* Use one side stream per process
* Handle multiple GPUs per process
* Reset stream when not found
* Address review comments
* Fix missing mutex initializer
[ROCm/rccl commit: 185e78a8f0]
* Fail the job if compiler flag HIP_HOST_UNCACHED_MEMORY is not turned on on mi350x
Place the check after initTransportsRank as the GPU arch info in comm->topo->nodes info is populated after that.
* Update src/init.cc to use ERROR instead of WARN
Co-authored-by: Nilesh M Negi <Nilesh.Negi@amd.com>
[ROCm/rccl commit: 05f914c997]
* Enhance logging in NCCL initialization
It's convenient to log comms obj and default channels together for debugging
* Add opCount to collDevWork and update increment logic
Added opCount to collDevWork and incremented it when proxyOpQueue is empty (e.g., for intra-node comms)
* Clarify opCount increment logic in enqueue.cc
Updated comment to clarify incrementing opCount for intranode communications.
* Refactor NCCL_INIT logging format
Updated logging format for NCCL_INIT to improve clarity.
* Remove duplicate INFO logging in init.cc
[ROCm/rccl commit: b00ee4c83c]
* Added ERROR message class to handle fatal error messages.
New ERROR message class will print the message in all debug level,
including none.
Change some of the fatal error message to be in ERROR instead of WARN.
Added new error handler function to print out more meaningful error
message in the future.
* Added CHANGELOG entry.
* Update CHANGELOG.md
Co-authored-by: Jeffrey Novotny <jnovotny@amd.com>
* Change to no longer reuse NONE as ERROR. ERROR is now a separated class.
* Update CHANGELOG.md
Co-authored-by: Jeffrey Novotny <jnovotny@amd.com>
---------
Co-authored-by: Jeffrey Novotny <jnovotny@amd.com>
[ROCm/rccl commit: 1ce83d5cc0]
* Add initial commit to increase tb size to 512
* Fix LL perf issue when subset of NCCL_MAX_NTHREADS is used
Adding a constant to barrier_generic logic from using fallback logic when nthreads < NCCL_MAX_NTHREADS and nthreads == blockDim.X
* Adjust nthreads for LL
* Opt threads for reduce_scatter upper small range
* Add macro for single node
* Restrict MSCCL to 256 threads to prevent mem access fault
* Support pre-MI350 compatibility
* Partially refactor threadblock size override
* Use const macros instead of numerals
* opt out of unused function
[ROCm/rccl commit: 12f51ba8bf]
* Gate based on ROCM version, safe for ROCm 7.0.2 and beyond.
* Updates naming to gfx9CheapFenceOff since we use this for gfx942 and gfx950. Thanks Nilesh.
* Add info logging statement to NCCL_INIT to print whether enabled when INFO logging is enabled.
[ROCm/rccl commit: c70f5b4621]
In this commit it disabled by default and can be enabled via
`RCCL_ENABLE_CONTEXT_TRACKING=1` for both (CDNA, RDNA)
Original PR https://github.com/ROCm/rccl/pull/1927
[ROCm/rccl commit: 00a42c80f3]
* Revert disabling of context tracking for Radeon
Original commit df3b7e47
`Disable context tracking for the current version. (#1839)`
* Add env variable for disabling of context tracking for Radeon
`export NCCL_DISABLE_CONTEXT_TRACKING=1` to force disable of context tracking
* Update docs/how-to/rccl-usage-tips.rst
Fix grammar, thanks @amd-jnovotny
Co-authored-by: Jeffrey Novotny <jnovotny@amd.com>
* Rename NCCL_DISABLE_CONTEXT_TRACKING -> RCCL_DISABLE_CONTEXT_TRACKING
* Revert changes in includes and rename util function
---------
Co-authored-by: Jeffrey Novotny <jnovotny@amd.com>
[ROCm/rccl commit: 07925ec027]
* [rocm_regression] Return errors when HSA_NO_SCRATCH_RECLAIM=1 even for rocm >= 6.4.0
* [rocm_regression] Check firmware version
* [rocm_regression] Resolve review comments
* [rocm_regression] Move hsa env checking into init once func
* [rocm_regression] Prevent hot fix version in firmware
* [rocm_regression] Improve unit tests
[ROCm/rccl commit: 361d596229]
* add direct allgather algorithm
* minor fix
* add debug print for memory allocation tracker
* add message size threshold for direct allgather
* scatter transfers across ranks
* update changelog
* minor fix
* Update CHANGELOG.md
Co-authored-by: Jeffrey Novotny <jnovotny@amd.com>
* enable direct AG when pxn is ON on MI300X or MI350
---------
Co-authored-by: Jeffrey Novotny <jnovotny@amd.com>
[ROCm/rccl commit: 5e7937effb]
Leverages the traits of extended-scope fine-grain memory to get rid of a device-scope acquire-release fence. This improves throughput for single node workloads on gfx942 and gfx950 for some input sizes (e.g., ~32 MiB to about 256 MiB) when using the simple protocol. Multinode workloads on MI300X see a smaller but statistically significant uplift for some message sizes. Runtime disablement is supported via setting the environment variable RCCL_GFX942_CHEAP_FENCE_ON to 0.
