto-detection build (default)**
195
196
```bash
204
-# Auto-detects all CPU features from the build system
205
-docker build -f docker/Dockerfile.cpu \
206
- --tag vllm-cpu-env \
207
- --target vllm-openai .
197
+docker build -f docker/Dockerfile.cpu --tag vllm-cpu-env --target vllm-openai .
198
```
199
210
-**Example 2: Cross-compilation for AVX512 deployment**
211
-
212
-Build an AVX512 image on any x86_64 system (even without AVX512):
200
+**Cross-compile for AVX512**
201
202
```bash
203
docker build -f docker/Dockerfile.cpu \
@@ -220,9 +208,7 @@ docker build -f docker/Dockerfile.cpu \
208
--target vllm-openai .
209
```
210
223
-**Example 3: Cross-compilation for AVX2 deployment**
224
-
225
-Build an AVX2 image for older CPUs:
211
+**Cross-compile for AVX2**
212
213
```bash
214
docker build -f docker/Dockerfile.cpu \