Power & Thermal Management

Understanding power management is essential for infrastructure engineers. Whether optimizing cloud costs, managing thermal throttling, or debugging performance issues, these concepts matter at scale.

Interview Frequency: Medium (important for infrastructure/performance roles)
Key Topics: cpufreq, cpuidle, thermal throttling, power governors
Time to Master: 6-8 hours

Why Power Management Matters

In the cloud era, power management directly impacts:

Cost: Power = money (cloud providers charge for compute)
Performance: Thermal throttling degrades performance
Reliability: Heat shortens hardware lifespan
Capacity planning: Power and cooling constraints limit density

Power Management Architecture

┌─────────────────────────────────────────────────────────────────────────────┐
│                    LINUX POWER MANAGEMENT STACK                              │
├─────────────────────────────────────────────────────────────────────────────┤
│                                                                              │
│  User Space                                                                  │
│  ┌─────────────────────────────────────────────────────────────────────────┐│
│  │  Tools: turbostat, cpupower, thermald, tuned                            ││
│  │  Interfaces: /sys/devices/system/cpu/, /sys/class/thermal/              ││
│  └─────────────────────────────────────────────────────────────────────────┘│
│                                                                              │
│  ════════════════════════════════════════════════════════════════════════   │
│                                                                              │
│  Kernel Frameworks                                                           │
│  ┌─────────────────────────────────────────────────────────────────────────┐│
│  │                                                                          ││
│  │  ┌─────────────┐  ┌─────────────┐  ┌─────────────┐  ┌─────────────┐    ││
│  │  │  cpufreq    │  │  cpuidle    │  │  thermal    │  │  PM QoS     │    ││
│  │  │ (frequency) │  │  (C-states) │  │  (cooling)  │  │  (latency)  │    ││
│  │  └──────┬──────┘  └──────┬──────┘  └──────┬──────┘  └──────┬──────┘    ││
│  │         │                │                │                │            ││
│  │         └────────────────┼────────────────┼────────────────┘            ││
│  │                          │                │                              ││
│  │                          ▼                ▼                              ││
│  │         ┌────────────────────────────────────────────────┐              ││
│  │         │           Power Management Core                 │              ││
│  │         │  - ACPI interface                               │              ││
│  │         │  - Intel/AMD specific drivers                   │              ││
│  │         │  - Device PM                                    │              ││
│  │         └────────────────────────────────────────────────┘              ││
│  │                                                                          ││
│  └─────────────────────────────────────────────────────────────────────────┘│
│                                                                              │
│  ════════════════════════════════════════════════════════════════════════   │
│                                                                              │
│  Hardware                                                                    │
│  ┌─────────────────────────────────────────────────────────────────────────┐│
│  │  CPU: P-states (frequency), C-states (idle), Turbo Boost               ││
│  │  Sensors: Temperature, power consumption, voltage                       ││
│  │  Cooling: Fans, passive cooling                                         ││
│  └─────────────────────────────────────────────────────────────────────────┘│
│                                                                              │
└─────────────────────────────────────────────────────────────────────────────┘

CPU Frequency Scaling (cpufreq)

P-States: Performance States

┌─────────────────────────────────────────────────────────────────────────────┐
│                    CPU P-STATES (Performance States)                         │
├─────────────────────────────────────────────────────────────────────────────┤
│                                                                              │
│  P-State    Frequency      Voltage       Power          Use Case            │
│  ────────────────────────────────────────────────────────────────────────   │
│  P0         3.6 GHz       1.2 V         ~100W          Max performance      │
│  P1         3.2 GHz       1.1 V         ~75W           High performance     │
│  P2         2.8 GHz       1.0 V         ~50W           Normal use           │
│  P3         2.4 GHz       0.9 V         ~35W           Power saving         │
│  P4         2.0 GHz       0.85V         ~25W           Low power            │
│  ...        ...           ...           ...            ...                   │
│  Pn         800 MHz       0.7 V         ~5W            Minimum              │
│                                                                              │
│  Turbo Boost (above base frequency):                                         │
│  • Single-core turbo: Up to 4.8 GHz                                         │
│  • All-core turbo: Up to 4.2 GHz                                            │
│  • Depends on: Temperature, power budget, active cores                      │
│                                                                              │
└─────────────────────────────────────────────────────────────────────────────┘

