8.2. PSCI Performance Measurements on Arm Juno Development Platform

This document summarises the findings of performance measurements of key operations in the Trusted Firmware-A Power State Coordination Interface (PSCI) implementation, using the in-built Performance Measurement Framework (PMF) and runtime instrumentation timestamps.

8.2.1. Method

We used the Juno R1 platform for these tests, which has 4 x Cortex-A53 and 2 x Cortex-A57 clusters running at the following frequencies:

Domain

Frequency (MHz)

Cortex-A57

900 (nominal)

Cortex-A53

650 (underdrive)

AXI subsystem

533

Juno supports CPU, cluster and system power down states, corresponding to power levels 0, 1 and 2 respectively. It does not support any retention states.

Given that runtime instrumentation using PMF is invasive, there is a small (unquantified) overhead on the results. PMF uses the generic counter for timestamps, which runs at 50MHz on Juno.

The following source trees and binaries were used:

Please see the Runtime Instrumentation Testing Methodology page for more details.

8.2.2. Procedure

  1. Build TFTF with runtime instrumentation enabled:

    make CROSS_COMPILE=aarch64-none-elf- PLAT=juno \
        TESTS=runtime-instrumentation all
    
  2. Fetch Juno’s SCP binary from TF-A’s archive:

    curl --fail --connect-timeout 5 --retry 5 -sLS -o scp_bl2.bin \
        https://downloads.trustedfirmware.org/tf-a/css_scp_2.12.0/juno/release/juno-bl2.bin
    
  3. Build TF-A with the following build options:

    make CROSS_COMPILE=aarch64-none-elf- PLAT=juno \
        BL33="/path/to/tftf.bin" SCP_BL2="scp_bl2.bin" \
        ENABLE_RUNTIME_INSTRUMENTATION=1 fiptool all fip
    
  4. Load the following images onto the development board: fip.bin, scp_bl2.bin.

8.2.3. Results

8.2.3.1. CPU_SUSPEND to deepest power level

CPU_SUSPEND latencies (µs) to deepest power level in parallel (v2.9)

Cluster

Core

Powerdown

Wakeup

Cache Flush

0

0

104.58

241.20

5.26

0

1

384.24

22.50

138.76

1

0

244.56

22.18

5.16

1

1

670.56

18.58

4.44

1

2

809.36

269.28

4.44

1

3

984.96

219.70

79.62

CPU_SUSPEND latencies (µs) to deepest power level in parallel (v2.10)

Cluster

Core

Powerdown

Wakeup

Cache Flush

0

0

242.66 (+132.03%)

245.1

5.4

0

1

522.08 (+35.87%)

26.24

138.32

1

0

104.36 (-57.33%)

27.1

5.32

1

1

382.56 (-42.95%)

23.34

4.42

1

2

807.74

271.54

4.64

1

3

981.36

221.8

79.48

CPU_SUSPEND latencies (µs) to deepest power level in serial (v2.9)

Cluster

Core

Powerdown

Wakeup

Cache Flush

0

0

236.56

23.24

138.18

0

1

236.86

23.28

138.10

1

0

281.04

22.80

77.24

1

1

100.28

18.52

4.54

1

2

100.12

18.78

4.50

1

3

100.36

18.94

4.44

CPU_SUSPEND latencies (µs) to deepest power level in serial (v2.10)

Cluster

Core

Powerdown

Wakeup

Cache Flush

0

0

236.84

27.1

138.36

0

1

236.96

27.1

138.32

1

0

280.06

26.94

77.5

1

1

100.76

23.42

4.36

1

2

100.02

23.42

4.44

1

3

100.08

23.2

4.4

8.2.3.2. CPU_SUSPEND to power level 0

CPU_SUSPEND latencies (µs) to power level 0 in parallel (v2.9)

Cluster

Core

Powerdown

Wakeup

Cache Flush

0

0

662.34

15.22

8.08

0

1

802.00

15.50

8.16

1

0

385.22

15.74

7.88

1

1

106.16

16.06

7.44

1

2

524.38

15.64

7.34

1

3

246.00

15.78

7.72

CPU_SUSPEND latencies (µs) to power level 0 in parallel (v2.10)

Cluster

Core

Powerdown

Wakeup

Cache Flush

0

0

801.04

18.66

8.22

0

1

661.28

19.08

7.88

1

0

105.9 (-72.51%)

20.3

7.58

1

1

383.58 (+261.32%)