[ROCm/rccl commit: 1aa2570b48]
* Support fused all reduce and elementwise operations
Add additional "acc" parameter to RCCL Replayer logs
Add flag which indicates availability of new API
* Fix Recorder json parsing
* Remove unreachable code
* Remove extra acc pointer check
* .
* Revert "[DEVICE] Adding ability to choose unroll factor at runtime (#1734)"
This reverts commit 4cadf3597c.
* Use noinline to reduce kernels linking time
* Don't use noinline for gfx942 and gfx950 to avoid perf regression
---------
Co-authored-by: AtlantaPepsi <timhu102@amd.com>
Co-authored-by: BertanDogancay <bertan.dogancay@gmail.com>
[ROCm/rccl commit: 9a4213356d]
Boosts single node bfloat16 allreduce performance by up to 20% for some data sizes and provides gating with the RCCL_GFX942_CHEAP_FENCE_OFF environment variable
[ROCm/rccl commit: 11fabf1de1]
Improve support for DirectNIC (CX8)
* Add support for XDR speed detection.
* When DirectNIC is enabled, report only the RDMA interfaces.
Extend the P2C (PXN over C2C) support to send/receive operations.
Support compilation with GCC 14 (Issues #1743, #1751).
Fix the unloading of network plugins that also provide tuner capability.
Fix the change of the current device across the calls to ncclCommDestroy()
and ncclCommAbort().
A note for users on MNNVL systems: please ensure an adequate stack size for
NCCL threads. While the default Linux stack size limit of 8192 KB is known
to be sufficient, we've seen crashes if the limit is changed to
"unlimited", as it causes the glibc library to unexpectedly *decrease* the
stack size of NCCL's background threads to just 2048 KB. Use "ulimit -s"
in bash to print the current limit; if needed, reset it to 8192 KB using
"ulimit -s 8192" (one also needs to ensure that the new setting is
propagated to other nodes when launching a multi-node NCCL job).
[ROCm/rccl commit: 7c12c627c6]
This feature tracks the proxy events and status of each send/recv op. ProxyTrace keeps a fixed number of active ops in host mem and dumps the status of each op when the program crashes or hangs.
[ROCm/rccl commit: 020dcf0a7c]
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.
[ROCm/rccl commit: 3ea7eedf3b]
Symmetric memory API and symmetric kernels
* Redesign from the ground up, enabling major latency and bandwidth
improvements.
* Add new API calls to register user-allocated memory among communicator
ranks into a NCCL window: ncclCommWindowRegister() and
ncclCommWindowDeregister(). The calls currently support symmetric
registration for P2P and NVLS, and require VMM memory buffers (i.e.,
CUMEM must be operational).
* Implement specialized kernels taking advantage of symmetrically
registered memory, with performance gains expected particularly for
small to medium message sizes.
* The kernels support 32 bit floating point types and smaller, and sum as
the reduction operator, with no more than one collective operation per
group.