Viewing Frequency Information

# View available frequencies
cat /sys/devices/system/cpu/cpu0/cpufreq/scaling_available_frequencies
# 800000 1000000 1200000 1400000 1600000 1800000 2000000 2200000 2400000

# Current frequency
cat /sys/devices/system/cpu/cpu0/cpufreq/scaling_cur_freq
# 2400000 (2.4 GHz)

# Hardware limits
cat /sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_min_freq  # 800000
cat /sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_max_freq  # 3600000

# Current governor
cat /sys/devices/system/cpu/cpu0/cpufreq/scaling_governor
# schedutil

# Using cpupower tool
cpupower frequency-info
cpupower frequency-set -g performance  # Set governor

# Real-time monitoring with turbostat
sudo turbostat --interval 1

Frequency Governors

Governor	Description	Use Case
`performance`	Always max frequency	Latency-sensitive, benchmarks
`powersave`	Always min frequency	Battery saving, idle servers
`ondemand`	Scale with CPU load	General purpose (legacy)
`conservative`	Scale gradually	Smooth frequency transitions
`schedutil`	Scheduler-integrated	Modern default, best for most
`userspace`	Manual control	Custom tuning

Setting Frequency

# Set governor for all CPUs
for cpu in /sys/devices/system/cpu/cpu[0-9]*; do
    echo performance > $cpu/cpufreq/scaling_governor
done

# Using cpupower
cpupower frequency-set -g performance

# Set specific frequency (userspace governor)
cpupower frequency-set -g userspace
cpupower frequency-set -f 2.4GHz

# Set min/max frequencies
echo 2400000 > /sys/devices/system/cpu/cpu0/cpufreq/scaling_min_freq
echo 3600000 > /sys/devices/system/cpu/cpu0/cpufreq/scaling_max_freq

Intel P-State Driver

# Check if intel_pstate is active
cat /sys/devices/system/cpu/cpu0/cpufreq/scaling_driver
# intel_pstate

# Intel pstate specific controls
ls /sys/devices/system/cpu/intel_pstate/
# max_perf_pct  min_perf_pct  no_turbo  status  turbo_pct

# Disable turbo boost
echo 1 > /sys/devices/system/cpu/intel_pstate/no_turbo

# Set performance limits (percentage of max)
echo 100 > /sys/devices/system/cpu/intel_pstate/max_perf_pct
echo 50 > /sys/devices/system/cpu/intel_pstate/min_perf_pct

# Check turbo frequency range
cat /sys/devices/system/cpu/intel_pstate/turbo_pct

CPU Idle States (cpuidle)

C-States: Idle States

┌─────────────────────────────────────────────────────────────────────────────┐
│                    CPU C-STATES (Idle States)                                │
├─────────────────────────────────────────────────────────────────────────────┤
│                                                                              │
│  C-State    Name           Power        Exit Latency    Description          │
│  ────────────────────────────────────────────────────────────────────────   │
│  C0         Active         High         0 μs            CPU executing        │
│  C1         Halt           Low          1-2 μs          Clock gated         │
│  C1E        Enhanced Halt  Lower        2-5 μs          + voltage reduction │
│  C3         Sleep          Very Low     50-100 μs       L1/L2 flushed       │
│  C6         Deep Sleep     Minimal      100-200 μs      Core off            │
│  C7/C8/C9   Deeper         Ultra Low    200-500 μs      Package states      │
│                                                                              │
│  Trade-off:                                                                  │
│  • Deeper states = more power savings                                       │
│  • Deeper states = higher wake-up latency                                   │
│  • Must balance power vs. latency requirements                              │
│                                                                              │
│  For latency-sensitive apps: Limit to C1/C1E                                │
│  For power efficiency: Allow all C-states                                   │
│                                                                              │
└─────────────────────────────────────────────────────────────────────────────┘

Viewing C-State Information

# Available idle states
ls /sys/devices/system/cpu/cpu0/cpuidle/
# state0/ state1/ state2/ state3/

# State details
cat /sys/devices/system/cpu/cpu0/cpuidle/state2/name    # C6
cat /sys/devices/system/cpu/cpu0/cpuidle/state2/latency # 200
cat /sys/devices/system/cpu/cpu0/cpuidle/state2/usage   # 1234567
cat /sys/devices/system/cpu/cpu0/cpuidle/state2/time    # 98765432

