NCCL 2.27.5-1

Improvements for GB200 systems
* Optimize the network performance by alternating the direction of the
  rings and the NIC to GPU assignment across communicators to limit
  unnecessary sharing.
* Fix the detection of C2C links in case GPU Direct RDMA is disabled
  between a GPU and a NIC.
* Fix PXN support on MNNVL systems, where NCCL would try (and fail) to
  share regular host memory across multiple nodes.
* Fix P2C (PXN over C2C), which is now preferred over regular PXN.  This
  support is currently preliminary and is disabled by default; use
  NCCL_PXN_C2C=1 to enable.

Further reduce the overheads of CUDA graph capturing, which increased in
NCCL 2.26.2 for large graphs.

Optimize the network performance on DGX B200 systems by adjusting the
bandwidths provided to the graph search algorithm.

Enable fp8 reductions in symmetric kernels on Blackwell with CUDA 12.8.

Restore the plugin name handling logic to make it possible to specify a
path to the plugin (Issue #1732).

Restore the ability to change NCCL_COLLNET_ENABLE during execution
(Issue #1741).

Add an example tuner plugin with CSV-based overrides.

Remove an x86 dependency from the example profiler.

src/device/Makefile

@@ -36,9 +36,8 @@ define COMPILE
  $(call COMPILE$(or $3,$(suffix $2)),$1,$2)
 endef
 
-ifeq ($(shell echo "$$((1000*$(CUDA_MAJOR) + 10*$(CUDA_MINOR) >= 12080))"),1)
-	NVCC_GENCODE_LDMC_FP8 = -gencode=arch=compute_100a,code=sm_100a \
-                     -gencode=arch=compute_120a,code=sm_120a
+ifeq ($(shell echo "$$((1000*$(CUDA_MAJOR) + 10*$(CUDA_MINOR) >= 12090))"),1)
+	NVCC_GENCODE_LDMC_FP8 = -gencode=arch=compute_100f,code=sm_100f
 else ifeq ($(shell echo "$$((1000*$(CUDA_MAJOR) + 10*$(CUDA_MINOR) >= 12070))"),1)
   NVCC_GENCODE_LDMC_FP8 = -gencode=arch=compute_100a,code=sm_100a
 else

src/device/reduce_kernel.h

@@ -1009,7 +1009,7 @@ struct Apply_LoadMultimem {
     DEFINE_Apply_LoadMultimem_minmax_v4_and_xparts(__nv_bfloat16, bf16x2, 4)
   #endif
 
-  #if NCCL_CUDA_ARCH_FAMILY_SPECIFIC == 1000 || NCCL_CUDA_ARCH_FAMILY_SPECIFIC == 1010 || NCCL_CUDA_ARCH_SPECIFIC == 1200 || NCCL_CUDA_ARCH_SPECIFIC == 1210
+  #if NCCL_CUDA_ARCH_SPECIFIC == 1000 || NCCL_CUDA_ARCH_SPECIFIC == 1010 || NCCL_CUDA_ARCH_FAMILY_SPECIFIC == 1000 || NCCL_CUDA_ARCH_FAMILY_SPECIFIC == 1010 || NCCL_CUDA_ARCH_SPECIFIC == 1200 || NCCL_CUDA_ARCH_SPECIFIC == 1210
     DEFINE_Apply_LoadMultimem_sum_v4_and_xparts(__nv_fp8_e4m3, e4m3x4, 4)
     DEFINE_Apply_LoadMultimem_minmax_v4_and_xparts(__nv_fp8_e4m3, e4m3x4, 4)
     DEFINE_Apply_LoadMultimem_sum_v4_and_xparts(__nv_fp8_e5m2, e5m2x4, 4)

src/device/symmetric/generate.py

@@ -108,7 +108,7 @@ def required_cuda(k):
       if k.algo in ldmc_algos:
         cudart = 12070
         arch = None
-        specific_sms = [100, 120]
+        specific_sms = ["100a", "101a", "100f", "101f", "120a", "121a"]
   return (cudart, arch, specific_sms)
 
 ################################################################################
@@ -145,7 +145,7 @@ def kernel_conds(k):
   if not specific_sms:
     arch_cond = "__CUDA_ARCH__ >= %d"%arch
   else:
-    arch_cond = " || ".join(["0"] + ["NCCL_CUDA_ARCH_SPECIFIC==%d"%(10*sm) for sm in specific_sms])
+    arch_cond = " || ".join(["0"] + ["NCCL_CUDA_ARCH_%sSPECIFIC==%d"%("FAMILY_" if sm[-1] == "f" else "", 10*int(sm.replace('a', '').replace('f', ''))) for sm in specific_sms])
   return cudart_cond, arch_cond
 
 def instantiate(k):