forked from ThomasKaiser/sbc-bench
-
Notifications
You must be signed in to change notification settings - Fork 0
/
3XzI.txt
835 lines (718 loc) · 43.9 KB
/
3XzI.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
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
sbc-bench v0.9.6 Firefly ITX-3588J HDMI(Linux) (Thu, 12 May 2022 08:32:25 +0000)
Distributor ID: Ubuntu
Description: Ubuntu 20.04.4 LTS
Release: 20.04
Codename: focal
Architecture: arm64
/usr/bin/gcc (Ubuntu 9.4.0-1ubuntu1~20.04.1) 9.4.0
Uptime: 08:32:25 up 1:50, 3 users, load average: 0.08, 0.21, 1.07
Linux 5.10.66 (firefly) 05/12/22 _aarch64_ (8 CPU)
avg-cpu: %user %nice %system %iowait %steal %idle
8.39 0.08 0.47 0.02 0.00 91.04
Device tps kB_read/s kB_wrtn/s kB_dscd/s kB_read kB_wrtn kB_dscd
mmcblk1 5.54 119.28 182.18 0.00 793195 1211460 0
mmcblk1boot0 0.02 0.08 0.00 0.00 548 0 0
mmcblk1boot1 0.02 0.08 0.00 0.00 548 0 0
total used free shared buff/cache available
Mem: 7.5Gi 661Mi 6.5Gi 41Mi 305Mi 6.7Gi
Swap: 0B 0B 0B
##########################################################################
Checking cpufreq OPP for cpu0-cpu3 (Cortex-A55):
No cpufreq support available. Measured on cpu1: 915 Mhz (913.539/913.509/913.448)
Checking cpufreq OPP for cpu4-cpu5 (Cortex-A76):
No cpufreq support available. Measured on cpu5: 980 Mhz (977.693/977.623/977.589)
Checking cpufreq OPP for cpu6-cpu7 (Cortex-A76):
No cpufreq support available. Measured on cpu7: 985 Mhz (982.738/982.668/982.622)
##########################################################################
Executing benchmark on cpu0 (Cortex-A55):
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 : 2699.2 MB/s (0.3%)
C copy backwards (32 byte blocks) : 2648.8 MB/s (0.3%)
C copy backwards (64 byte blocks) : 2690.0 MB/s (0.2%)
C copy : 3227.0 MB/s
C copy prefetched (32 bytes step) : 1875.2 MB/s (0.3%)
C copy prefetched (64 bytes step) : 3377.4 MB/s
C 2-pass copy : 1506.8 MB/s (0.2%)
C 2-pass copy prefetched (32 bytes step) : 1184.1 MB/s
C 2-pass copy prefetched (64 bytes step) : 1660.4 MB/s
C fill : 7061.7 MB/s
C fill (shuffle within 16 byte blocks) : 6311.5 MB/s
C fill (shuffle within 32 byte blocks) : 6310.6 MB/s
C fill (shuffle within 64 byte blocks) : 6156.5 MB/s
---
standard memcpy : 3670.4 MB/s
standard memset : 11118.8 MB/s
---
NEON LDP/STP copy : 3040.4 MB/s
NEON LDP/STP copy pldl2strm (32 bytes step) : 1367.8 MB/s
NEON LDP/STP copy pldl2strm (64 bytes step) : 2181.1 MB/s
NEON LDP/STP copy pldl1keep (32 bytes step) : 2218.3 MB/s
NEON LDP/STP copy pldl1keep (64 bytes step) : 2799.8 MB/s
NEON LD1/ST1 copy : 2935.5 MB/s
NEON STP fill : 11087.8 MB/s
NEON STNP fill : 7867.2 MB/s (0.4%)
ARM LDP/STP copy : 3041.3 MB/s
ARM STP fill : 11083.4 MB/s
ARM STNP fill : 7866.8 MB/s (0.3%)
==========================================================================
== 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. ==
==========================================================================
NEON LDP/STP copy (from framebuffer) : 282.6 MB/s
NEON LDP/STP 2-pass copy (from framebuffer) : 260.1 MB/s
NEON LD1/ST1 copy (from framebuffer) : 77.2 MB/s (0.1%)