# Current governor
cat /sys/devices/system/cpu/cpuidle/current_governor
# menu

# Using turbostat to see C-state residency
sudo turbostat --show Core,CPU,Busy%,Avg_MHz,C1%,C3%,C6%,C7% --interval 1

Limiting C-States for Low Latency

# Disable deeper C-states (disable state 2 and beyond)
for cpu in /sys/devices/system/cpu/cpu*/cpuidle/state[2-9]; do
    echo 1 > $cpu/disable
done

# Kernel parameter to limit C-states
# Add to GRUB_CMDLINE_LINUX:
intel_idle.max_cstate=1
processor.max_cstate=1

# Using PM QoS to set latency constraint
# Request max 10 microsecond latency
echo 10 > /dev/cpu_dma_latency  # Keeps file open

# Programmatically:
int fd = open("/dev/cpu_dma_latency", O_RDWR);
int latency = 10;  // microseconds
write(fd, &latency, sizeof(latency));
// Keep fd open as long as low latency needed

Thermal Management

Thermal Zones

# List thermal zones
ls /sys/class/thermal/thermal_zone*/
# thermal_zone0/ thermal_zone1/ ...

# View temperature (millidegrees Celsius)
cat /sys/class/thermal/thermal_zone0/temp
# 45000 (45°C)

# View zone type
cat /sys/class/thermal/thermal_zone0/type
# x86_pkg_temp

# View trip points
cat /sys/class/thermal/thermal_zone0/trip_point_0_temp  # 85000
cat /sys/class/thermal/thermal_zone0/trip_point_0_type  # passive
cat /sys/class/thermal/thermal_zone0/trip_point_1_temp  # 95000
cat /sys/class/thermal/thermal_zone0/trip_point_1_type  # critical

Thermal Trip Points

┌─────────────────────────────────────────────────────────────────────────────┐
│                    THERMAL TRIP POINTS                                       │
├─────────────────────────────────────────────────────────────────────────────┤
│                                                                              │
│  Temperature                                                                 │
│       │                                                                      │
│  95°C ┼───────────────────────────────────────────── CRITICAL               │
│       │                                              • Emergency shutdown   │
│       │                                              • Prevent damage       │
│       │                                                                      │
│  85°C ┼───────────────────────────────────────────── PASSIVE                │
│       │                                              • CPU throttling       │
│       │                                              • Reduce power/freq    │
│       │                                                                      │
│  75°C ┼───────────────────────────────────────────── ACTIVE                 │
│       │                                              • Increase fan speed   │
│       │                                                                      │
│       │                                                                      │
│  45°C ┼───────────────────────────────────────────── NORMAL                 │
│       │                                              • Normal operation     │
│       │                                                                      │
│       └─────────────────────────────────────────────────────────────────►   │
│                                                                              │
└─────────────────────────────────────────────────────────────────────────────┘

Cooling Devices

# List cooling devices
ls /sys/class/thermal/cooling_device*/

# View cooling device type
cat /sys/class/thermal/cooling_device0/type
# intel_powerclamp
# Processor
# Fan

# Current and max cooling state
cat /sys/class/thermal/cooling_device0/cur_state    # 0
cat /sys/class/thermal/cooling_device0/max_state    # 10

# Which zone is this device bound to?
cat /sys/class/thermal/thermal_zone0/cdev0/type

Monitoring Thermal Throttling

# Using turbostat
sudo turbostat --show Core,Busy%,Avg_MHz,PkgTmp,PkgWatt --interval 1

# Check for throttling
# MSR_THERM_STATUS (0x19C)
sudo rdmsr 0x19C
# Bit 0: Thermal status (currently throttling)
# Bit 1: Thermal status log (has throttled)

# perf for thermal events
perf stat -e power/energy-pkg/,power/energy-cores/ sleep 10

# Using sensors (lm-sensors package)
sensors
# coretemp-isa-0000
# Core 0:       +45.0°C  (high = +85.0°C, crit = +95.0°C)

Power Monitoring

RAPL (Running Average Power Limit)

# View power zones
ls /sys/class/powercap/intel-rapl/

# Package power
cat /sys/class/powercap/intel-rapl/intel-rapl:0/name           # package-0
cat /sys/class/powercap/intel-rapl/intel-rapl:0/energy_uj      # Cumulative energy

# DRAM power
cat /sys/class/powercap/intel-rapl/intel-rapl:0/intel-rapl:0:0/name  # dram

# Power constraints (limits)
cat /sys/class/powercap/intel-rapl/intel-rapl:0/constraint_0_power_limit_uw
cat /sys/class/powercap/intel-rapl/intel-rapl:0/constraint_0_time_window_us

# Using perf for power events
perf stat -e power/energy-pkg/,power/energy-cores/,power/energy-ram/ ./myapp

# Using turbostat
sudo turbostat --show PkgWatt,CorWatt,RAMWatt --interval 1

Power Capping

# Set power limit (microwatts)
echo 65000000 > /sys/class/powercap/intel-rapl/intel-rapl:0/constraint_0_power_limit_uw

# Set time window
echo 1000000 > /sys/class/powercap/intel-rapl/intel-rapl:0/constraint_0_time_window_us

Performance vs Power Trade-offs

Latency-Sensitive Workloads

# Maximum performance configuration
# /etc/tuned/profiles/latency-performance/

# 1. Set performance governor
cpupower frequency-set -g performance

# 2. Disable turbo (for consistent performance)
echo 1 > /sys/devices/system/cpu/intel_pstate/no_turbo

# 3. Disable deep C-states
for state in /sys/devices/system/cpu/cpu*/cpuidle/state[2-9]; do
    echo 1 > $state/disable
done

# 4. Disable frequency scaling (lock to max)
for cpu in /sys/devices/system/cpu/cpu*/cpufreq; do
    cat $cpu/cpuinfo_max_freq > $cpu/scaling_min_freq
done

# 5. Isolate CPUs for application
# GRUB: isolcpus=2-7 nohz_full=2-7

Power-Efficient Workloads

# Power saving configuration
# 1. Set schedutil governor (balances power/performance)
cpupower frequency-set -g schedutil

# 2. Enable all C-states
for state in /sys/devices/system/cpu/cpu*/cpuidle/state*; do
    echo 0 > $state/disable 2>/dev/null
done

# 3. Enable turbo (uses power when needed, saves when not)
echo 0 > /sys/devices/system/cpu/intel_pstate/no_turbo

# 4. Set power cap if needed
echo 35000000 > /sys/class/powercap/intel-rapl/intel-rapl:0/constraint_0_power_limit_uw

Using tuned Profiles

# Install tuned
yum install tuned  # or apt install tuned

# List available profiles
tuned-adm list

# Common profiles:
# balanced           - Default, balance power and performance
# throughput-performance - Maximum throughput
# latency-performance    - Minimum latency
# powersave             - Maximum power savings
# virtual-guest         - Optimized for VMs
# virtual-host          - Optimized for hypervisors

# Set profile
tuned-adm profile latency-performance

# Check current profile
tuned-adm active

Cloud and Container Considerations

Container Power Impact

# Containers don't directly manage power, but:
# 1. CPU limits affect power indirectly
# 2. Throttled containers waste cycles

# Check if container is being throttled
cat /sys/fs/cgroup/cpu/docker/<container>/cpu.stat
# nr_throttled 12345
# throttled_time 67890123456

# Power-aware scheduling (Kubernetes)
apiVersion: v1
kind: Pod
metadata:
  name: power-aware
spec:
  containers:
  - name: app
    resources:
      requests:
        cpu: "2"       # Request specific cores
      limits:
        cpu: "2"       # Limit to prevent throttling
    # Pin to specific NUMA node for efficiency
    nodeSelector:
      kubernetes.io/numa-node: "0"

Cloud Instance Types

┌─────────────────────────────────────────────────────────────────────────────┐
│                    CLOUD POWER CONSIDERATIONS                                │
├─────────────────────────────────────────────────────────────────────────────┤
│                                                                              │
│  Instance Type        Power Profile          Notes                          │
│  ────────────────────────────────────────────────────────────────────────   │
│  General Purpose      Balanced               C-states enabled, schedutil   │
│  (m5, n2)                                                                   │
│                                                                              │
│  Compute Optimized    Max Performance        Often performance governor    │
│  (c5, c2)             Higher power draw                                    │
│                                                                              │
│  Memory Optimized     Balanced               Large memory, moderate CPU    │
│  (r5, m2)                                                                   │
│                                                                              │
│  Burstable            Power efficient        Credits for bursts           │
│  (t3, e2-micro)       Deep C-states                                       │
│                                                                              │
│  High Frequency       Max single-thread      Turbo boost, high power      │
│  (z1d, c6i.metal)                                                          │
│                                                                              │
│  Spot/Preemptible     Variable               Reclaimed for capacity       │
│                                                                              │
│  Key: Match instance to workload power profile                              │
│                                                                              │
└─────────────────────────────────────────────────────────────────────────────┘

