Skip to content

Commit

Permalink
Few more results and SoC guessing.
Browse files Browse the repository at this point in the history 
Closes #37, closes #38
  • Loading branch information
@ThomasKaiser
ThomasKaiser committed Jan 22, 2022
1 parent 69a909b commit d6658d7
Show file tree
Hide file tree
Showing 16 changed files with 2,284 additions and 434 deletions.
24 changes: 13 additions & 11 deletions README.md
View file
Original file line number Diff line number Diff line change
Expand Up @@ -14,21 +14,23 @@ It has five **entirely different** usage modes:
* Plot thermal/consumption charts with `-p` to [measure efficiency of settings/devices](Efficiency_Graphing.md)
* Provide basic CLI monitoring functionality through the `-m` switch

The monitoring now also displays some SoC and CPU core information when starting:
The monitoring now also displays some hardware information when starting:

tk@odroidxu4:~$ sbc-bench -m
Samsung Exynos EXYNOS5800 rev 1, Kernel: armv7l, Userland: armhf
CPU topology (clusters, cpufreq members, clockspeeds)
Samsung Exynos EXYNOS5800 rev 1, Exynos 5422, Kernel: armv7l, Userland: armhf
CPU sysfs topology (clusters, cpufreq members, clockspeeds)
cpufreq min max
CPU cluster policy speed speed core type
0 1 0 200 1400 Cortex-A7
1 1 0 200 1400 Cortex-A7
2 1 0 200 1400 Cortex-A7
3 1 0 200 1400 Cortex-A7
4 0 4 200 2000 Cortex-A15
5 0 4 200 2000 Cortex-A15
6 0 4 200 2000 Cortex-A15
7 0 4 200 2000 Cortex-A15
0 1 0 200 1400 Cortex-A7 / r0p3
1 1 0 200 1400 Cortex-A7 / r0p3
2 1 0 200 1400 Cortex-A7 / r0p3
3 1 0 200 1400 Cortex-A7 / r0p3
4 0 4 200 2000 Cortex-A15 / r2p3
5 0 4 200 2000 Cortex-A15 / r2p3
6 0 4 200 2000 Cortex-A15 / r2p3
7 0 4 200 2000 Cortex-A15 / r2p3

Thermal source: /sys/devices/virtual/thermal/thermal_zone0/ (cpu0-thermal)

