From d86686fffda05d558a4c103ca835bb3da231b7a3 Mon Sep 17 00:00:00 2001
From: Daniel Koepping <elkoled@gmail.com>
Date: Wed, 5 Aug 2026 07:21:58 -0700
Subject: [PATCH] log chestnut metrics (#38527)

* log chestnut state

* chestnut metrics

* more metrics

* fix chestnut metrics

* cache chestnut power limit

* use valid

* move to class
---
 openpilot/cereal/log.capnp           | 11 +++++++
 openpilot/cereal/services.py         |  1 +
 openpilot/selfdrive/modeld/modeld.py | 44 +++++++++++++++++++++++++++-
 3 files changed, 55 insertions(+), 1 deletion(-)

diff --git a/openpilot/cereal/log.capnp b/openpilot/cereal/log.capnp
index 10b968b039..9c04599843 100644
--- a/openpilot/cereal/log.capnp
+++ b/openpilot/cereal/log.capnp
@@ -707,6 +707,16 @@ struct UsbState {
   }
 }
 
+struct ChestnutState {
+  tempC @0 :Float32;
+  memoryTempC @1 :Float32;
+  powerDrawW @2 :Float32;
+  powerLimitW @3 :Float32;
+  gpuUsagePercent @4 :UInt8;
+  gpuClockMhz @5 :UInt16;
+  fanSpeedRpm @6 :UInt16;
+}
+
 struct RadarState @0x9a185389d6fdd05f {
   mdMonoTime @6 :UInt64;  # for debugging
   radarErrors @13 :Car.RadarData.Error;
@@ -2570,6 +2580,7 @@ struct Event {
     procLog @33 :ProcLog;
     clocks @35 :Clocks;
     deviceState @6 :DeviceState;
+    chestnutState @152 :ChestnutState;
     logMessage @18 :Text;
     errorLogMessage @85 :Text;
 
diff --git a/openpilot/cereal/services.py b/openpilot/cereal/services.py
index 42669688b0..a41a890f97 100755
--- a/openpilot/cereal/services.py
+++ b/openpilot/cereal/services.py
@@ -25,6 +25,7 @@ _services: dict[str, tuple] = {
   "accelerometer": (True, 104., 104),
   "temperatureSensor": (True, 2., 200),
   "deviceState": (True, 2., 1),
+  "chestnutState": (True, 0.1, 1),
   "touch": (True, 20., 1),
   "can": (True, 100., 2053, QueueSize.BIG),  # decimation gives ~3 msgs in a full segment
   "controlsState": (True, 100., 10, QueueSize.MEDIUM),
diff --git a/openpilot/selfdrive/modeld/modeld.py b/openpilot/selfdrive/modeld/modeld.py
index d7ddce60b9..8b2bb40d12 100755
--- a/openpilot/selfdrive/modeld/modeld.py
+++ b/openpilot/selfdrive/modeld/modeld.py
@@ -1,7 +1,9 @@
 #!/usr/bin/env python3
+from functools import cached_property
 import os
 os.environ['GMMU'] = '0' # for usbgpu fast loading, noop for qcom
 from tinygrad.tensor import Tensor
+from tinygrad.device import Device
 import threading
 import time
 import numpy as np
@@ -9,6 +11,7 @@ import openpilot.cereal.messaging as messaging
 from openpilot.cereal import log
 from opendbc.car.structs import car
 from openpilot.cereal.messaging import PubMaster, SubMaster
+from openpilot.cereal.services import SERVICE_LIST
 from msgq.visionipc import VisionIpcClient, VisionStreamType, VisionBuf
 from opendbc.car.car_helpers import get_demo_car_params
 from openpilot.common.swaglog import cloudlog
@@ -64,6 +67,40 @@ def get_action_from_model(model_output: dict[str, np.ndarray], prev_action: log.
                                 shouldStop=bool(stop))
 
 
+class ChestnutState:
+  # only modeld can access chestnut
+  def __init__(self, pm: PubMaster):
+    self.pm = pm
+    self.valid = True
+
+  @cached_property
+  def power_limit(self) -> int:
+    smu = Device["AMD"].iface.dev_impl.smu
+    return smu._send_msg(smu.smu_mod.PPSMC_MSG_GetPptLimit, 0, read_back_arg=True)
+
+  def send(self) -> None:
+    msg = messaging.new_message('chestnutState')
+    state = msg.chestnutState
+    try:
+      smu = Device["AMD"].iface.dev_impl.smu
+      metrics = smu.read_table(smu.smu_mod.SmuMetricsExternal_t, smu.smu_mod.TABLE_SMU_METRICS).SmuMetrics
+      state.tempC = metrics.AvgTemperature[smu.smu_mod.TEMP_HOTSPOT]
+      state.memoryTempC = metrics.AvgTemperature[smu.smu_mod.TEMP_MEM]
+      state.powerDrawW = metrics.AverageSocketPower
+      state.powerLimitW = self.power_limit
+      state.gpuUsagePercent = metrics.AverageGfxActivity
+      state.gpuClockMhz = metrics.AverageGfxclkFrequencyPostDs
+      state.fanSpeedRpm = metrics.AvgFanRpm
+      self.valid = True
+    except Exception:
+      if self.valid:
+        cloudlog.exception("chestnut state read failed")
+      self.valid = False
+
+    msg.valid = self.valid
+    self.pm.send('chestnutState', msg)
+
+
 class FrameMeta:
   frame_id: int = 0
   timestamp_sof: int = 0
@@ -204,11 +241,13 @@ def main(demo=False):
   cloudlog.warning(f"models loaded in {time.monotonic() - st:.1f}s, modeld starting")
 
   # messaging
-  pm = PubMaster(["modelV2", "drivingModelData", "cameraOdometry"])
+  pub_socks = ["modelV2", "drivingModelData", "cameraOdometry"] + (["chestnutState"] if USBGPU else [])
+  pm = PubMaster(pub_socks)
   sm = SubMaster(["deviceState", "carState", "roadCameraState", "liveCalibration", "driverMonitoringState", "carControl", "liveDelay"])
 
   publish_state = PublishState()
   params = Params()
+  chestnut_state = ChestnutState(pm) if USBGPU else None
 
   # setup filter to track dropped frames
   frame_dropped_filter = FirstOrderFilter(0., 10., 1. / ModelConstants.MODEL_RUN_FREQ)
@@ -353,6 +392,9 @@ def main(demo=False):
       pm.send('cameraOdometry', posenet_send)
     last_vipc_frame_id = meta_main.frame_id
 
+    if chestnut_state is not None and run_count % round(ModelConstants.MODEL_RUN_FREQ / SERVICE_LIST['chestnutState'].frequency) == 0:
+      chestnut_state.send()
+
 
 if __name__ == "__main__":
   try: