chore: update nixl to 0.8.0 and ucx 1.20 (#5007)

Signed-off-by: Nate Mailhot <nmailhot@nvidia.com>

Cargo.lock

@@ -6769,9 +6769,9 @@ dependencies = [
 
 [[package]]
 name = "nixl-sys"
-version = "0.7.1"
+version = "0.8.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "6d80bd4b5b8363cfd933000a8757a453e58ee10ee6e400c38ae31db512444a31"
+checksum = "06518de4709489510ec1fa8938f1f510f8c9c1f154b096c3c0913ab2feade9b0"
 dependencies = [
  "bindgen 0.71.1",
  "cc",

container/build.sh

@@ -123,8 +123,8 @@ SGLANG_BASE_IMAGE_TAG="25.01-cuda12.8-devel-ubuntu24.04"
 SGLANG_FRAMEWORK_IMAGE="nvcr.io/nvidia/cuda"
 SGLANG_FRAMEWORK_IMAGE_TAG="${SGLANG_CUDA_VERSION}-cudnn-devel-ubuntu24.04"
 
-NIXL_REF=0.7.1
-NIXL_UCX_REF=v1.19.0
+NIXL_REF=0.8.0
+NIXL_UCX_REF=v1.20.0-rc1
 NIXL_GDRCOPY_REF=v2.5.1
 NIXL_LIBFABRIC_REF=v2.3.0
 

container/deps/trtllm/install_nixl.sh

@@ -23,7 +23,7 @@ set -ex
 
 GITHUB_URL="https://github.com"
 
-UCX_VERSION="v1.19.1"
+UCX_VERSION="v1.20.0-rc1"
 UCX_INSTALL_PATH="/usr/local/ucx/"
 CUDA_PATH="/usr/local/cuda"
 

docs/reference/support-matrix.md

@@ -54,7 +54,7 @@ If you are using a **GPU**, the following GPU models and architectures are suppo
 | :----------------- | :---------- | :------------------------------------ | :----------- |
 | ai-dynamo          | 0.8.0       | >=2.28                                |              |
 | ai-dynamo-runtime  | 0.8.0       | >=2.28 (Python 3.12 has known issues) |              |
-| NIXL               | 0.7.1       | >=2.27                                | >=11.8       |
+| NIXL               | 0.8.0       | >=2.27                                | >=11.8       |
 
 ### Build Dependency
 
@@ -65,7 +65,7 @@ The following table shows the dependency versions included with each Dynamo rele
 | SGLang         | 0.5.6          | 0.5.6.post1             | 0.5.3.post4| 0.5.3.post4      | 0.5.3.post4|
 | TensorRT-LLM   | 1.2.0rc5       | 1.2.0rc6                | 1.2.0rc3   | 1.2.0rc3         | 1.2.0rc2   |
 | vLLM           | 0.12.0         | 0.12.0                  | 0.11.0     | 0.11.0           | 0.11.0     |
-| NIXL           | 0.8.0          | 0.8.0                   | 0.7.1      | 0.7.1            | 0.7.1      |
+| NIXL           | 0.8.0          | 0.8.0                   | 0.8.0      | 0.8.0            | 0.8.0      |
 
 > [!Note]
 > **main (ToT)** reflects the current development branch. **v0.8.0** is the upcoming release (planned for January 14, 2025) and not yet available.

lib/bindings/kvbm/Cargo.lock

@@ -4137,9 +4137,9 @@ dependencies = [
 
 [[package]]
 name = "nixl-sys"
-version = "0.7.1"
+version = "0.8.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "6d80bd4b5b8363cfd933000a8757a453e58ee10ee6e400c38ae31db512444a31"
+checksum = "06518de4709489510ec1fa8938f1f510f8c9c1f154b096c3c0913ab2feade9b0"
 dependencies = [
  "bindgen 0.71.1",
  "cc",

lib/bindings/kvbm/pyproject.toml

@@ -26,7 +26,7 @@ license = { text = "Apache-2.0" }
 license-files = ["LICENSE"]
 requires-python = ">=3.10"
 dependencies = [
-    "nixl[cu12]==0.7.1",
+    "nixl[cu12]==0.8.0",
     "pydantic>=2.0"
 ]
 classifiers = [
@@ -45,8 +45,8 @@ classifiers = [
 keywords = ["llm", "genai", "inference", "nvidia", "kvcache", "dynamo"]
 