20.4

7.42

1

2

523.52

20.1

7.74

1

3

244.5

20.16

7.56

CPU_SUSPEND latencies (µs) to power level 0 in serial (v2.9)

Cluster

Core

Powerdown

Wakeup

Cache Flush

0

0

99.80

15.94

5.42

0

1

99.76

15.80

5.24

1

0

278.26

16.16

4.58

1

1

96.88

16.00

4.52

1

2

96.80

16.12

4.54

1

3

96.88

16.12

4.54

CPU_SUSPEND latencies (µs) to power level 0 in serial (v2.10)

Cluster

Core

Powerdown

Wakeup

Cache Flush

0

0

99.84

18.86

5.54

0

1

100.2

18.82

5.66

1

0

278.12

20.56

4.48

1

1

96.68

20.62

4.3

1

2

96.94

20.14

4.42

1

3

96.68

20.46

4.32

8.2.3.3. CPU_OFF on all non-lead CPUs

CPU_OFF on all non-lead CPUs in sequence then, CPU_SUSPEND on the lead core to the deepest power level.

CPU_OFF latencies (µs) on all non-lead CPUs (v2.9)

Cluster

Core

Powerdown

Wakeup

Cache Flush

0

0

235.76

26.14

137.80

0

1

235.40

25.72

137.62

1

0

174.70

22.40

77.26

1

1

100.92

24.04

4.52

1

2

100.68

22.44

4.36

1

3

101.36

22.70

4.52

CPU_OFF latencies (µs) on all non-lead CPUs (v2.10)

test_rt_instr_cpu_off_serial (latest)

Cluster

Core

Powerdown

Wakeup

Cache Flush

0

0

236.04

30.02

137.9

0

1

235.38

29.7

137.72

1

0

175.18

26.96

77.26

1

1

100.56

28.34

4.32

1

2

100.38

26.82

4.3

1

3

100.86

26.98

4.42

8.2.3.4. CPU_VERSION in parallel

CPU_VERSION latency (µs) in parallel on all cores (2.9)

Cluster

Core

Latency

0

0

1.48

0

1

1.04

1

0

0.56

1

1

0.92

1

2

0.96

1

3

0.96

CPU_VERSION latency (µs) in parallel on all cores (2.10)

Cluster

Core

Latency

0

0

1.1 (-25.68%)

0

1

1.06

1

0

0.58

1

1

0.88

1

2

0.92

1

3

0.9

8.2.4. Annotated Historic Results

The following results are based on the upstream TF master as of 31/01/2017. TF-A was built using the same build instructions as detailed in the procedure above.

In the results below, CPUs 0-3 refer to CPUs in the little cluster (A53) and CPUs 4-5 refer to CPUs in the big cluster (A57). In all cases CPU 4 is the lead CPU.

PSCI_ENTRY corresponds to the powerdown latency, PSCI_EXIT the wakeup latency, and CFLUSH_OVERHEAD the latency of the cache flush operation.

8.2.4.1. CPU_SUSPEND to deepest power level on all CPUs in parallel

CPU

PSCI_ENTRY (us)

PSCI_EXIT (us)

CFLUSH_OVERHEAD (us)

0

27

20

5

1

114

86

5

2

202

58

5

3

375

29

94

4

20

22

6

5

290

18

206

A large variance in PSCI_ENTRY and PSCI_EXIT times across CPUs is observed due to TF PSCI lock contention. In the worst case, CPU 3 has to wait for the 3 other CPUs in the cluster (0-2) to complete PSCI_ENTRY and release the lock before proceeding.

The CFLUSH_OVERHEAD times for CPUs 3 and 5 are higher because they are the last CPUs in their respective clusters to power down, therefore both the L1 and L2 caches are flushed.

The CFLUSH_OVERHEAD time for CPU 5 is a lot larger than that for CPU 3 because the L2 cache size for the big cluster is lot larger (2MB) compared to the little cluster (1MB).

8.2.4.2. CPU_SUSPEND to power level 0 on all CPUs in parallel

CPU

PSCI_ENTRY (us)

PSCI_EXIT (us)

CFLUSH_OVERHEAD (us)

0

116

14

8

1

204

14

8

2

287

13

8

3

376

13

9

4

29

15

7

5

21

15

8

There is no lock contention in TF generic code at power level 0 but the large variance in PSCI_ENTRY times across CPUs is due to lock contention in Juno platform code. The platform lock is used to mediate access to a single SCP communication channel. This is compounded by the SCP firmware waiting for each AP CPU to enter WFI before making the channel available to other CPUs, which effectively serializes the SCP power down commands from all CPUs.

