This resolves missing GPU names due to deviceID only being computed for
AMD gpus. It also filters out display adapters so there is less noise in
the logs when scanning cards.
This introduces support to overlay FCAT markers on top of other HUD elements.
For more information about nVidia's FCAT tool see: https://nvidia.com/en-us/geforce/technologies/fcat/technology
This patch includes support for the existing parameter infrastructure and it is fully dynamic.
Squashed commit:
Refactor FCAT to use existing parameter handling infrastructure
With this patch the FCAT overlay should be hooked up into the normal
parameter handling infrastructure. This includes support for
configuration file options, which are also part of this commit.
Most of the API does not modify it, so annotate it as const.
This makes it earlier to reason what's happening.
Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
Above all, we really don't need the gpu_busy_percent node, if the GPU
exposes a gpu_metrics node,
Although looking closer, the gpu_busy_percent check is meant for
something else - to distinguish between the card node (cardX) and the
card output node (cardX-output-foo).
To top it all up, the check at the very end implies that we can get a
case where gpu_metrics and gpu_busy_perfect is missing ... that's not
possible.
So instead, drop the early gpu_busy_perfect check and properly mandate
it later on.
Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
Currently we open the standalone busy, temp, gpu/memory clock and
power_usage nodes, even if gpu_metrics is present.
At the same time, we correctly ignore them when doing the read-only.
So just avoid opening them all together.
Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
Currently we can get the load and temp stats either from the standalone
nodes or from the gpu_metrics (binary) sysfs node.
Fix the next to handle that.
Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
Currently we pipeline a bunch of commands alike cat | grep | sed, there
we can do all that job with a single sed invocation - use that.
Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
Currently we readout the data as string only to convert it to hex. In
the error case, we leak the fd for the given nodes.
Read the data as hex directly, error out if that fails and avoid the
leaks.
Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
Using a vector with erase(begin()) results in a copy of the full array
every time. This is inefficient and in the case of HUD graphs, was having
quadratic complexity which can noticeably affect performance.
This patch replaces most of the operations with an O(1) alternative,
keeping the total cost under a linear bound. While on this, also refactor
the size of graph array into a constant.
Calling std::unique_lock in update() will cause a delay if
update_hw_info() takes longer than the period when update() is called.
To fix this, a try_to_lock tag is added. update() will only update if
the lock is aquired, else the update call is skipped.
Signed-off-by: Stephan Lachnit <stephanlachnit@debian.org>
FlightlessMango
Emil Velikov
Alessandro Toia
Kurt Kartaltepe
jackun
PMunkes
Emil Velikov
flightlessmango
Joshua Ashton
Tatsuyuki Ishi
Stephan Lachnit