spmdBroadcast — Broadcast a value from one SPMD rank to every rank.
spmdBroadcast returns the designated source rank's value on every rank in the active gang.
Syntax
value = spmdBroadcast(source, value)Inputs
lab- Type
IntegerScalar- Required
- Yes
- Default
- —
- Description
- One-based lab index.
value- Type
Any- Required
- No
- Default
- —
- Description
- Optional value supplied by the designated lab.
Returns
result- Type
Any- Description
- Result produced by the parallel operation.
Errors
RunMat:parallel:LoweringRequired- When
- The operation is invoked without an active compiler-owned SPMD or distributed execution context.
- Message
parallel operation requires executor-aware lowering
SPMD execution
Collective and point-to-point functions run inside an active spmd region. Ranks are one-based within the admitted worker gang. Every rank follows the compiler-owned operation order, which keeps matching calls deterministic across local, browser, isolated-process, and remote execution.
The spmd* names are the current surface. The corresponding lab*, gplus, gcat, and gop names remain compatible spellings with the same execution and value semantics.
Values and messages
Messages use RunMat's canonical execution codec. Supported values retain their class, shape, container structure, and exact integer payload; a transfer does not route typed data through double.
Source and destination ranks must identify members of the active gang. Tags are exact nonnegative integers and default to zero when omitted. A receive may select a particular source and tag; spmdProbe and labProbe test the same selection without consuming a message.
Aggregate results
An aggregate without a destination returns the result to every rank. With a destination, the result is delivered to that rank and the other ranks receive the operation's documented empty result. Concatenation uses the requested one-based dimension, and functional reductions require an associative binary callable.
Failures and cancellation
A collective used outside compiler-owned SPMD execution returns a structured lowering error. Invalid ranks, tags, dimensions, reducers, mismatched operation order, worker failure, cancellation, and an unsatisfied communication dependency terminate the gang with a structured parallel error instead of waiting indefinitely.
Example
Broadcast a typed value from the first rank
pool = parpool(2);
spmd
shared = spmdBroadcast(1, uint64([spmdIndex() 0x0020000000000001u64]));
endExpected output:
both ranks receive the first rank's uint64 vectorUsing spmdBroadcast with coding agents
Open a RunMat example with live inputs, then ask the agent to explain how spmdBroadcast changes the result.
Run a small spmdBroadcast example, explain the result, then change one input and compare the output.
FAQ
Where can these functions run?⌄
Inside an active spmd region. They are lowered with the region's worker-gang identity and operation sequence.
Are ranks zero-based?⌄
No. Rank values are one-based within the active SPMD gang.
Do messages preserve integer classes?⌄
Yes. The execution codec preserves supported typed values, including exact fixed-width integer payloads.
What happens when destination is omitted from an aggregate?⌄
The aggregate result is returned to every rank.
Can an unmatched communication wait forever?⌄
No. RunMat detects an unsatisfied gang communication dependency and reports a structured parallel error.
Related Parallel functions
codistributed · codistributed.build · codistributor · codistributor1d · codistributor2dbc · distributed · fetchNext · fetchOutputs · gcat · gcp · getCodistributor · getCurrentJob · getCurrentTask · getCurrentWorker · getLocalPart · globalIndices · gop · gplus · iscodistributed · isComplete · labBarrier · labBroadcast · labindex · labProbe · labReceive · labSend · labSendReceive · numlabs · parfeval · parfevalOnAll · parpool · redistribute · spmdBarrier · spmdCat · spmdIndex · spmdPlus · spmdProbe · spmdReceive · spmdReduce · spmdSend · spmdSendReceive · spmdSize
Related guides
Source you can read
Every RunMat function's source is public. Read exactly how spmdBroadcast is executed, line by line, in Rust.
- View the source for spmdBroadcast in Rust on GitHub
- Learn how the RunMat runtime works
- Found a bug? Open an issue with a minimal reproduction.
About RunMat
RunMat is a runtime that executes MATLAB-syntax code fast on any GPU. The runtime is free for individuals, including freelancers; students and faculty at an academic institution, for teaching or research there; contributors to an open-source project, for work on that project; organizations with 250 or fewer people and US$10M or less in annual revenue; and anyone, for non-production use. See runmat.com/license.
- RunMat automatically optimizes your math for GPU execution on Apple, Nvidia, and AMD hardware. No code changes needed. Simulations that took hours now take minutes.
- Start running code in seconds. RunMat runs in the browser, on the desktop, or from the CLI. No license server, no IT ticket.