forked from ThomasKaiser/sbc-bench
-
Notifications
You must be signed in to change notification settings - Fork 0
/
3MGN.txt
352 lines (288 loc) · 15.9 KB
/
3MGN.txt
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
sbc-bench v0.9.1 Raspberry Pi Model B Rev 2 (Tue, 18 Jan 2022 16:08:48 +0000)
Distributor ID: Raspbian
Description: Raspbian GNU/Linux 11 (bullseye)
Release: 11
Codename: bullseye
Architecture: armhf
Raspberry Pi ThreadX version:
Oct 29 2021 10:49:08
Copyright (c) 2012 Broadcom
version b8a114e5a9877e91ca8f26d1a5ce904b2ad3cf13 (clean) (release) (start)
ThreadX configuration (/boot/config.txt):
dtparam=audio=on
camera_auto_detect=1
display_auto_detect=1
dtoverlay=vc4-kms-v3d
max_framebuffers=2
disable_overscan=1
[cm4]
otg_mode=1
[all]
[pi4]
arm_boost=1
[all]
Actual ThreadX settings:
aphy_params_current=547
arm_freq_min=700
audio_pwm_mode=514
camera_auto_detect=1
config_hdmi_boost=2
disable_auto_turbo=1
disable_commandline_tags=2
disable_overscan=1
display_auto_detect=1
display_hdmi_rotate=-1
display_lcd_rotate=-1
dphy_params_current=547
dvfs=3
enable_tvout=1
enable_uart=1
force_pwm_open=1
framebuffer_ignore_alpha=1
framebuffer_swap=1
ignore_lcd=1
init_uart_clock=0x2dc6c00
mask_gpu_interrupt0=3072
mask_gpu_interrupt1=26370
max_framebuffers=2
pause_burst_frames=1
program_serial_random=1
total_mem=512
hdmi_force_cec_address:0=65535
hdmi_force_cec_address:1=65535
hdmi_pixel_freq_limit:0=0x9a7ec80
/usr/bin/gcc (Raspbian 10.2.1-6+rpi1) 10.2.1 20210110
Uptime: 16:08:50 up 7 min, 1 user, load average: 0.54, 0.41, 0.26
Linux 5.10.63+ (raspberrypi) 01/18/22 _armv6l_ (1 CPU)
avg-cpu: %user %nice %system %iowait %steal %idle
10.67 0.12 8.16 1.66 0.00 79.39
Device tps kB_read/s kB_wrtn/s kB_dscd/s kB_read kB_wrtn kB_dscd
mmcblk0 13.82 293.35 48.41 0.00 130185 21484 0
total used free shared buff/cache available
Mem: 429Mi 34Mi 349Mi 0.0Ki 44Mi 347Mi
Swap: 99Mi 0B 99Mi
Filename Type Size Used Priority
/var/swap file 102396 0 -2
##########################################################################
Checking cpufreq OPP (ARM1176):
Cpufreq OPP: 700 ThreadX: 700 Measured: 690 @ 1.2V
##########################################################################
Executing benchmark on cpu0 (ARM1176):
tinymembench v0.4.9 (simple benchmark for memory throughput and latency)
==========================================================================
== Memory bandwidth tests ==
== ==
== Note 1: 1MB = 1000000 bytes ==
== Note 2: Results for 'copy' tests show how many bytes can be ==
== copied per second (adding together read and writen ==
== bytes would have provided twice higher numbers) ==
== Note 3: 2-pass copy means that we are using a small temporary buffer ==
== to first fetch data into it, and only then write it to the ==
== destination (source -> L1 cache, L1 cache -> destination) ==
== Note 4: If sample standard deviation exceeds 0.1%, it is shown in ==
== brackets ==
==========================================================================
C copy backwards : 119.8 MB/s (1.0%)
C copy backwards (32 byte blocks) : 120.7 MB/s (1.3%)
C copy backwards (64 byte blocks) : 123.1 MB/s (1.4%)
C copy : 122.3 MB/s (0.7%)
