nvidia-cuda-proxy-control(1) NVIDIA nvidia-cuda-proxy-control(1)NAME
nvidia-cuda-proxy-control - NVIDIA CUDA proxy management program
SYNOPSIS
nvidia-cuda-proxy-control [-d]
DESCRIPTION
CUDA proxy is a feature that allows multiple CUDA processes to share a
single GPU context. A CUDA program runs in proxy mode if the proxy
control daemon is running on the system. When a CUDA program starts,
it tries to connect to the daemon, which will then create a proxy
server process for the connecting client if one does not exist for the
user (UID) who launched the client. Each user (UID) has its own proxy
server process. The proxy server creates the shared GPU context, man‐
ages its clients, and issues work to the GPU on behalf of its clients.
The proxy mode should be transparent to CUDA programs.
Currently, proxy is available on 64-bit Linux only, and requires a
device that supports Unified Virtual Address (UVA). Applications
requiring pre-CUDA 4.0 APIs are not supported under proxy mode.
OPTIONS-d
Start the proxy control daemon, assuming the user has enough privilege
(e.g. root).
<no arguments>
Start the front-end management user interface to the daemon, which
needs to be started first. The front-end UI keeps reading commands from
stdin until EOF. Commands are separated by the newline character. If
an invalid command is issued and rejected, an error message will be
printed to stdout. The exit status of the front-end UI is zero if com‐
munication with the daemon is successful. A non-zero value is returned
if the daemon is not found or connection to the daemon is broken unex‐
pectedly. See the "quit" command below for more information about the
exit status.
Commands supported by the proxy control daemon:
get_server_list
Print out a list of PIDs of all proxy server processes.
start_server -uid UID
Start a new proxy server process for the specified user (UID).
shutdown_server PID [-f]
Shutdown the proxy server with given PID. The proxy server will
not accept any new client connections and it exits when all cur‐
rent clients disconnect. -f is forced immediate shutdown. If a
client launches a faulty kernel that runs forever, a forced
shutdown of the proxy server may be required, since the proxy
server creates and issues GPU work on behalf of its clients.
get_client_list PID
Print out a list of PIDs of all clients connected to the server
with given PID.
quit [-t TIMEOUT]
Shutdown the proxy control daemon process and all proxy servers.
The proxy control daemon stops accepting new clients while wait‐
ing for current proxy servers and proxy clients to finish. If
TIMEOUT is specified (in seconds), the daemon will force proxy
servers to shutdown if they are still running after TIMEOUT sec‐
onds.
This command is synchronous. The front-end UI waits for the dae‐
mon to shutdown, then returns the daemon's exit status. The exit
status is zero iff all proxy servers have exited gracefully.
ENVIRONMENT
CUDA_PROXY_PIPE_DIRECTORY
Specify the directory that contains the named pipes used for
communication among proxy control, proxy server, and proxy
clients. The value of this environment variable should be con‐
sistent in the proxy control daemon and all proxy client pro‐
cesses. Default directory is /tmp/nvidia-proxy
CUDA_PROXY_LOG_DIRECTORY
Specify the directory that contains the proxy log files. This
variable is used by the proxy control daemon only. Default
directory is /var/log/nvidia-proxy
FILES
Log files created by the proxy control daemon in the specified direc‐
tory
control.log
Record startup and shutdown of proxy control daemon, user com‐
mands issued with their results, and status of proxy servers.
server.log
Record startup and shutdown of proxy servers, and status of
proxy clients.
nvidia-cuda-proxy-control 2012-04-10 nvidia-cuda-proxy-control(1)
