2.22.3-1

Rework core for NVIDIA Trusted Computing
 * Compress work structs so that they are shared between channels
 * Utilize the full amount of kernel argument space permitted (4k)
   before resorting to work fifo.
 * Rework the task preprocessing phase.
 * Use a separate abortDevFlag which is kept in sync with abortFlag
   using cudaMemcpy operations.
 * Rename src/include/align.h to src/include/bitops.h

Add lazy connection establishment for collective operations
 * Move buffer allocation and connection establishment to the first
   collective operation using that algorithm.
 * Accelerate init time and reduce memory usage.
 * Avoid allocating NVLS buffers if all calls are registered.
 * Compute algo/proto in ncclLaunchCollTasksInfo early on.
 * Connect peers in ncclCollPreconnectFunc if not connected already.
 * Also move shared buffer creation to the first send/recv call.

Accelerate intra-node NVLink detection
 * Make each rank only detect NVLinks attached to its GPU.
 * Fuse XMLs to reconstruct the full NVLink topology

Add init profiling to report time spend in different init phases.
 * Report timings of bootstrap, allgather, search, connect, etc.
 * Add new "PROFILE" category for NCCL_DEBUG_SUBSYS.

Add support for PCI p2p on split PCI switches
 * Detect split PCI switches through a kernel module exposing
   switch information.
 * Update the topology XML and graph to add those inter-switch
   connections.

Add cost estimation API
 * Add a new ncclGroupEndSimulate primitive to return the estimated
   time a group would take.

Net/IB: Add separate traffic class for fifo messages
 * Add NCCL_IB_FIFO_TC to control the traffic class of fifo messages
   independently from NCCL_IB_TC.
   Merges PR #1194

Net/IB: Add support for IB router
 * Use flid instead of lid if subnets do not match
 * Warn if flid is 0

Optimizations and fixes for device network offload (unpack)
 * Double the default number of channels
 * Cache netDeviceType
 * Fix save/increment head logic to enable Tree support.

Support ncclGroupStart/End for ncclCommAbort/Destroy
 * Allow Abort/Destroy to be called within a group when managing
   multiple GPUs with a single process.

Improve Tuner API
 * Provide to the plugin the original cost table so that the plugin
   can leave unknown or disabled algo/proto combinations untouched.
 * Remove nvlsSupport and collnetSupport.

Do not print version to stdout when using a debug file
 * Also print version from all processes with INFO debug level.
   Fixes issue #1271

Fix clang warnings in NVTX headers
 * Update NVTX headers to the latest version
   Fixes issue #1270

Disable port fusion in heterogeneous systems
 * Do not fuse ports if a mix of multi-port and single port are detected.

Fix NVLS graphs search for dual NICs.
 * Fix NVLS graph search when we have more than one NIC per GPU.

Fix crash with collnetDirect
 * Add separate graph search for collnetDirect, testing alltoall paths
   and working similarly to the NVLS search.

Fix hang when nodes have different CPU types
 * Add the CPU type to the rank peer info.
 * Align all ranks on the CPU type after the first allgather.
 * Only use the aligned CPU type for all tuning operations.
   Fixes issue #1136
   Fixes issue #1184

Fix performance of registered send/recv operations
 * Allow for single full size operations
 * Add INFO to confirm the registration of send/recv buffers.

Move all sync ops to finalize stage
 * Ensure ncclCommDestroy is non-blocking if ncclCommFinalize has
   been called.

Improve error reporting during SHM segment creation

Improve support of various compilers
   Merges PR #1177
   Merges PR #1228

Allow net and tuner plugins to be statically linked
 * Search for ncclNet or ncclTuner symbols in the main binary.
   Merges PR #979

Plugin examples includes cleanup
 * Harmonize err.h and common.h usage.
 * Add mixed plugin with both net and tuner.

src/device/reduce_kernel.h

@@ -37,6 +37,7 @@ template<typename T>
 struct FuncSum  { using EltType = T; __device__ FuncSum(uint64_t opArg=0) {}; };
 template<typename T>
 struct FuncProd { using EltType = T; __device__ FuncProd(uint64_t opArg=0) {}; };
+
 template<typename T>
 struct FuncMinMax {
   using EltType = T;
@@ -47,9 +48,30 @@ struct FuncMinMax {
     isMinNotMax = (opArg&1)==0;
   }
 };
+
 template<typename T> struct FuncPreMulSum;
 template<typename T> struct FuncSumPostDiv;
 
+////////////////////////////////////////////////////////////////////////////////
+// Trait class for handling the reduction argument.
+
+template<typename Fn>
+struct RedOpArg { // default case: no argument
+  static constexpr bool ArgUsed = false;
+  __device__ static uint64_t loadArg(void *ptr) { return 0; }
+};
+
+template<typename T>
+struct RedOpArg<FuncMinMax<T>> {
+  static constexpr bool ArgUsed = true;
+  __device__ static uint64_t loadArg(void *ptr) {
+    union { uint64_t u64; T val; };
+    u64 = 0;
+    val = *(T*)ptr;
+    return u64;
+  }
+};
+
 ////////////////////////////////////////////////////////////////////////////////
 // Trait classes for reduction functions. Given a function (FuncSum, etc.)
 // and a number of elements in a pack, will reduce, preOp, or postOp a pack
@@ -356,6 +378,17 @@ struct Apply_PostOp<Fn, /*EltPerPack=*/0> {
 ////////////////////////////////////////////////////////////////////////////////
 // FuncPreMulSum
 
+template<typename T>
+struct RedOpArg<FuncPreMulSum<T>> {
+  static constexpr bool ArgUsed = true;
+  __device__ static uint64_t loadArg(void *ptr) {
+    union { uint64_t u64; T val; };
+    u64 = 0;
+    val = *(T*)ptr;
+    return u64;
+  }
+};
+
 // General definition for all integral types, float, and double.
 template<typename T>
 struct FuncPreMulSum {
@@ -486,6 +519,14 @@ struct Apply_PreOp<FuncPreMulSum<half>, /*EltPerPack=*/1> {
 ////////////////////////////////////////////////////////////////////////////////
 // FuncSumPostDiv
 
+template<typename T>
+struct RedOpArg<FuncSumPostDiv<T>> {
+  static constexpr bool ArgUsed = true;
+  __device__ static uint64_t loadArg(void *ptr) {
+    return *(uint64_t*)ptr;
+  }
+};
+
 template<typename T, bool IsFloating=IsFloatingPoint<T>::value>
 struct FuncSumPostDiv_IntOnly;
 
@@ -658,7 +699,7 @@ struct Apply_LoadMultimem {
     static constexpr bool IsFloat = IsFloatingPoint<T>::value;
     static constexpr int BigPackSize =
       IsFloat && IsSum && sizeof(T) < 8 ? 16 :
-      IsFloat && IsSum ? 8 :
+      IsFloat && IsSum ? sizeof(T) :
       IsFloat && IsMinMax && sizeof(T)==2 ? 16 :
       !IsFloat && (IsSum||IsMinMax) && sizeof(T)>=4 ? sizeof(T) :
       /*multimem.ld_reduce not supported:*/ 0;