C copy prefetched (32 bytes step) : 358.1 MB/s (23.6%)
C copy prefetched (64 bytes step) : 201.5 MB/s (0.7%)
C 2-pass copy : 132.9 MB/s (1.6%)
C 2-pass copy prefetched (32 bytes step) : 240.6 MB/s (1.7%)
C 2-pass copy prefetched (64 bytes step) : 204.3 MB/s (1.0%)
C fill : 735.6 MB/s (0.2%)
C fill (shuffle within 16 byte blocks) : 735.7 MB/s (0.6%)
C fill (shuffle within 32 byte blocks) : 735.4 MB/s (0.9%)
C fill (shuffle within 64 byte blocks) : 735.1 MB/s (0.9%)
---
standard memcpy : 355.0 MB/s (1.4%)
standard memset : 1424.4 MB/s (1.1%)
---
VFP copy : 157.8 MB/s (1.0%)
VFP 2-pass copy : 141.0 MB/s (1.1%)
ARM fill (STRD) : 735.3 MB/s (0.2%)
ARM fill (STM with 8 registers) : 1088.3 MB/s
ARM fill (STM with 4 registers) : 1442.9 MB/s (0.9%)
ARM copy prefetched (incr pld) : 355.9 MB/s
ARM copy prefetched (wrap pld) : 186.5 MB/s
ARM 2-pass copy prefetched (incr pld) : 282.2 MB/s (0.9%)
ARM 2-pass copy prefetched (wrap pld) : 224.2 MB/s (1.3%)
==========================================================================
== Framebuffer read tests. ==
== ==
== Many ARM devices use a part of the system memory as the framebuffer, ==
== typically mapped as uncached but with write-combining enabled. ==
== Writes to such framebuffers are quite fast, but reads are much ==
== slower and very sensitive to the alignment and the selection of ==
== CPU instructions which are used for accessing memory. ==
== ==
== Many x86 systems allocate the framebuffer in the GPU memory, ==
== accessible for the CPU via a relatively slow PCI-E bus. Moreover, ==
== PCI-E is asymmetric and handles reads a lot worse than writes. ==
== ==
== If uncached framebuffer reads are reasonably fast (at least 100 MB/s ==
== or preferably >300 MB/s), then using the shadow framebuffer layer ==
== is not necessary in Xorg DDX drivers, resulting in a nice overall ==
== performance improvement. For example, the xf86-video-fbturbo DDX ==
== uses this trick. ==
==========================================================================
VFP copy (from framebuffer) : 164.3 MB/s (0.3%)
VFP 2-pass copy (from framebuffer) : 150.0 MB/s (1.3%)
ARM copy (from framebuffer) : 146.2 MB/s (0.4%)
ARM 2-pass copy (from framebuffer) : 160.8 MB/s (0.2%)
==========================================================================
== Memory latency test ==
== ==
== Average time is measured for random memory accesses in the buffers ==
== of different sizes. The larger is the buffer, the more significant ==
== are relative contributions of TLB, L1/L2 cache misses and SDRAM ==
== accesses. For extremely large buffer sizes we are expecting to see ==
== page table walk with several requests to SDRAM for almost every ==
== memory access (though 64MiB is not nearly large enough to experience ==
== this effect to its fullest). ==
== ==
== Note 1: All the numbers are representing extra time, which needs to ==
== be added to L1 cache latency. The cycle timings for L1 cache ==
== latency can be usually found in the processor documentation. ==
== Note 2: Dual random read means that we are simultaneously performing ==
== two independent memory accesses at a time. In the case if ==