Debugging Power Issues

High CPU Temperature

# 1. Check current temperature
cat /sys/class/thermal/thermal_zone*/temp

# 2. Check for throttling
sudo turbostat --show Core,Busy%,Avg_MHz,PkgTmp,PkgWatt --interval 1

# 3. Check which process is causing load
top -o %CPU

# 4. Check cooling devices
for cooling in /sys/class/thermal/cooling_device*; do
    echo "$cooling: $(cat $cooling/type) state: $(cat $cooling/cur_state)/$(cat $cooling/max_state)"
done

# 5. Check for thermal trips in logs
dmesg | grep -i thermal

Unexpected Frequency Drops

# 1. Check if throttling due to temperature
cat /sys/devices/system/cpu/cpu0/thermal_throttle/core_throttle_count

# 2. Check power limits
cat /sys/class/powercap/intel-rapl/intel-rapl:0/constraint_0_power_limit_uw

# 3. Check governor
cat /sys/devices/system/cpu/cpu0/cpufreq/scaling_governor

# 4. Check if turbo is disabled
cat /sys/devices/system/cpu/intel_pstate/no_turbo

# 5. Monitor in real-time
sudo turbostat --show Core,CPU,Avg_MHz,Busy%,Bzy_MHz,TSC_MHz,PkgTmp --interval 1

High Power Consumption

# 1. Monitor power draw
sudo turbostat --show PkgWatt,CorWatt,RAMWatt --interval 1

# 2. Check which processes are using CPU
perf top

# 3. Check for inefficient C-state usage
cat /sys/devices/system/cpu/cpu*/cpuidle/state*/usage

# 4. Check for unnecessary activity
sudo perf record -g -a sleep 10
sudo perf report

Interview Questions

Q: How would you optimize a latency-sensitive application's power management?

Answer:For latency-sensitive workloads, minimize wake-up latency:

Disable deep C-states: Prevent 100+ μs wake latencies

intel_idle.max_cstate=1
# or disable via /sys/devices/system/cpu/cpu*/cpuidle/state*/disable

Lock CPU frequency: Prevent P-state transitions

cpupower frequency-set -g performance
# Set min freq = max freq

Consider disabling turbo: Turbo varies, causing jitter

echo 1 > /sys/devices/system/cpu/intel_pstate/no_turbo

Use PM QoS: Programmatically request low latency

int fd = open("/dev/cpu_dma_latency", O_RDWR);
int latency = 0;  // Request zero latency
write(fd, &latency, sizeof(latency));

Isolate CPUs: Prevent kernel interference

isolcpus=2-7 nohz_full=2-7

Trade-off: Higher power consumption, but consistent low latency.

Q: What causes thermal throttling and how do you detect it?

Answer:Causes:

High CPU utilization exceeding TDP
Inadequate cooling (failed fan, blocked vents)
High ambient temperature
Too many active cores at high frequency

Detection:

# Check temperature
cat /sys/class/thermal/thermal_zone0/temp  # millidegrees

# Check throttle count
cat /sys/devices/system/cpu/cpu0/thermal_throttle/package_throttle_count

# Real-time monitoring
sudo turbostat --show Core,CPU,Avg_MHz,PkgTmp --interval 1
# If Avg_MHz drops when temp is high → throttling

# Check dmesg for thermal warnings
dmesg | grep -i "thermal\|throttl"

Solutions:

Improve cooling (check fans, thermal paste)
Reduce power limit via RAPL
Spread work across more cores (lower per-core heat)
Check for runaway processes

Q: Explain the difference between P-states and C-states

Answer:P-states (Performance States):

Control frequency and voltage when CPU is active
Higher P-number = lower frequency = lower power
Transition time: ~10-100 μs
Managed by: cpufreq governors