On platforms with a more efficient CPU power down mechanism, it should be possible to make the PSCI_ENTRY times smaller and consistent.

The PSCI_EXIT times are consistent across all CPUs because TF does not require locks at power level 0.

The CFLUSH_OVERHEAD times for all CPUs are small and consistent since only the cache associated with power level 0 is flushed (L1).

8.2.4.3. CPU_SUSPEND to deepest power level on all CPUs in sequence

CPU

PSCI_ENTRY (us)

PSCI_EXIT (us)

CFLUSH_OVERHEAD (us)

0

114

20

94

1

114

20

94

2

114

20

94

3

114

20

94

4

195

22

180

5

21

17

6

The CFLUSH_OVERHEAD times for lead CPU 4 and all CPUs in the non-lead cluster are large because all other CPUs in the cluster are powered down during the test. The CPU_SUSPEND call powers down to the cluster level, requiring a flush of both L1 and L2 caches.

The CFLUSH_OVERHEAD time for CPU 4 is a lot larger than those for the little CPUs because the L2 cache size for the big cluster is lot larger (2MB) compared to the little cluster (1MB).

The PSCI_ENTRY and CFLUSH_OVERHEAD times for CPU 5 are low because lead CPU 4 continues to run while CPU 5 is suspended. Hence CPU 5 only powers down to level 0, which only requires L1 cache flush.

8.2.4.4. CPU_SUSPEND to power level 0 on all CPUs in sequence

CPU

PSCI_ENTRY (us)

PSCI_EXIT (us)

CFLUSH_OVERHEAD (us)

0

22

14

5

1

22

14

5

2

21

14

5

3

22

14

5

4

17

14

6

5

18

15

6

Here the times are small and consistent since there is no contention and it is only necessary to flush the cache to power level 0 (L1). This is the best case scenario.

The PSCI_ENTRY times for CPUs in the big cluster are slightly smaller than for the CPUs in little cluster due to greater CPU performance.

The PSCI_EXIT times are generally lower than in the last test because the cluster remains powered on throughout the test and there is less code to execute on power on (for example, no need to enter CCI coherency)

8.2.4.5. CPU_OFF on all non-lead CPUs in sequence then CPU_SUSPEND on lead CPU to deepest power level

The test sequence here is as follows:

  1. Call CPU_ON and CPU_OFF on each non-lead CPU in sequence.

  2. Program wake up timer and suspend the lead CPU to the deepest power level.

  3. Call CPU_ON on non-lead CPU to get the timestamps from each CPU.

CPU

PSCI_ENTRY (us)

PSCI_EXIT (us)

CFLUSH_OVERHEAD (us)

0

110

28

93

1

110

28

93

2

110

28

93

3

111

28

93

4

195

22

181

5

20

23

6

The CFLUSH_OVERHEAD times for all little CPUs are large because all other CPUs in that cluster are powerered down during the test. The CPU_OFF call powers down to the cluster level, requiring a flush of both L1 and L2 caches.

The PSCI_ENTRY and CFLUSH_OVERHEAD times for CPU 5 are small because lead CPU 4 is running and CPU 5 only powers down to level 0, which only requires an L1 cache flush.

The CFLUSH_OVERHEAD time for CPU 4 is a lot larger than those for the little CPUs because the L2 cache size for the big cluster is lot larger (2MB) compared to the little cluster (1MB).

The PSCI_EXIT times for CPUs in the big cluster are slightly smaller than for CPUs in the little cluster due to greater CPU performance. These times generally are greater than the PSCI_EXIT times in the CPU_SUSPEND tests because there is more code to execute in the “on finisher” compared to the “suspend finisher” (for example, GIC redistributor register programming).

8.2.4.6. PSCI_VERSION on all CPUs in parallel

Since very little code is associated with PSCI_VERSION, this test approximates the round trip latency for handling a fast SMC at EL3 in TF.

CPU

TOTAL TIME (ns)

0

3020

1

2940

2

2980

3

3060

4

520

5

720

The times for the big CPUs are less than the little CPUs due to greater CPU performance.

We suspect the time for lead CPU 4 is shorter than CPU 5 due to subtle cache effects, given that these measurements are at the nano-second level.


Copyright (c) 2019-2023, Arm Limited and Contributors. All rights reserved.