NEON LD1/ST1 2-pass copy (from framebuffer) : 75.1 MB/s
ARM LDP/STP copy (from framebuffer) : 148.7 MB/s
ARM LDP/STP 2-pass copy (from framebuffer) : 142.8 MB/s
==========================================================================
== 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 : 0.0 ns / 0.0 ns
4096 : 0.0 ns / 0.0 ns
8192 : 0.0 ns / 0.0 ns
16384 : 0.1 ns / 0.1 ns
32768 : 1.1 ns / 1.9 ns
65536 : 2.9 ns / 5.3 ns
131072 : 6.5 ns / 10.9 ns
262144 : 15.7 ns / 23.3 ns
524288 : 22.6 ns / 29.6 ns
1048576 : 26.4 ns / 31.6 ns
2097152 : 28.6 ns / 32.9 ns
4194304 : 58.0 ns / 81.4 ns
8388608 : 102.4 ns / 135.9 ns
16777216 : 127.2 ns / 155.6 ns
33554432 : 141.9 ns / 167.7 ns
67108864 : 154.9 ns / 184.2 ns
Executing benchmark on cpu4 (Cortex-A76):
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 : 8064.1 MB/s
C copy backwards (32 byte blocks) : 7610.5 MB/s
C copy backwards (64 byte blocks) : 7609.8 MB/s
C copy : 8054.7 MB/s
C copy prefetched (32 bytes step) : 6880.0 MB/s
C copy prefetched (64 bytes step) : 7278.1 MB/s
C 2-pass copy : 2584.1 MB/s (0.2%)
C 2-pass copy prefetched (32 bytes step) : 3203.2 MB/s
C 2-pass copy prefetched (64 bytes step) : 3370.7 MB/s
C fill : 15037.9 MB/s (0.3%)
C fill (shuffle within 16 byte blocks) : 15039.3 MB/s (0.3%)
C fill (shuffle within 32 byte blocks) : 15033.7 MB/s (0.4%)
C fill (shuffle within 64 byte blocks) : 14990.3 MB/s (0.3%)
---
standard memcpy : 10189.0 MB/s (0.1%)
standard memset : 15066.4 MB/s (0.5%)
---
NEON LDP/STP copy : 10153.3 MB/s (0.1%)
NEON LDP/STP copy pldl2strm (32 bytes step) : 7427.3 MB/s
NEON LDP/STP copy pldl2strm (64 bytes step) : 7768.2 MB/s
NEON LDP/STP copy pldl1keep (32 bytes step) : 7760.8 MB/s
NEON LDP/STP copy pldl1keep (64 bytes step) : 7760.2 MB/s
NEON LD1/ST1 copy : 10179.1 MB/s (0.2%)
NEON STP fill : 15073.8 MB/s (0.4%)
NEON STNP fill : 15075.4 MB/s (0.5%)
ARM LDP/STP copy : 10164.9 MB/s (0.1%)
ARM STP fill : 15071.5 MB/s (0.4%)
ARM STNP fill : 15070.2 MB/s (0.4%)
==========================================================================
== 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. ==
==========================================================================
NEON LDP/STP copy (from framebuffer) : 1477.5 MB/s
NEON LDP/STP 2-pass copy (from framebuffer) : 1246.9 MB/s
NEON LD1/ST1 copy (from framebuffer) : 1487.3 MB/s (0.1%)
NEON LD1/ST1 2-pass copy (from framebuffer) : 1249.3 MB/s (0.2%)
ARM LDP/STP copy (from framebuffer) : 1429.1 MB/s
ARM LDP/STP 2-pass copy (from framebuffer) : 1243.7 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 : 0.0 ns / 0.0 ns
4096 : 0.0 ns / 0.0 ns
8192 : 0.0 ns / 0.0 ns
16384 : 0.0 ns / 0.0 ns
32768 : 0.0 ns / 0.0 ns
65536 : 0.0 ns / 0.0 ns
131072 : 2.6 ns / 3.6 ns
262144 : 4.9 ns / 6.5 ns
524288 : 10.5 ns / 14.0 ns
1048576 : 21.2 ns / 27.7 ns
2097152 : 27.3 ns / 32.4 ns
4194304 : 52.7 ns / 72.5 ns
8388608 : 93.7 ns / 121.7 ns
16777216 : 122.9 ns / 145.7 ns
33554432 : 138.2 ns / 155.2 ns
67108864 : 149.3 ns / 162.0 ns
Executing benchmark on cpu6 (Cortex-A76):