* Floating point summation is always done in fp32 accumulators (with the
exception of fp8 on NVLS, where it uses fp16 inside the switch). Thus,
the accuracy with fp8 and fp16 data types should be much improved.
* This initial implementation supports non-network communicators only (P2P
and NVLS transports).
* To explore this functionality users need to use the new memory
registration API calls with the NCCL_WIN_COLL_SYMMETRIC flag and all
ranks of a communicator must pass buffers at the same offset in the same
registration when invoking a collective NCCL operation.
Add support for DGX Spark.
Add support for DirectNIC (CX8) to the internal IB plugin.
Add a new ncclCommShrink() API call
* It is a non-collective call similar to ncclCommSplit(), which makes it
possible to exclude some (possibly unresponsive) ranks from the parent
communicator.
Add support for loading multiple network plugins
* This enables the creation of generic containers that can work across a
range of providers.
* Allow NCCL_NET_PLUGIN to accept a comma-separated list of plugins to
load.
NVLink SHARP (NVLS) improvements
* Implement NVLS+IB SHARP support for AllGather and ReduceScatter with
user buffer registration. This improves performance and reduces the
number of CTAs needed to achieve peak bandwidth.
* Gracefully fall back by default to other transports if NVLS
initialization fails (the old behavior of returning an error code from a
NCCL call can be preserved by setting NCCL_NVLS_ENABLE=1).
* Decrease the NVLS channel count to 24 on Blackwell systems with multiple
NVLink domains per communicator.
* Enable fine-tuning of NCCL behavior per communicator using new
"ncclConfig_t" members "collnetEnable", "CTAPolicy", and "nvlsCTAs".
Profiler improvements
* Extend the init function by adding communicator name, comm id (hash),
rank, number of ranks, number of nodes, and the NCCL log function to the
argument list. This makes the name and the comm id available to all
events in the communicator without explicitly passing them to each
individual event. Add the communicator id and rank to the profiler trace
filename. Now, the communicator name can be set via a new "ncclConfig_t"
member "commName".
* Improve the accuracy of the GPU kernel events by providing GPU-generated
timestamps for the start and stop of every NCCL operation.
* Harmonize proxy events, removing overlaps between ProxyOp and ProxyStep
states.
* Add support for network-defined event updates (through
"recordEventState").
* Report the correct number of channels used by every collective/p2p
operation (used to be set to nMaxChannels for collectives and absent for
p2ps).
* Fix the logic on proxyCtrl Idle/Active events (Issue #1162).
* Fix an issue where the network proxy profiler could lose track of an
event identifier (Issue #1682).
* Improve the backward compatibility with plugins older than v4.
* Ensure that the work counters are 0-initialized.
* Fix a potential race condition in the network profiler that could result
in an event being linked to a wrong parent.
MNNVL improvements
* Increase to 16 the number of NICs used to communicate between MNNVL
domains on GB200 systems, to optimize the performance of collective
operations.
* Add support for more complex MNNVL topologies with up to 32 NICs per
node.
* If the MNNVL fabric initialization was unsuccessful, NCCL will now fail
by default, so as to avoid inadvertently falling back to a potentially
much slower network transport. Such failures are typically due to a
misconfigured IMEX support on the system. To continue without MNNVL,
restart the job with NCCL_MNNVL_ENABLE=0.
* Fix a potential hang in alltoall-like communication patterns at a scale
of over 80 ranks.
* Make NCCL_P2P_DISABLE=1 imply NCCL_MNNVL_ENABLE=0 (so the latter no
longer needs to be specified on MNNVL systems).
* Fix an initialization failure when NCCL_TOPO_FILE is used on MNNVL
systems.
* Fix the graph search to exclude non-local NICs.
* Fix the SHM transport to use fabric handles on MNNVL systems.
NIC Fusion improvements
* Disable the creation of fused NICs for physical devices that haven't
been merged.
* Flatten multiple ports to a single PCI device within the internal IB
plugin and reparent dual-port NICs under the first PCI parent. If the
parent is not a PCI switch, PCI devices for fused NICs won't be
duplicated.
* Route traffic on GB200-CX8 systems through DirectNIC, not the host
interface.