Time big.LITTLE load %cpu %sys %usr %nice %io %irq Temp
18:18:33: 800/ 500MHz 0.00 18% 0% 17% 0% 0% 0% 25.0°C
Expand Down
8 changes: 5 additions & 3 deletions Results.md
View file
Original file line number Diff line number Diff line change
Expand Up @@ -19,6 +19,7 @@ So do **not** rely on collected numbers unless you carefully read through all th
| [Cubietruck](http://ix.io/1u3W) | 960 MHz | 4.19 | Stretch armhf | 985 | 15260 | 18590 | 380 | 1380 | - |
| [EspressoBin](http://ix.io/1kt2) | 800 MHz | 4.17 | Stretch arm64 | 1138 | 54290 | 368330 | 1040 | 2490 | 1.23 |
| [EspressoBin](http://ix.io/1lCe) | 1200 MHz | 4.18 | Stretch arm64 | 1630 | 81900 | 544240 | 1000 | 2400 | 1.82 |
| [Gigabyte H270-T70](http://ix.io/3N5c) | 2000 Mhz | 5.16 | Sid arm64 | 107180 | 110340 | 340750 | 4180 | 17130 | - |
| [Helios4](http://ix.io/1jCy) | 1600 MHz | 4.14 | Stretch armhf | 2210 | 44785 *1280 | 42500 *98560 | 910 | 4840 | - |
| [Hugsun X99](http://ix.io/2ICt) | 2088/1800 MHz | 5.9 | Focal arm64 | 7710 | 412105 | 1184306 | 2270 | 5970 | - |
| [HummingBoard Pulse i.MX8M Quad](http://ix.io/27FC) | 1500 MHz | 4.19 | Buster arm64 | 4330 | 97380 | 695540 | 2230 | 9900 | - |
Expand All @@ -44,7 +45,7 @@ So do **not** rely on collected numbers unless you carefully read through all th
| [NanoPi Fire3](http://ix.io/3GmP) | 1400 MHz | 4.14 | Focal arm64 | 7360 | 118460 | 653770 | 1530 | 4590 | 11.27 |
| [NanoPi K1 Plus](http://ix.io/3EA8) | 1370 MHz | 5.10 | Focal arm64 | 3520 | 115020 | 639060 | 1060 | 3680 | 5.48 |
| [NanoPi K2](http://ix.io/1iT1) | 1480 MHz | 4.14 | Stretch arm64 | 3850 | 43020 | 50370 | 1660 | 3870 | 4.61 |
| [NanoPi M1 Plus](http://ix.io/3GnV) | 1370 MHz | 4.19 | Bionic armhf | 3030 | 26040 | 26580 | 830 | 3450 | - |
| [NanoPi M1 Plus](http://ix.io/3N2z) | 1370 MHz | 4.19 | Bionic armhf | 3030 | 25720 | 26660 | 830 | 3450 | - |
| [NanoPi M4](http://ix.io/1lzP) | 2000/1500 MHz | 4.19 | Stretch arm64 | 6400 | 334650 | 1128330 | 4080 | 8270 | 8.86 |
| [NanoPi M4v2](http://ix.io/3MAK) | 2015/1510 MHz | 5.10 | Bullseye arm64 | 6680 | 230410 | 921980 | 3110 | 7640 | - |
| [NanoPi NEO4](http://ix.io/1oho) | 2000/1500 MHz | 4.4| Stretch arm64 | 6510 | 320600 | 1128860 | 2260 | 4770 | 8.71 |
Expand Down Expand Up @@ -117,7 +118,7 @@ So do **not** rely on collected numbers unless you carefully read through all th
| [RockPro64](http://ix.io/1iFp) | 1800/1400 MHz | 4.18 | Stretch arm64 | 6300 | 237700 | 1021500 | 3650 | 8450 | 8.20 |
| [RockPro64](http://ix.io/2sZH) | 2010/1510 MHz | 5.4 | Focal arm64 | 6920 | 397740 | 1145300 | 3700 | 8430 | 11.55 |
| [RockPro64](http://ix.io/2yIx) | 2010/1510 MHz | 5.8 | Bullseye arm64 | 7000 | 369820 | 1144950 | 3690 | 8360 | 11.08 |
| [SBC2D70 (SSD202D)](http://ix.io/3MZF) | 1296 MHz | 5.16 | Sid armhf | 1540 | 19920 | 25250 | 830 | 3150 | - |
| [SBC2D70 (SSD202D)](http://ix.io/3N1U) | 1700 MHz | 5.16 | Sid armhf | 1960 | 26240 | 33120 | 770 | 3190 | - |
| [Teres-I](http://ix.io/1tJg) | 1050 MHz | 4.19 | Stretch arm64 | 2785 | 72261 | 491590 | 1080 | 2820 | - |
| [Tinkerboard](http://ix.io/1iSX) | 1730 MHz | 4.14 | Stretch armhf | 5350 | 63150 | 66600 | 1480 | 3900 | - |
| [Tinkerboard](http://ix.io/3Lir) | 1800 MHz | 4.4 | Buster armhf | 5440 | 62410 | 66300 | 1340 | 3510 | - |
Expand Down Expand Up @@ -151,6 +152,7 @@ So do **not** rely on collected numbers unless you carefully read through all th
* Cubietruck numbers are more or less representative for all other Allwinner A20 devices, same with Lime for Allwinner A10, Olimex Teres-I for Allwinner A64, Orange Pi "PC Plus" and "Plus 2" for Allwinner H2+/H3 and Orange Pi "PC 2" and Prime for Allwinner H5.
* Clearfog Pro and Helios4 use exactly same SoC (Armada 385), kernel and clockspeeds and the only reason why OpenSSL numbers differ is since Helios4 numbers were made using [Marvell's CESA crypto accelerator via cryptodev](https://forum.armbian.com/topic/7763-benchmarking-cpus/?do=findComment&comment=59569) which provides nice speed improvements with larger block sizes but also some initialization overhead with tiny block sizes. Also CPU utilization is way lower so the SoC is free for other stuff while performing better at the same time.
* EspressoBin's boot BLOB claims to run at up to 1GHz while real clockspeeds are lower maxing out with this setting at 790MHz (obviously a kernel bug -- see [details](https://forum.armbian.com/topic/4089-espressobin-support-development-efforts/?do=findComment&comment=60082))
* Gigabyte H270-T70 numbers are for one [blade module equipped with two Cavium ThunderX CN8890 (48 cores each)](https://twitter.com/linux_chenxing/status/1449603003057532930).