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 : 8104.3 MB/s
C copy backwards (32 byte blocks) : 7651.2 MB/s
C copy backwards (64 byte blocks) : 7648.7 MB/s
C copy : 8097.5 MB/s
C copy prefetched (32 bytes step) : 6916.6 MB/s
C copy prefetched (64 bytes step) : 7316.2 MB/s
C 2-pass copy : 2604.2 MB/s
C 2-pass copy prefetched (32 bytes step) : 3223.9 MB/s
C 2-pass copy prefetched (64 bytes step) : 3411.9 MB/s
C fill : 15061.2 MB/s (0.5%)
C fill (shuffle within 16 byte blocks) : 15062.9 MB/s (0.3%)
C fill (shuffle within 32 byte blocks) : 15066.5 MB/s (0.3%)
C fill (shuffle within 64 byte blocks) : 15002.4 MB/s (0.3%)
---
standard memcpy : 10188.3 MB/s (0.2%)
standard memset : 15078.5 MB/s (0.4%)
---
NEON LDP/STP copy : 10148.7 MB/s (0.1%)
NEON LDP/STP copy pldl2strm (32 bytes step) : 7469.0 MB/s
NEON LDP/STP copy pldl2strm (64 bytes step) : 7800.5 MB/s (0.1%)
NEON LDP/STP copy pldl1keep (32 bytes step) : 7791.3 MB/s
NEON LDP/STP copy pldl1keep (64 bytes step) : 7797.3 MB/s
NEON LD1/ST1 copy : 10170.2 MB/s (0.2%)
NEON STP fill : 15076.8 MB/s (0.5%)
NEON STNP fill : 15078.3 MB/s (0.5%)
ARM LDP/STP copy : 10160.6 MB/s (0.2%)
ARM STP fill : 15077.1 MB/s (0.5%)
ARM STNP fill : 15077.7 MB/s (0.5%)
==========================================================================
== 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. ==
==========================================================================
NEON LDP/STP copy (from framebuffer) : 1479.1 MB/s
NEON LDP/STP 2-pass copy (from framebuffer) : 1249.9 MB/s
NEON LD1/ST1 copy (from framebuffer) : 1487.6 MB/s
NEON LD1/ST1 2-pass copy (from framebuffer) : 1250.1 MB/s
ARM LDP/STP copy (from framebuffer) : 1437.5 MB/s
ARM LDP/STP 2-pass copy (from framebuffer) : 1247.1 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 : 0.0 ns / 0.0 ns
4096 : 0.0 ns / 0.0 ns
8192 : 0.0 ns / 0.0 ns
16384 : 0.0 ns / 0.0 ns
32768 : 0.0 ns / 0.0 ns
65536 : 0.0 ns / 0.0 ns
131072 : 2.6 ns / 3.6 ns
262144 : 5.2 ns / 6.7 ns
524288 : 8.6 ns / 10.8 ns
1048576 : 21.1 ns / 27.7 ns
2097152 : 28.1 ns / 33.1 ns
4194304 : 52.9 ns / 73.6 ns
8388608 : 98.1 ns / 127.7 ns
16777216 : 125.2 ns / 148.3 ns
33554432 : 139.4 ns / 155.9 ns
67108864 : 148.8 ns / 161.5 ns
##########################################################################
Executing ramlat on cpu0 (Cortex-A55), results in ns:
size: 1x32 2x32 1x64 2x64 1xPTR 2xPTR 4xPTR
4k: 10.13 13.62 8.074 13.62 5.817 7.050 13.62
8k: 10.13 13.62 8.074 13.62 5.817 7.052 13.62
16k: 10.13 13.62 8.075 13.62 5.817 7.051 13.63
32k: 10.29 13.89 8.198 13.86 5.894 7.178 13.91
64k: 23.10 30.14 21.18 30.15 20.05 27.42 40.67
128k: 32.64 41.00 30.31 40.98 29.55 39.45 67.35
256k: 37.88 61.04 35.70 61.04 33.95 54.66 102.0
512k: 40.03 70.56 37.71 70.47 35.48 62.81 120.6
1024k: 40.26 71.19 37.95 71.07 35.70 63.41 132.2
2048k: 42.39 74.62 39.75 80.05 37.51 66.89 139.0
4096k: 77.05 148.9 78.10 137.7 71.01 169.5 343.1
8192k: 121.3 207.7 119.2 207.7 138.4 219.4 425.9
16384k: 145.5 236.3 139.9 236.3 150.0 247.6 521.4
Executing ramlat on cpu4 (Cortex-A76), results in ns:
size: 1x32 2x32 1x64 2x64 1xPTR 2xPTR 4xPTR
