rocm-smi --showproductname will not show "Card series" in its output if
product_name exported by Kernel is empty string. This has been raised a
regression by customer.
BUG: SWDEV-297228
Signed-off-by: Harish Kasiviswanathan <Harish.Kasiviswanathan@amd.com>
Change-Id: I9aae24778e2d3a30aa661d8f338278c1666590fb
During the tail end when process is terminating, subprocess module fails
to find the process. This results in extraneous printing of a line with
char 'b'. Fix this.
BUG: SWDEV-296409
Signed-off-by: Harish Kasiviswanathan <Harish.Kasiviswanathan@amd.com>
Change-Id: I39aacf8ae948a5acec0aa93296cc0e0aec88b3ef
Python's default 'print' implementation is not thread safe, causing
empty lines to be printed during multithreaded code execution.
This fixes the --showevents output for multi-GPU systems.
Signed-off-by: Ori Messinger <Ori.Messinger@amd.com>
Change-Id: I72f7341cdf4401f1fed4cd8f7d7a4a90bf9a3a4c
Use zero padding for the hexadecimal value 'device_model' inside
showProductName with a padding length of 4.
Signed-off-by: Ori Messinger <Ori.Messinger@amd.com>
Change-Id: I962b94d414c6ba050d951486ad9e7559123f8850
Specify that timestamp resolution is in ns in header file.
Signed-off-by: Harish Kasiviswanathan <Harish.Kasiviswanathan@amd.com>
Change-Id: I4db00a07c0b5c43ae23c98213f2fbbcf93110234
gpu_metrics version 1.2 provides atomic timestamp. Use this timestamp.
Signed-off-by: Harish Kasiviswanathan <Harish.Kasiviswanathan@amd.com>
Change-Id: I7a1a675f53b93718f34b1f2979173e9064e0ef93
Chnage to RSMI_GPU_METRICS_API_CONTENT_VER_1. In preparation for
supporting additional formats
Signed-off-by: Harish Kasiviswanathan <Harish.Kasiviswanathan@amd.com>
Change-Id: I4367a2622a0fa41e6b05bc4436ecd24b8c4e30e2
Implement default GPU power cap functionality in rsmitst.
It is available in the "rsmitstReadOnly.TestPowerRead" test, and
is displayed as: "Default Power Cap: #uW" (where uW is microwatts).
Signed-off-by: Ori Messinger <Ori.Messinger@amd.com>
Change-Id: I564ea3785f1a93dfd30587634057516549fa762c
Since device is a list, we need to pass a single item to the isAmdGpu
function.
Fixes: c7c2ac5559 "rocm_smi.py: Don't try to reset non-AMD GPUs"
Signed-off-by: Kent Russell <kent.russell@amd.com>
Change-Id: I19a74377636ff4589f11d092f41e1d35c1acb307
Instead of throwing "Unsupported clock" errors for ASICs that don't
support a certain clock type (e.g. dcefclk on MI-series), just dump the
warning to logging.debug and don't try to read the clock
Signed-off-by: Kent Russell <kent.russell@amd.com>
Change-Id: If3cb9a472b03aa535a76fc24bcd9f77122090634
Use default power cap exposed via sysfs to determine when to
show 'Out of Spec" warning.
Signed-off-by: Ori Messinger <Ori.Messinger@amd.com>
Change-Id: I0fa3612b50e230856b0d5a390f876b35268d9587
Implement default GPU power cap functionality in the LIB.
Signed-off-by: Ori Messinger <Ori.Messinger@amd.com>
Change-Id: Ia6b3420beb0e4df5559c3e6d11d0667972590b53
Implement showevent functionality in the ROCm SMI Python CLI.
It can be called using --showevents with any combination of:
VM_FAULT, THERMAL_THROTTLE, and/or GPU_RESET
For example:
./rocm-smi --showevents VM_FAULT, THERMAL_THROTTLE, GPU_RESET
Signed-off-by: Ori Messinger <Ori.Messinger@amd.com>
Change-Id: I905fd9c949e91423b79833a04ab89d6ba3760e62
Many data center cards are fanless. Don't show warning if unable to get
fan speed. The fan speed will be reported as 0
Signed-off-by: Harish Kasiviswanathan <Harish.Kasiviswanathan@amd.com>
Change-Id: I53efe67ac88fb0824cf4820430b46c18bc7692df
This won't work for obvious reasons, so exit with an error instead of
trying to access a file that doesn't exist and segfaulting
Change-Id: Id1230922fa6e9a19e9394280faad88a43c7d2e34
The rsmi_dev_temp_metric_get() can also support the HBM
temperatures which is retrieved from gpu_metrics.
Change-Id: I96b979296e90cf881523627b41b1a02849676416
Previously, RSMI assumed that the event counter values returned
from perf were only new events. But in fact, when we read the
counter values, they are running totals. To account for this, we
now record the value we read and take the difference between the
current value and the previously recorded value.
Change-Id: I1e04b514e89c7c4d4719889f2dae3a1283864e7f
rocm_smi.py --set<m|s>clk was treating the freq as a string.
This causes problems in parsing when the index is more than 1
digit. Now, treat the indexes as integers.
Change-Id: Ia0d859d33b685fe90689a86ff1c83980808b1514
Previously, during the rsmi_init discovery process, the existence
of an hwmon# directory was used to distinguish between gpus nodes
and non-gpu nodes. This isn't reliable in some scenarios. Instead,
the existence of the vbios_version file is used as an
indicator that the node is indeed a gpu.
Change-Id: Icfbe5c42ed0970077b05f25c3d209308a31bec85
The purpose of this patch is to fix a power cap bug for --setpoweroverdrive.
This bug occurs when the user attempts to set a lower wattage than the current
or default wattage, which displays an unnecessary warning message.
Signed-off-by: Ori Messinger <Ori.Messinger@amd.com>
Change-Id: I730d2c6031b7d7c4af5acf32ecd28da5ca21ab12
Elena Sakhnovitch
Harish Kasiviswanathan
Bill(Shuzhou) Liu
Icarus Sparry
Ori Messinger
Divya Shikre
Kent Russell
Cole Nelson
Chris Freehill