C-states (Idle States):

Control power when CPU is idle
Higher C-number = deeper sleep = lower power
Deeper states take longer to wake from
Managed by: cpuidle governors

When each matters:

High load: P-states matter (CPU always active)
Low/variable load: C-states matter (CPU frequently idle)
Latency-sensitive: Limit both (avoid transitions)

Example: Web server handling requests

Between requests: CPU enters C-states (saves power)
During request: CPU at appropriate P-state
With C6 disabled: Faster wake-up, higher power draw

Summary

Mechanism	Purpose	Key Files
cpufreq	CPU frequency scaling	`/sys/devices/system/cpu/cpu*/cpufreq/`
cpuidle	Idle state management	`/sys/devices/system/cpu/cpu*/cpuidle/`
thermal	Temperature monitoring	`/sys/class/thermal/`
RAPL	Power limiting/monitoring	`/sys/class/powercap/intel-rapl/`
PM QoS	Latency constraints	`/dev/cpu_dma_latency`
tuned	Profile-based optimization	`/etc/tuned/profiles/`

Key Tools

# Monitor all at once
sudo turbostat --interval 1

# Set power/performance profile
tuned-adm profile <profile-name>

# CPU frequency control
cpupower frequency-set -g performance

# View temperatures
sensors

# Power statistics
perf stat -e power/energy-pkg/ ./myapp

Next Steps

Process Subsystem → - CPU scheduling and affinity
Tracing & Profiling → - Power analysis with perf
Interview Questions → - Performance debugging scenarios

Testing & Code Quality

Crash Courses

AI Engineering

Math for ML - Understanding Linear Algebra

Probability & Statistics for ML

Math for ML - Understanding Calculus

ML Mastery

Deep Learning Mastery

NestJS Mastery

Microservices Mastery

Low Level Design

OOP Concepts

SOLID Principles

Design Patterns

LLD Case Studies

System Design (HLD)

Senior Level (L5+/Staff)

HLD Case Studies

Engineering Fundamentals

DevOps & Operations

AWS Cloud

AWS Monitoring & Observability

AWS Security Services

AWS Serverless

AWS Operations

AWS Advanced

AWS Case Studies

DevOps Tools

Database Engineering

HIPAA Compliance Mastery

Operating Systems

Linux Internals

Distributed Systems

Networking Mastery

Build Your Own X

Go Lang Mastery

C Programming

Distributed System Tools

​Power & Thermal Management

​Why Power Management Matters

​Power Management Architecture

​CPU Frequency Scaling (cpufreq)

​P-States: Performance States

​Viewing Frequency Information

​Frequency Governors

​Setting Frequency

​Intel P-State Driver

​CPU Idle States (cpuidle)

​C-States: Idle States

​Viewing C-State Information

​Limiting C-States for Low Latency

​Thermal Management

​Thermal Zones

​Thermal Trip Points

​Cooling Devices

​Monitoring Thermal Throttling

​Power Monitoring

​RAPL (Running Average Power Limit)

​Power Capping

​Performance vs Power Trade-offs

​Latency-Sensitive Workloads

​Power-Efficient Workloads

​Using tuned Profiles

​Cloud and Container Considerations

​Container Power Impact

​Cloud Instance Types

​Debugging Power Issues

​High CPU Temperature

​Unexpected Frequency Drops

​High Power Consumption

​Interview Questions

​Summary

​Key Tools

​Next Steps

Power & Thermal Management

Why Power Management Matters

Power Management Architecture

CPU Frequency Scaling (cpufreq)

P-States: Performance States

Viewing Frequency Information

Frequency Governors

Setting Frequency

Intel P-State Driver

CPU Idle States (cpuidle)

C-States: Idle States

Viewing C-State Information

Limiting C-States for Low Latency

Thermal Management

Thermal Zones

Thermal Trip Points

Cooling Devices

Monitoring Thermal Throttling

Power Monitoring

RAPL (Running Average Power Limit)

Power Capping

Performance vs Power Trade-offs

Latency-Sensitive Workloads

Power-Efficient Workloads

Using tuned Profiles

Cloud and Container Considerations

Container Power Impact

Cloud Instance Types

Debugging Power Issues

High CPU Temperature

Unexpected Frequency Drops

High Power Consumption

Interview Questions

Summary

Key Tools

Next Steps