4k: 11.08 11.81 11.07 12.13 9.857 11.41 10.03
8k: 11.08 11.81 11.05 12.13 9.864 11.40 10.03
16k: 11.06 11.81 11.07 12.13 9.857 11.40 10.03
32k: 11.02 11.81 11.07 12.13 9.859 11.40 10.03
64k: 11.12 11.87 11.12 12.18 9.901 11.08 10.15
128k: 17.51 20.93 17.51 20.87 17.10 18.61 20.63
256k: 21.52 23.49 21.53 23.45 20.08 22.49 23.14
512k: 28.35 32.08 28.25 32.05 26.44 31.36 32.73
1024k: 43.91 47.31 43.88 47.21 41.79 47.09 53.15
2048k: 46.26 49.78 45.64 49.67 43.66 49.95 57.44
4096k: 89.25 89.21 80.80 86.07 93.58 92.23 87.47
8192k: 131.5 128.7 134.6 126.1 127.0 125.0 117.2
16384k: 153.5 146.3 161.9 150.9 151.0 146.3 150.7
Executing ramlat on cpu6 (Cortex-A76), results in ns:
size: 1x32 2x32 1x64 2x64 1xPTR 2xPTR 4xPTR
4k: 10.97 11.75 11.02 12.07 9.810 11.35 9.975
8k: 10.99 11.75 11.02 12.07 9.812 11.35 9.983
16k: 10.98 11.75 11.02 12.07 9.810 11.35 9.981
32k: 10.99 11.75 11.02 12.07 9.812 11.35 9.979
64k: 11.05 11.82 11.07 12.13 9.866 11.07 10.15
128k: 17.31 20.68 17.31 20.54 16.92 18.40 20.45
256k: 21.49 23.09 21.50 23.11 20.02 22.48 23.05
512k: 27.93 30.94 28.03 30.94 26.00 30.43 32.75
1024k: 44.53 47.15 44.48 47.07 42.24 46.94 52.65
2048k: 47.55 51.82 47.05 51.75 44.89 51.66 57.85
4096k: 81.66 91.45 81.77 85.32 94.19 83.40 79.31
8192k: 130.7 126.1 145.9 127.0 129.3 125.1 117.5
16384k: 159.9 148.0 153.4 147.8 154.7 147.1 146.2
##########################################################################
Executing benchmark on each cluster individually
OpenSSL 1.1.1f, built on 31 Mar 2020
type 16 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes 16384 bytes
aes-128-cbc 87990.15k 253062.91k 486757.63k 627099.99k 685861.55k 690350.76k
aes-128-cbc 293923.67k 553489.49k 708496.13k 758075.39k 777254.23k 779627.18k
aes-128-cbc 295437.31k 556587.65k 712456.96k 761737.56k 781170.01k 783510.19k
aes-192-cbc 83513.35k 223986.24k 391992.66k 481601.19k 516131.50k 518810.28k
aes-192-cbc 274177.48k 484878.02k 600051.63k 627351.55k 648481.45k 650035.20k
aes-192-cbc 275746.59k 487192.30k 603059.03k 630418.77k 651687.25k 653241.00k
aes-256-cbc 79842.01k 203770.41k 336984.83k 401778.69k 425866.58k 427753.47k
aes-256-cbc 256100.59k 430974.17k 520201.64k 545776.30k 556258.65k 557421.91k
aes-256-cbc 257136.12k 433054.95k 522670.85k 548495.70k 559035.73k 560196.27k
##########################################################################
Executing benchmark single-threaded on cpu0 (Cortex-A55)
7-Zip (a) [64] 16.02 : Copyright (c) 1999-2016 Igor Pavlov : 2016-05-21
p7zip Version 16.02 (locale=C,Utf16=off,HugeFiles=on,64 bits,8 CPUs LE)
LE
CPU Freq: - - - - 128000000 256000000 - - -
RAM size: 7674 MB, # CPU hardware threads: 8
RAM usage: 435 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: 788 100 767 767 | 11124 100 950 950
23: 743 100 757 757 | 10993 100 952 952
24: 731 100 786 786 | 10841 100 952 952
25: 715 100 817 817 | 10636 100 947 947
---------------------------------- | ------------------------------
Avr: 100 782 782 | 100 950 950
Tot: 100 866 866
Executing benchmark single-threaded on cpu4 (Cortex-A76)
7-Zip (a) [64] 16.02 : Copyright (c) 1999-2016 Igor Pavlov : 2016-05-21
