fi_rxm(7) Libfabric Programmer's Manual
fi_rxm - The RxM (RDM over MSG) Utility Provider
The RxM provider (ofi_rxm) is an utility provider that supports FI_EP_RDM type endpoint emulated over FI_EP_MSG type endpoint(s) of an underlying core provider. FI_EP_RDM endpoints have a reliable unconnected messaging interface and RxM emulates this by hiding the connection management of underlying FI_EP_MSG endpoints from the user. Additionally, RxM can hide memory registration requirement from a core provider like verbs if the apps don’t support it.
Requirements for core provider
RxM provider requires the core provider to support the following features:
MSG endpoints (FI_EP_MSG)
RMA read/write (FI_RMA) - Used for implementing rendezvous protocol for large messages.
FI_OPT_CM_DATA_SIZE of at least 48 bytes
Requirements for applications
Since RxM emulates RDM endpoints by hiding connection management and connections are established only on-demand (when app tries to send data), the first several data transfer calls would return EAGAIN. Applications should be aware of this and retry until the operation succeeds.
If an application has chosen manual progress for data progress, it should also read the CQ so that the connection establishment progresses. Not doing so would result in a stall. See also the ERRORS section in fi_msg(3).
The RxM provider currently supports FI_MSG, FI_TAGGED and FI_RMA capabilities.
- Endpoint types
- The provider supports only FI_EP_RDM.
- Endpoint capabilities
- The following data transfer interface is supported: FI_MSG, FI_TAGGED, FI_RMA.
- The RxM provider supports both FI_PROGRESS_MANUAL and FI_PROGRESS_AUTO. The former is more optimal.
- Addressing Formats
- FI_SOCKADDR, FI_SOCKADDR_IN
- Memory Region
- FI_MR_VIRT_ADDR, FI_MR_ALLOCATED, FI_MR_PROV_KEY MR mode bits would be required from the app in case the core provider requires it.
When using RxM provider, some limitations from the underlying MSG provider could also show up. Please refer to the corresponding MSG provider man pages to find about those limitations.
RxM provider does not support the following features:
Application error data buffers
Reporting unknown source addr data as part of completions
When sending large messages, an app doing an sread or waiting on the CQ file descriptor may not get a completion when reading the CQ after being woken up from the wait. The app has to do sread or wait on the file descriptor again. This is needed because RxM uses a rendezvous protocol for large message sends. An app would get woken up from waiting on CQ fd when rendezvous protocol request completes but it would have to wait again to get an ACK from the receiver indicating completion of large message transfer by remote RMA read.
The ofi_rxm provider checks for the following environment variables.
- Defines the transmit buffer size / inject size. Messages of size less than this would be transmitted via an eager protocol and those above would be transmitted via a rendezvous or SAR (Segmentation And Reassembly) protocol. Transmit data would be copied up to this size (default: ~16k).
- Defines the maximum number of MSG provider CQ entries (default: 1) that would be read per progress (RxM CQ read).
- Set this environment variable to control the RxM SAR (Segmentation And Reassembly) protocol. Messages of size greater than this (default: 256 Kb) would be transmitted via rendezvous protocol.
- Set this to 1 to use shared receive context from MSG provider. This reduces overall memory usage but there may be a slight increase in latency (default: 0).
- Defines default TX context size (default: 1024)
- Defines default RX context size (default: 1024)
- Defines FI_EP_MSG TX size that would be requested (default: 128).
- Defines FI_EP_MSG RX size that would be requested (default: 128).
- Defines the expected number of ranks / peers an endpoint would communicate with (default: 256).
To optimize for bandwidth, ensure you use higher values than default for FI_OFI_RXM_TX_SIZE, FI_OFI_RXM_RX_SIZE, FI_OFI_RXM_MSG_TX_SIZE, FI_OFI_RXM_MSG_RX_SIZE subject to memory limits of the system and the tx and rx sizes supported by the MSG provider.
FI_OFI_RXM_SAR_LIMIT is another knob that can be experimented with to optimze for bandwidth.
To conserve memory, ensure FI_UNIVERSE_SIZE set to what is required. Similarly check that FI_OFI_RXM_TX_SIZE, FI_OFI_RXM_RX_SIZE, FI_OFI_RXM_MSG_TX_SIZE and FI_OFI_RXM_MSG_RX_SIZE env variables are set to only required values.