== the memory subsystem can't handle multiple outstanding ==
== requests, dual random read has the same timings as two ==
== single reads performed one after another. ==
==========================================================================
block size : single random read / dual random read
1024 : 0.0 ns / 0.0 ns
2048 : 0.2 ns / 0.5 ns
4096 : 0.0 ns / 0.0 ns
8192 : 0.1 ns / 0.0 ns
16384 : 1.7 ns / 2.6 ns
32768 : 32.5 ns / 52.9 ns
65536 : 52.7 ns / 75.8 ns
131072 : 68.8 ns / 92.7 ns
262144 : 133.5 ns / 208.9 ns
524288 : 228.7 ns / 399.9 ns
1048576 : 276.5 ns / 498.9 ns
2097152 : 300.3 ns / 550.0 ns
4194304 : 312.8 ns / 575.6 ns
8388608 : 319.4 ns / 590.7 ns
16777216 : 333.4 ns / 606.7 ns
33554432 : 351.9 ns / 656.6 ns
67108864 : 383.3 ns / 719.5 ns
##########################################################################
Executing benchmark twice on cluster 0 (ARM1176)
OpenSSL 1.1.1k, built on 25 Mar 2021
type 16 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes 16384 bytes
aes-128-cbc 8351.64k 12832.98k 14582.27k 15186.26k 15245.31k 14783.83k
aes-128-cbc 7452.71k 12068.37k 14681.17k 15373.31k 15605.76k 15357.27k
aes-192-cbc 5841.53k 9809.92k 12228.61k 12979.54k 13101.74k 12927.57k
aes-192-cbc 6020.05k 10209.94k 12424.11k 13238.27k 13494.95k 13331.11k
aes-256-cbc 6475.07k 9686.12k 10880.68k 11452.76k 11408.73k 11512.49k
aes-256-cbc 6648.35k 9974.95k 11279.53k 11582.12k 11856.55k 11747.33k
##########################################################################
Executing benchmark single-threaded on cpu0 (ARM1176)
7-Zip (a) [32] 16.02 : Copyright (c) 1999-2016 Igor Pavlov : 2016-05-21
p7zip Version 16.02 (locale=C,Utf16=off,HugeFiles=on,32 bits,1 CPU LE)
LE
CPU Freq: 9142857 9142857 12800000 12800000 16000000 32000000 56888888 128000000 292571428
RAM size: 429 MB, # CPU hardware threads: 1
RAM usage: 219 MB, # Benchmark threads: 1
Compressing | Decompressing
Dict Speed Usage R/U Rating | Speed Usage R/U Rating
KiB/s % MIPS MIPS | KiB/s % MIPS MIPS
22: 225 99 223 220 | 4976 99 429 425
23: 209 98 217 214 | 4710 95 427 408
24: 206 99 225 222 | 4777 98 426 419
---------------------------------- | ------------------------------
Avr: 99 222 218 | 98 427 417
Tot: 98 324 318
Compression: 218
Decompression: 417
Total: 318
##########################################################################
Testing clockspeeds again. System health now:
Time fake/real load %cpu %sys %usr %nice %io %irq Temp VCore
16:34:06: 700/ 700MHz 1.28 100% 2% 97% 0% 0% 0% 53.0°C 1.2V
Checking cpufreq OPP (ARM1176):
Cpufreq OPP: 700 ThreadX: 700 Measured: 335 @ 1.2V
##########################################################################
Thermal source: /sys/devices/virtual/thermal/thermal_zone0/ (cpu-thermal)
System health while running tinymembench:
Time fake/real load %cpu %sys %usr %nice %io %irq Temp VCore
16:09:12: 700/ 700MHz 0.76 22% 7% 12% 0% 1% 0% 41.2°C 1.2V
16:10:11: 700/ 700MHz 1.02 100% 1% 98% 0% 0% 0% 42.8°C 1.2V
16:11:10: 700/ 700MHz 1.19 100% 0% 99% 0% 0% 0% 46.5°C 1.2V
16:12:10: 700/ 700MHz 1.33 100% 0% 99% 0% 0% 0% 46.5°C 1.2V
16:13:10: 700/ 700MHz 1.38 100% 0% 99% 0% 0% 0% 46.5°C 1.2V
16:14:08: 700/ 700MHz 1.36 100% 0% 99% 0% 0% 0% 46.5°C 1.2V