p7zip Version 16.02 (locale=C,Utf16=off,HugeFiles=on,64 bits,8 CPUs LE)
LE
CPU Freq: 64000000 - - - - - - - 2048000000
RAM size: 7674 MB, # CPU hardware threads: 8
RAM usage: 435 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: 1673 100 1628 1628 | 16130 100 1377 1377
23: 1606 100 1638 1637 | 15941 100 1380 1380
24: 1564 100 1683 1682 | 15755 100 1383 1383
25: 1522 100 1739 1739 | 15512 100 1381 1381
---------------------------------- | ------------------------------
Avr: 100 1672 1672 | 100 1380 1380
Tot: 100 1526 1526
Executing benchmark single-threaded on cpu6 (Cortex-A76)
7-Zip (a) [64] 16.02 : Copyright (c) 1999-2016 Igor Pavlov : 2016-05-21
p7zip Version 16.02 (locale=C,Utf16=off,HugeFiles=on,64 bits,8 CPUs LE)
LE
CPU Freq: - - - - - - - 1024000000 -
RAM size: 7674 MB, # CPU hardware threads: 8
RAM usage: 435 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: 1690 100 1645 1645 | 16220 100 1385 1385
23: 1621 100 1652 1652 | 16077 100 1392 1392
24: 1570 100 1689 1689 | 15886 100 1395 1395
25: 1529 100 1747 1747 | 15634 100 1392 1392
---------------------------------- | ------------------------------
Avr: 100 1683 1683 | 100 1391 1391
Tot: 100 1537 1537
##########################################################################
Executing benchmark 3 times multi-threaded
7-Zip (a) [64] 16.02 : Copyright (c) 1999-2016 Igor Pavlov : 2016-05-21
p7zip Version 16.02 (locale=C,Utf16=off,HugeFiles=on,64 bits,8 CPUs LE)
LE
CPU Freq: - - 64000000 - 128000000 - - - -
RAM size: 7674 MB, # CPU hardware threads: 8
RAM usage: 1765 MB, # Benchmark threads: 8
Compressing | Decompressing
Dict Speed Usage R/U Rating | Speed Usage R/U Rating
KiB/s % MIPS MIPS | KiB/s % MIPS MIPS
22: 9463 749 1228 9206 | 94633 685 1179 8072
23: 8861 732 1234 9029 | 94873 692 1186 8210
24: 8890 759 1260 9559 | 93571 693 1186 8213
25: 8564 758 1290 9779 | 91870 692 1181 8176
---------------------------------- | ------------------------------
Avr: 750 1253 9393 | 690 1183 8168
Tot: 720 1218 8780
7-Zip (a) [64] 16.02 : Copyright (c) 1999-2016 Igor Pavlov : 2016-05-21
p7zip Version 16.02 (locale=C,Utf16=off,HugeFiles=on,64 bits,8 CPUs LE)
LE
CPU Freq: 64000000 64000000 - - - - - - -
RAM size: 7674 MB, # CPU hardware threads: 8
RAM usage: 1765 MB, # Benchmark threads: 8
Compressing | Decompressing
Dict Speed Usage R/U Rating | Speed Usage R/U Rating
KiB/s % MIPS MIPS | KiB/s % MIPS MIPS
22: 9481 748 1233 9224 | 96062 692 1184 8194
23: 8779 715 1251 8945 | 94904 693 1185 8213
24: 8628 736 1260 9277 | 93708 694 1185 8225
25: 8396 743 1290 9586 | 92749 699 1181 8254
---------------------------------- | ------------------------------
Avr: 736 1259 9258 | 694 1184 8221
Tot: 715 1221 8740
7-Zip (a) [64] 16.02 : Copyright (c) 1999-2016 Igor Pavlov : 2016-05-21
p7zip Version 16.02 (locale=C,Utf16=off,HugeFiles=on,64 bits,8 CPUs LE)
LE
CPU Freq: - - - - 128000000 - - - -
RAM size: 7674 MB, # CPU hardware threads: 8
RAM usage: 1765 MB, # Benchmark threads: 8
Compressing | Decompressing
Dict Speed Usage R/U Rating | Speed Usage R/U Rating
KiB/s % MIPS MIPS | KiB/s % MIPS MIPS