 [project.optional-dependencies]
-cu12 = ["nixl[cu12]==0.7.1"]
-cu13 = ["nixl[cu13]==0.7.1"]
+cu12 = ["nixl[cu12]==0.8.0"]
+cu13 = ["nixl[cu13]==0.8.0"]
 test = [
     "pytest>=8.3.4",
     "pytest-mypy",

lib/bindings/python/Cargo.lock

@@ -561,7 +561,7 @@ dependencies = [
  "bitflags 2.10.0",
  "cexpr",
  "clang-sys",
- "itertools 0.11.0",
+ "itertools 0.12.1",
  "proc-macro2",
  "quote",
  "regex",
@@ -579,7 +579,7 @@ dependencies = [
  "bitflags 2.10.0",
  "cexpr",
  "clang-sys",
- "itertools 0.11.0",
+ "itertools 0.12.1",
  "log",
  "prettyplease",
  "proc-macro2",
@@ -1483,7 +1483,7 @@ dependencies = [
  "libc",
  "option-ext",
  "redox_users",
- "windows-sys 0.61.2",
+ "windows-sys 0.59.0",
 ]
 
 [[package]]
@@ -1963,7 +1963,7 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "39cab71617ae0d63f51a36d69f866391735b51691dbda63cf6f96d042b63efeb"
 dependencies = [
  "libc",
- "windows-sys 0.61.2",
+ "windows-sys 0.52.0",
 ]
 
 [[package]]
@@ -3316,7 +3316,7 @@ dependencies = [
  "portable-atomic",
  "portable-atomic-util",
  "serde_core",
- "windows-sys 0.61.2",
+ "windows-sys 0.52.0",
 ]
 
 [[package]]
@@ -4175,9 +4175,9 @@ dependencies = [
 
 [[package]]
 name = "nixl-sys"
-version = "0.7.1"
+version = "0.8.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "6d80bd4b5b8363cfd933000a8757a453e58ee10ee6e400c38ae31db512444a31"
+checksum = "06518de4709489510ec1fa8938f1f510f8c9c1f154b096c3c0913ab2feade9b0"
 dependencies = [
  "bindgen 0.71.1",
  "cc",
@@ -4259,7 +4259,7 @@ version = "0.50.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "7957b9740744892f114936ab4a57b3f487491bbeafaf8083688b16841a4240e5"
 dependencies = [
- "windows-sys 0.61.2",
+ "windows-sys 0.59.0",
 ]
 
 [[package]]
@@ -6065,7 +6065,7 @@ dependencies = [
  "errno",
  "libc",
  "linux-raw-sys",
- "windows-sys 0.61.2",
+ "windows-sys 0.52.0",
 ]
 
 [[package]]
@@ -6861,7 +6861,7 @@ dependencies = [
  "getrandom 0.3.4",
  "once_cell",
  "rustix",
- "windows-sys 0.61.2",
+ "windows-sys 0.52.0",
 ]
 
 [[package]]
@@ -8089,7 +8089,7 @@ version = "0.1.11"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "c2a7b1c03c876122aa43f3020e6c3c3ee5c05081c9a00739faf7503aeba10d22"
 dependencies = [
- "windows-sys 0.61.2",
+ "windows-sys 0.48.0",
 ]
 
 [[package]]

lib/llm/Cargo.toml

@@ -103,7 +103,7 @@ dialoguer = { version = "0.11", default-features = false, features = [
 
 # block_manager
 aligned-vec = { version = "0.6.4", optional = true }
-nixl-sys = { version = "=0.7.1", optional = true }
+nixl-sys = { version = "=0.8.0", optional = true }
 cudarc = { workspace = true, optional = true }
 nix = { version = "0.26", optional = true }
 

lib/memory/Cargo.toml

@@ -26,7 +26,7 @@ dynamo-config = { workspace = true }
 
 anyhow = { workspace = true }
 cudarc = { workspace = true }
-nixl-sys = { version = "0.7" }
+nixl-sys = { version = "0.8" }
 serde = { workspace = true}
 thiserror = { workspace = true }
 tracing = { workspace = true }

pyproject.toml

@@ -55,13 +55,13 @@ trtllm =[
 
 vllm = [
     "uvloop",
-    "nixl[cu12]<=0.7.1",
+    "nixl[cu12]<=0.8.0",
     "vllm[flashinfer,runai]==0.12.0",
 ]
 
 sglang = [
     "uvloop",
-    "nixl[cu12]<=0.7.1",
+    "nixl[cu12]<=0.8.0",
     "sglang==0.5.6",
 ]