Improve support for platforms with C2C connectivity (e.g., GB200)
* Enable GPUDirect RDMA for the NICs by default.
* Add support for P2C (PXN over C2C) and the LL128 protocol.
Extend NCCL fault tolerance in multithreaded scenarios
* Support the creation of multiple nonblocking communicators within a
single group and polling in parallel for the completion using multiple
threads (one per communicator).
Enable ncclImplicitOrderLaunch for CUDA 12.9+
* This can potentially speed up NCCL_IMPLICIT_LAUNCH_ORDER.
Improve the netSocket transport latency and control
* Provide finer control over the size of the socket send/receive buffers,
the task size, and the number of sockets that a single peer can open.
* Add support for the inlining of small messages behind the header when
using multiple sockets per connection.
Improve the readability of the CPU affinity in the debug output
* Print it as a range string rather than a bitmask.
Fix a potential race condition in graph execution
* A contention could arise when mixing graph and non-graph execution.
Improve PXN connection code
* Avoid duplicate and unused connections.
RAS fixes
* Fix a memory corruption at job termination time in case of a previously
failed initialization of a RAS socket connection.
* Fix a race condition leading to a crash when generating a RAS report
during communicator initialization (Issues #1669, #1718).
* Fix a potential race condition when gathering data for a RAS status
report.
Fix a potential memory corruption in ncclCommSplit()
* Memory could get corrupted when resource sharing was in use and the size
of the NVLink domain in the new communicator was smaller than in the old
one.
Fix asynchronous graph upload
* Fix a small memory leak.
* Fix oversychronization.
Add a check for out-of-memory conditions in ncclMemAlloc()
Clean up the NCCL socket code
* accept() will retry also if just reading the magic failed (Issue #1613).
* connect() will retry also if poll() did not return a POLLOUT event
(Issue #1618).
* Add error checking in a few instances (Issue #1539).
* Fix the loop condition in ncclFindInterfaceMatchSubnet() (Issue #1574).
* Clean up the debug output, downgrading WARN messages to INFO in
non-critical cases, and printing the peer's address where relevant.
Switch NCCL_DEBUG_FILE to line buffering
* This should help avoid mixed-up partial output lines in multithreaded
cases.
Other minor fixes
* Improve the checks for buffer overflows in the graph code (Issue #1585).
* Extend logging and state clearing to all four events in the internal IB
plugin (Issue #1650).
* Fix the error path in case IB communication is not ready (Issue #1489).
* Add ECE logging for IB fabric.
* Fix various minor issues in the graph module (Issue #1635).
* Clean up the debug output in the graph code, downgrading WARN messages
to INFO in non-critical cases.
* Add a missing argument to a directSend() call (Issue #1628).
* Remove duplicate code in sendProxySetup() (Issue #1420).
* Fix the order of arguments of cudaDeviceCanAccessPeer() (Issue #1507).
* Fix compiler warnings with GCC 14.
* Fix a typo in a comment (Issue #1236).
[ROCm/rccl commit: 72d2432094]
Using a single slice rather than the typical two provides about 5% speedup (sometimes more or less) on some GFX9 systems for single node.
[ROCm/rccl commit: 2f6b20c00a]
* Detect if HSA_NO_SCRATCH_RECLAIM is set after initEnv()
For rocm older than 6.4, we need to set HSA_NO_SCRATCH_RECLAIM=1 to use LL128 protocol.
This Env is set outside of RCCL, add the logging to detect whether its set during runtime.
* check hip runtime ver via hipRuntimeGetVersion
* move the detection to ncclinit func
* correct rocm version integer
* update warning message
* avoid unnecessary info msg on hsa_no_scratch_reclaim detection
[ROCm/rccl commit: 51f87fbb43]
Wenkai Du
Dingming Wu
Bertan Dogancay
alex-breslow-amd
Arm Patinyasakdikul
Nilesh M Negi
Mustafa Abduljabbar
Surya Periaswamy
Artem Kuzmitckii
BertanDogancay
amd-jiali
Bhuvan Mital
ycui1984
Nusrat Islam
Avinash
Kamil Iskra
Dingming Wu