22: 9510 754 1227 9251 | 96193 693 1184 8205
23: 8303 685 1235 8460 | 94683 691 1186 8194
24: 8903 765 1251 9573 | 93500 692 1186 8206
25: 8638 765 1289 9863 | 91817 691 1182 8171
---------------------------------- | ------------------------------
Avr: 742 1250 9287 | 692 1184 8194
Tot: 717 1217 8740
Compression: 9393,9258,9287
Decompression: 8168,8221,8194
Total: 8780,8740,8740
##########################################################################
Testing clockspeeds again. System health now:
Time CPU n/a load %cpu %sys %usr %nice %io %irq Temp
09:03:27: --- 7.95 96% 2% 94% 0% 0% 0% 34.2°C
Checking cpufreq OPP for cpu0-cpu3 (Cortex-A55):
No cpufreq support available. Measured on cpu1: 915 Mhz (911.102/910.931/910.801)
Checking cpufreq OPP for cpu4-cpu5 (Cortex-A76):
No cpufreq support available. Measured on cpu5: 980 Mhz (975.961/975.869/975.615)
Checking cpufreq OPP for cpu6-cpu7 (Cortex-A76):
No cpufreq support available. Measured on cpu7: 985 Mhz (980.721/980.372/978.839)
##########################################################################
Thermal source: /sys/devices/virtual/thermal/thermal_zone0/ (soc-thermal)
System health while running tinymembench:
Time CPU n/a load %cpu %sys %usr %nice %io %irq Temp
08:32:30: --- 0.15 8% 0% 8% 0% 0% 0% 31.5°C
08:34:30: --- 0.99 13% 0% 12% 0% 0% 0% 31.5°C
08:36:31: --- 1.00 12% 0% 12% 0% 0% 0% 31.5°C
08:38:31: --- 1.00 12% 0% 12% 0% 0% 0% 33.3°C
08:40:31: --- 1.01 12% 0% 12% 0% 0% 0% 32.4°C
08:42:31: --- 1.00 12% 0% 12% 0% 0% 0% 31.5°C
08:44:31: --- 1.04 13% 0% 12% 0% 0% 0% 34.2°C
08:46:31: --- 1.05 13% 0% 12% 0% 0% 0% 32.4°C
08:48:31: --- 1.03 12% 0% 12% 0% 0% 0% 31.5°C
System health while running ramlat:
Time CPU n/a load %cpu %sys %usr %nice %io %irq Temp
08:48:52: --- 1.09 9% 0% 8% 0% 0% 0% 31.5°C
08:49:01: --- 1.08 12% 0% 12% 0% 0% 0% 31.5°C
08:49:10: --- 1.07 13% 0% 12% 0% 0% 0% 31.5°C
08:49:19: --- 1.06 13% 0% 12% 0% 0% 0% 31.5°C
08:49:28: --- 1.05 12% 0% 12% 0% 0% 0% 31.5°C
08:49:37: --- 1.04 13% 0% 12% 0% 0% 0% 31.5°C
08:49:46: --- 1.04 13% 0% 12% 0% 0% 0% 31.5°C
08:49:55: --- 1.03 12% 0% 12% 0% 0% 0% 31.5°C
System health while running OpenSSL benchmark:
Time CPU n/a load %cpu %sys %usr %nice %io %irq Temp
08:50:02: --- 1.03 9% 0% 8% 0% 0% 0% 32.4°C
08:50:18: --- 1.02 12% 0% 12% 0% 0% 0% 31.5°C
08:50:34: --- 1.02 12% 0% 12% 0% 0% 0% 31.5°C
08:50:50: --- 1.01 12% 0% 12% 0% 0% 0% 31.5°C
08:51:06: --- 1.01 12% 0% 12% 0% 0% 0% 31.5°C
08:51:22: --- 1.00 12% 0% 12% 0% 0% 0% 31.5°C
08:51:38: --- 1.00 12% 0% 12% 0% 0% 0% 31.5°C
08:51:54: --- 1.00 12% 0% 12% 0% 0% 0% 31.5°C
08:52:10: --- 1.00 12% 0% 12% 0% 0% 0% 31.5°C
08:52:26: --- 1.00 12% 0% 12% 0% 0% 0% 31.5°C
08:52:42: --- 1.00 12% 0% 12% 0% 0% 0% 31.5°C
System health while running 7-zip single core benchmark:
Time CPU n/a load %cpu %sys %usr %nice %io %irq Temp
08:52:44: --- 1.00 9% 0% 9% 0% 0% 0% 31.5°C
08:52:59: --- 1.00 12% 0% 12% 0% 0% 0% 31.5°C
08:53:14: --- 1.00 12% 0% 12% 0% 0% 0% 31.5°C
08:53:29: --- 1.00 12% 0% 12% 0% 0% 0% 31.5°C
08:53:44: --- 1.00 12% 0% 12% 0% 0% 0% 31.5°C
08:53:59: --- 1.00 12% 0% 12% 0% 0% 0% 31.5°C