16:15:07: 700/ 700MHz 1.26 100% 0% 99% 0% 0% 0% 46.5°C 1.2V
16:16:07: 700/ 700MHz 1.28 100% 0% 99% 0% 0% 0% 46.5°C 1.2V
16:17:06: 700/ 700MHz 1.28 100% 0% 99% 0% 0% 0% 47.6°C 1.2V
16:18:06: 700/ 700MHz 1.28 100% 0% 99% 0% 0% 0% 47.6°C 1.2V
16:19:06: 700/ 700MHz 1.44 100% 1% 98% 0% 0% 0% 48.7°C 1.2V
16:20:05: 700/ 700MHz 1.42 100% 0% 99% 0% 0% 0% 48.7°C 1.2V
16:21:04: 700/ 700MHz 1.38 100% 0% 99% 0% 0% 0% 49.8°C 1.2V
16:22:04: 700/ 700MHz 1.32 100% 0% 99% 0% 0% 0% 49.8°C 1.2V
16:23:03: 700/ 700MHz 1.24 100% 0% 99% 0% 0% 0% 49.8°C 1.2V
16:24:03: 700/ 700MHz 1.31 100% 0% 99% 0% 0% 0% 50.8°C 1.2V
16:25:05: 700/ 700MHz 1.40 100% 1% 97% 0% 0% 0% 51.4°C 1.2V
16:26:03: 700/ 700MHz 1.36 100% 0% 98% 0% 0% 0% 51.4°C 1.2V
16:27:03: 700/ 700MHz 1.31 100% 0% 99% 0% 0% 0% 51.9°C 1.2V
16:28:02: 700/ 700MHz 1.30 100% 0% 99% 0% 0% 0% 51.9°C 1.2V
System health while running OpenSSL benchmark:
Time fake/real load %cpu %sys %usr %nice %io %irq Temp VCore
16:29:08: 700/ 700MHz 1.53 78% 2% 75% 0% 0% 0% 51.9°C 1.2V
16:29:43: 700/ 700MHz 1.58 100% 0% 99% 0% 0% 0% 52.5°C 1.2V
16:30:18: 700/ 700MHz 1.54 100% 0% 99% 0% 0% 0% 53.0°C 1.2V
System health while running 7-zip single core benchmark:
Time fake/real load %cpu %sys %usr %nice %io %irq Temp VCore
16:30:58: 700/ 700MHz 1.72 80% 2% 76% 0% 0% 0% 53.0°C 1.2V
16:32:32: 700/ 700MHz 1.29 100% 0% 99% 0% 0% 0% 51.9°C 1.2V
16:34:06: 700/ 700MHz 1.28 100% 2% 97% 0% 0% 0% 53.0°C 1.2V
##########################################################################
Throttling statistics (time spent on each cpufreq OPP):
700 MHz: 1609.74 sec
##########################################################################
dmesg output while running the benchmarks:
[ 445.793743] process 'local/src/tinymembench/tinymembench' started with executable stack
##########################################################################
Linux 5.10.63+ (raspberrypi) 01/18/22 _armv6l_ (1 CPU)
avg-cpu: %user %nice %system %iowait %steal %idle
80.02 0.09 2.57 0.36 0.00 16.97
Device tps kB_read/s kB_wrtn/s kB_dscd/s kB_read kB_wrtn kB_dscd
mmcblk0 3.71 69.40 12.64 14149.89 142869 26032 29131088
total used free shared buff/cache available
Mem: 429Mi 34Mi 334Mi 0.0Ki 60Mi 347Mi
Swap: 99Mi 0B 99Mi
Filename Type Size Used Priority
/var/swap file 102396 0 -2
CPU sysfs topology (clusters, cpufreq members, clockspeeds)
cpufreq min max
CPU cluster policy speed speed core type
0 -1 0 700 700 ARM1176 / r0p7
Architecture: armv6l
Byte Order: Little Endian
CPU(s): 1
On-line CPU(s) list: 0
Thread(s) per core: 1
Core(s) per socket: 1
Socket(s): 1
Vendor ID: ARM
Model: 7
Model name: ARM1176
Stepping: r0p7
CPU max MHz: 700.0000
CPU min MHz: 700.0000
BogoMIPS: 697.95
Flags: half thumb fastmult vfp edsp java tls
SoC guess: BCM2835
Compiler: /usr/bin/gcc (Raspbian 10.2.1-6+rpi1/arm-linux-gnueabihf)
Userland: armhf
ThreadX: b8a114e5a9877e91ca8f26d1a5ce904b2ad3cf13 / Oct 29 2021 10:49:08
Kernel: 5.10.63+/armv6l
CONFIG_HZ=100
CONFIG_HZ_100=y
CONFIG_HZ_FIXED=0
CONFIG_PREEMPTIRQ_TRACEPOINTS=y
CONFIG_PREEMPT_VOLUNTARY=y
| Raspberry Pi Model B Rev 2 | 700 MHz | 5.10 | Bullseye armv6l/armhf | 110 | 7900 | 11630 | 360 | 1420 | - |