* HoneyComb LX2 numbers are made with pre-release settings (the NXP LX2160A's A72 cores clocked with 1.9 GHz instead of 2.0 GHz). Solid-Run's Jon Nettleton shared also slightly better [numbers with CPU cores at 2.4Ghz and RAM at 3200Mtps](https://www.cnx-software.com/2019/03/29/clearfog-itx-workstation-ultimate-arm-developer-platform/#comment-561998).
* Hugsun X99 is an [overclocked RK3399 TV box](https://github.com/ThomasKaiser/sbc-bench/issues/20), just to show the effect of overclocking the A53 cores to 1.8 GHz and the A72 to 2.1 GHz on an RK3399.
* Jetson Nano was [properly powered with 5V/5A via barrel plug](https://forum.armbian.com/topic/9921-nvidia-jetson-nano/?do=findComment&comment=78467) (when powering through Micro USB the board enters a lower consumption/performance profile)
Expand All @@ -167,7 +169,7 @@ So do **not** rely on collected numbers unless you carefully read through all th
* The highest clockspeeds listed for RPi 4/400 are 1800 MHz since this is what can be achieved with little overvoltage on early BCM2711B0 SoC revisions and is default _without overvolting_ on revision C0 or later. While it's possible to 'overclock' the RPi 4 to [2.15 GHz](http://ix.io/3Llk), [2.3 GHz](http://ix.io/3L37) and even [2.4 GHz](http://ix.io/3KVs) all of this requires overvolting the SoC up to `over_voltage=15`. Unfortunately the ThreadX DFVS ([Dynamic voltage and frequency scaling](https://developer.arm.com/documentation/100960/0100/Dynamic-Voltage-and-Frequency-Scaling)) implementation is rather primitive and as such this overvolting happens also when the CPU cores idle at low frequencies which pretty much fries the CPU cores all the time resulting in high(er) temperatures and consumption figures than necessary. That's why 1.8 GHz is considered the reasonable maximum here.
* Rock Pi 4B numbers are preliminary. Board has been tested *without* heatsink first so throttling occured as expected. Second time with higher cpufreq OPPs just a fan was added (fan without heatsink == pretty inefficient). Memory performance seems rather low but that's due to testing with vendor's **armhf** Linaro images -- see other RK3399 devices running same software stack, e.g. RockPro64 numbers above with kernel 4.4, armhf and also being limited to 1.8/1.4GHz.
* Last RockPro64 entry has been made after ayufan solved memory performance problem with Rockchip's 4.4 kernel on his images (see [discussion](https://forum.khadas.com/t/painlessly-usable-linux-distro/3124/24?u=tkaiser))
* [SBC2D70](http://linux-chenxing.org/infinity2/ido-sbc2d70/) results are representative for [SigmaStar SSD201/SSD202D dual Cortex-A7](http://linux-chenxing.org/infinity2/#ssd201ssd202d).
* [SBC2D70](http://linux-chenxing.org/infinity2/ido-sbc2d70/) results are somewhat representative for [SigmaStar SSD201/SSD202D dual Cortex-A7](http://linux-chenxing.org/infinity2/#ssd201ssd202d) in general even though clockspeeds exceeding 1.3 GHz are considered boost frequencies and require appropriate cooling.
* Vim2 is somewhat special: not a real big.LITTLE design but two A53 clusters controlled by a firmware BLOB that allows cluster 0 to clock up to 1414 MHz (reported falsely as 1512 MHz) and cluster 1 able to reach 1 GHz ([details](https://forum.khadas.com/t/cpu-frequency-up-to-2ghz/2010/23?u=tkaiser))
* The 'TRONFY MXQ S805' numbers should be similar to ODROID-C1/C1+ since same Amlogic S805 SoC at same clockspeed.
* x86 numbers are meant as comparison. x5-Z8300 numbers were made with an [UP Board](https://wiki.up-community.org/Hardware_Specification), 1st x5-Z8350 is an Atomic Pi and the 2nd a RockPi X, Celeron J3455 with an [ASRock J3455-ITX mainboard](https://forum.openmediavault.org/index.php/Thread/24093), Pentium N4200 on [UP2 Board](https://wiki.up-community.org/Hardware_Specification_UP2), Pentium J4205 on an [ASRock J4205-ITX](https://forum.openmediavault.org/index.php/Thread/24093-Efficient-low-cost-home-made-NAS/?postID=182578#post182578), Ryzen Embedded R1606G on [DFI GHF51 SBC](https://www.cnx-software.com/2020/08/10/amd-ryzen-embedded-sbc-review-with-ubuntu-20-04/), Celeron J4105 on two ODROID-H2 with different DDR4-PC19200 (2400MT/s) SO-DIMMs (remotely accessed via maze.odroid.com) and Celeron N4100 tested on an [ODROID-H2 engineering sample](https://forum.odroid.com/viewtopic.php?f=168&t=32911&p=239613#p239581) with single channel DRAM config, Pentium J5005 is in an [MINIX NEO J50C-4](https://www.cnx-software.com/2019/12/12/a-look-at-ubuntu-on-minix-neo-g41v-4-and-j50c-4-mini-pcs/), Pentium G4600 is inside a [TK Microserver MI106+](https://www.thomas-krenn.com/de/produkte/tower-systeme/silent-tower-server/microserver-mi106-plus.html).
Expand Down
Loading

0 comments on commit d6658d7

Please sign in to comment.