08:54:14: --- 1.00 12% 0% 12% 0% 0% 0% 31.5°C
08:54:29: --- 1.00 12% 0% 12% 0% 0% 0% 31.5°C
08:54:44: --- 1.00 12% 0% 12% 0% 0% 0% 31.5°C
08:54:59: --- 1.00 12% 0% 12% 0% 0% 0% 31.5°C
08:55:14: --- 1.00 13% 0% 12% 0% 0% 0% 31.5°C
08:55:30: --- 1.00 12% 0% 12% 0% 0% 0% 31.5°C
08:55:45: --- 1.00 12% 0% 12% 0% 0% 0% 31.5°C
08:56:00: --- 1.00 12% 0% 12% 0% 0% 0% 31.5°C
08:56:15: --- 1.00 13% 0% 12% 0% 0% 0% 31.5°C
08:56:30: --- 1.00 12% 0% 12% 0% 0% 0% 31.5°C
08:56:45: --- 1.00 12% 0% 12% 0% 0% 0% 31.5°C
08:57:00: --- 1.00 12% 0% 12% 0% 0% 0% 31.5°C
08:57:15: --- 1.08 13% 0% 12% 0% 0% 0% 31.5°C
08:57:30: --- 1.06 12% 0% 12% 0% 0% 0% 31.5°C
08:57:45: --- 1.05 13% 0% 12% 0% 0% 0% 31.5°C
08:58:00: --- 1.04 12% 0% 12% 0% 0% 0% 31.5°C
08:58:15: --- 1.03 12% 0% 12% 0% 0% 0% 31.5°C
08:58:30: --- 1.02 12% 0% 12% 0% 0% 0% 31.5°C
System health while running 7-zip multi core benchmark:
Time CPU n/a load %cpu %sys %usr %nice %io %irq Temp
08:58:37: --- 1.02 9% 0% 9% 0% 0% 0% 31.5°C
08:58:47: --- 2.31 94% 1% 93% 0% 0% 0% 33.3°C
08:58:57: --- 2.95 91% 0% 91% 0% 0% 0% 32.4°C
08:59:10: --- 3.88 81% 1% 79% 0% 0% 0% 34.2°C
08:59:21: --- 4.59 79% 0% 78% 0% 0% 0% 32.4°C
08:59:34: --- 5.16 93% 2% 91% 0% 0% 0% 33.3°C
08:59:45: --- 5.28 80% 0% 80% 0% 0% 0% 32.4°C
08:59:55: --- 5.34 85% 3% 82% 0% 0% 0% 34.2°C
09:00:08: --- 5.97 95% 2% 93% 0% 0% 0% 34.2°C
09:00:19: --- 6.00 81% 1% 80% 0% 0% 0% 34.2°C
09:00:32: --- 6.39 97% 0% 97% 0% 0% 0% 34.2°C
09:00:42: --- 6.58 77% 1% 75% 0% 0% 0% 33.3°C
09:00:55: --- 6.80 99% 0% 98% 0% 0% 0% 34.2°C
09:01:05: --- 6.93 72% 1% 71% 0% 0% 0% 34.2°C
09:01:18: --- 7.24 96% 1% 94% 0% 0% 0% 34.2°C
09:01:30: --- 7.26 70% 1% 68% 0% 0% 0% 34.2°C
09:01:40: --- 7.69 97% 2% 94% 0% 0% 0% 34.2°C
09:01:52: --- 7.22 94% 1% 93% 0% 0% 0% 34.2°C
09:02:05: --- 7.54 79% 1% 78% 0% 0% 0% 34.2°C
09:02:15: --- 7.61 92% 0% 91% 0% 0% 0% 33.3°C
09:02:29: --- 7.54 81% 1% 79% 0% 0% 0% 34.2°C
09:02:39: --- 7.39 79% 0% 78% 0% 0% 0% 32.4°C
09:02:52: --- 7.62 93% 2% 91% 0% 0% 0% 34.2°C
09:03:03: --- 7.45 81% 0% 81% 0% 0% 0% 33.3°C
09:03:14: --- 7.85 84% 2% 81% 0% 0% 0% 34.2°C
09:03:27: --- 7.95 96% 2% 94% 0% 0% 0% 34.2°C
##########################################################################
Linux 5.10.66 (firefly) 05/12/22 _aarch64_ (8 CPU)
avg-cpu: %user %nice %system %iowait %steal %idle
11.91 0.06 0.47 0.02 0.00 87.53
Device tps kB_read/s kB_wrtn/s kB_dscd/s kB_read kB_wrtn kB_dscd
mmcblk1 4.45 94.31 142.93 0.00 803915 1218352 0
mmcblk1boot0 0.02 0.06 0.00 0.00 548 0 0
mmcblk1boot1 0.02 0.06 0.00 0.00 548 0 0
total used free shared buff/cache available
Mem: 7.5Gi 872Mi 6.3Gi 41Mi 318Mi 6.5Gi
Swap: 0B 0B 0B
CPU sysfs topology (clusters, cpufreq members, clockspeeds)
cpufreq min max
CPU cluster policy speed speed core type
0 0 0 - - Cortex-A55 / r2p0
1 0 0 - - Cortex-A55 / r2p0
2 0 0 - - Cortex-A55 / r2p0
3 0 0 - - Cortex-A55 / r2p0
4 1 0 - - Cortex-A76 / r4p0
5 1 0 - - Cortex-A76 / r4p0
6 2 0 - - Cortex-A76 / r4p0
7 2 0 - - Cortex-A76 / r4p0
Architecture: aarch64
CPU op-mode(s): 32-bit, 64-bit
Byte Order: Little Endian
CPU(s): 8
On-line CPU(s) list: 0-7
Thread(s) per core: 1
Core(s) per socket: 2
Socket(s): 3
Vendor ID: ARM
Model: 0
Model name: Cortex-A55
Stepping: r2p0
BogoMIPS: 48.00
L1d cache: 256 KiB
L1i cache: 256 KiB
L2 cache: 1 MiB
L3 cache: 3 MiB
Vulnerability Itlb multihit: Not affected
Vulnerability L1tf: Not affected
Vulnerability Mds: Not affected
Vulnerability Meltdown: Not affected
Vulnerability Spec store bypass: Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1: Mitigation; __user pointer sanitization
Vulnerability Spectre v2: Not affected
Vulnerability Srbds: Not affected
Vulnerability Tsx async abort: Not affected
Flags: fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm lrcpc dcpop asimddp
SoC guess: Rockchip RK3588 (35880000)
Compiler: /usr/bin/gcc (Ubuntu 9.4.0-1ubuntu1~20.04.1/aarch64-linux-gnu)
Userland: arm64
Kernel: 5.10.66/aarch64
CONFIG_HZ=300
CONFIG_HZ_300=y
CONFIG_PREEMPT_VOLUNTARY=y
raid6: neonx8 gen() 6846 MB/s
raid6: neonx8 xor() 5276 MB/s
raid6: neonx4 gen() 7042 MB/s
raid6: neonx4 xor() 5353 MB/s
raid6: neonx2 gen() 6462 MB/s
raid6: neonx2 xor() 5075 MB/s
raid6: neonx1 gen() 5293 MB/s
raid6: neonx1 xor() 4380 MB/s
raid6: int64x8 gen() 1632 MB/s
raid6: int64x8 xor() 1031 MB/s
raid6: int64x4 gen() 2048 MB/s
raid6: int64x4 xor() 1126 MB/s
raid6: int64x2 gen() 3019 MB/s
raid6: int64x2 xor() 1660 MB/s
raid6: int64x1 gen() 2477 MB/s
raid6: int64x1 xor() 1184 MB/s
raid6: using algorithm neonx4 gen() 7042 MB/s
raid6: .... xor() 5353 MB/s, rmw enabled
raid6: using neon recovery algorithm
xor: measuring software checksum speed
xor: using function: arm64_neon (12928 MB/sec)
cpu0/index2: 128K, level: 2, type: Unified
cpu0/index0: 32K, level: 1, type: Data
cpu0/index3: 3072K, level: 3, type: Unified
cpu0/index1: 32K, level: 1, type: Instruction
cpu1/index2: 128K, level: 2, type: Unified
cpu1/index0: 32K, level: 1, type: Data
cpu1/index3: 3072K, level: 3, type: Unified
cpu1/index1: 32K, level: 1, type: Instruction
cpu2/index2: 128K, level: 2, type: Unified
cpu2/index0: 32K, level: 1, type: Data
cpu2/index3: 3072K, level: 3, type: Unified
cpu2/index1: 32K, level: 1, type: Instruction
cpu3/index2: 128K, level: 2, type: Unified
cpu3/index0: 32K, level: 1, type: Data
cpu3/index3: 3072K, level: 3, type: Unified
cpu3/index1: 32K, level: 1, type: Instruction
cpu4/index2: 512K, level: 2, type: Unified
cpu4/index0: 64K, level: 1, type: Data
cpu4/index3: 3072K, level: 3, type: Unified
cpu4/index1: 64K, level: 1, type: Instruction
cpu5/index2: 512K, level: 2, type: Unified
cpu5/index0: 64K, level: 1, type: Data
cpu5/index3: 3072K, level: 3, type: Unified
cpu5/index1: 64K, level: 1, type: Instruction
cpu6/index2: 512K, level: 2, type: Unified
cpu6/index0: 64K, level: 1, type: Data
cpu6/index3: 3072K, level: 3, type: Unified
cpu6/index1: 64K, level: 1, type: Instruction
cpu7/index2: 512K, level: 2, type: Unified
cpu7/index0: 64K, level: 1, type: Data
cpu7/index3: 3072K, level: 3, type: Unified
cpu7/index1: 64K, level: 1, type: Instruction
| Firefly ITX-3588J HDMI(Linux) | no cpufreq support | 5.10 | Focal arm64 | 8750 | 295440 | 560200 | 10190 | 15080 | - |