Milk-V Mars

[geerlingguy]

Basic information

• Board URL (official): https://milkv.io/mars
• Board purchased from: Provided by Milk-V, available at Arace
• Board purchase date: June 26, 2024
• Board specs (as tested): MV050-D8 (8GB RAM)
• Board price (as tested): $68.99

Linux/system information

# output of `screenfetch`
         _,met$$$$$gg.           user@milkv
      ,g$$$$$$$$$$$$$$$P.        OS: Debian  bookworm
    ,g$$P""       """Y$$.".      Kernel: riscv64 Linux 5.15.0
   ,$$P'              `$$$.      Uptime: 19m
  ',$$P       ,ggs.     `$$b:    Packages: 1289
  `d$$'     ,$P"'   .    $$$     Shell: bash 5.2.2
   $$P      d$'     ,    $$P     Disk: 3.1G / 7.3G (43%)
   $$:      $$.   -    ,d$$'     CPU: Unknown @ 4x 1.5GHz
   $$\;      Y$b._   _,d$P'      RAM: 474MiB / 7908MiB
   Y$$.    `.`"Y$$$$P"'         
   `$$b      "-.__              
    `Y$$                        
     `Y$$.                      
       `$$b.                    
         `Y$$b.                 
            `"Y$b._             
                `""""    

# output of `uname -a`
Linux milkv 5.15.0 #1 SMP Mon Nov 13 18:56:24 CST 2023 riscv64 GNU/Linux

Benchmark results

CPU

• Geekbench 6: (74 single / 218 multi - https://browser.geekbench.com/v6/cpu/6746810)
• 2.06 Gflops at 0.44 Gflops/W (geerlingguy/top500-benchmark HPL result)

Power

• Idle power draw (at wall): 3 W
• Maximum simulated power draw (stress-ng --matrix 0): 5.3 W
• During Geekbench multicore benchmark: 5.3 W
• During top500 HPL benchmark: 4.7 W

Disk

SanDisk Extreme A2 256 GB microSD Card

Benchmark	Result
iozone 4K random read	5.79 MB/s
iozone 4K random write	0.00 MB/s
iozone 1M random read	21.66 MB/s
iozone 1M random write	0.00 MB/s
iozone 1M sequential read	21.68 MB/s
iozone 1M sequential write	20.26 MB/s

wget https://raw.githubusercontent.com/geerlingguy/pi-cluster/master/benchmarks/disk-benchmark.sh
chmod +x disk-benchmark.sh
sudo MOUNT_PATH=/ TEST_SIZE=1g ./disk-benchmark.sh

Run benchmark on any attached storage device (e.g. eMMC, microSD, NVMe, SATA) and add results under an additional heading.

Also consider running PiBenchmarks.com script.

Network

iperf3 results:

• iperf3 -c $SERVER_IP: 924 Mbps
• iperf3 -c $SERVER_IP --reverse: 776 Mbps
• iperf3 -c $SERVER_IP --bidir: 798 Mbps up, 294 Mbps down

(Be sure to test all interfaces, noting any that are non-functional.)

GPU

glmark2-es2 results:

Heh... it _ran_, but it was painful. Got a score of "0" after many tests reported 0 FPS.

Note: This benchmark requires an active display on the device. Not all devices may be able to run glmark2-es2, so in that case, make a note and move on!

TODO: See this issue for discussion about a full suite of standardized GPU benchmarks.

Memory

tinymembench results:

Click to expand memory benchmark result

tinymembench v0.4.10 (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                                     :   1006.1 MB/s (0.7%)
 C copy backwards (32 byte blocks)                    :   1008.0 MB/s (1.3%)
 C copy backwards (64 byte blocks)                    :   1006.6 MB/s (0.8%)
 C copy                                               :   1023.0 MB/s
 C copy prefetched (32 bytes step)                    :   1023.8 MB/s
 C copy prefetched (64 bytes step)                    :   1026.2 MB/s (0.2%)
 C 2-pass copy                                        :    754.0 MB/s (1.4%)
 C 2-pass copy prefetched (32 bytes step)             :    763.0 MB/s (0.8%)
 C 2-pass copy prefetched (64 bytes step)             :    877.8 MB/s (4.9%)
 C fill                                               :    871.8 MB/s (2.6%)
 C fill (shuffle within 16 byte blocks)               :    904.8 MB/s (2.9%)
 C fill (shuffle within 32 byte blocks)               :    864.0 MB/s (2.1%)
 C fill (shuffle within 64 byte blocks)               :    861.3 MB/s (2.2%)
 ---
 standard memcpy                                      :   1024.9 MB/s
 standard memset                                      :    815.7 MB/s (0.8%)

==========================================================================
== 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 :    5.4 ns          /    10.5 ns 
    131072 :    8.1 ns          /    15.8 ns 
    262144 :    9.5 ns          /    18.6 ns 
    524288 :   14.8 ns          /    28.5 ns 
   1048576 :   16.4 ns          /    30.6 ns 
   2097152 :   23.2 ns          /    41.3 ns 
   4194304 :   84.7 ns          /   166.2 ns 
   8388608 :  122.4 ns          /   242.4 ns 
  16777216 :  143.6 ns          /   285.4 ns 
  33554432 :  157.3 ns          /   313.3 ns 
  67108864 :  167.7 ns          /   334.5 ns

sbc-bench results

Run sbc-bench and paste a link to the results here:

https://0x0.st/Xa9J.bin

Phoronix Test Suite

Results from pi-general-benchmark.sh:

• pts/encode-mp3: N/A (test does not compile)
• pts/x264 4K: N/A (test does not compile)
• pts/x264 1080p: N/A (test does not compile)
• pts/phpbench: 63136
• pts/build-linux-kernel (defconfig): N/A (test quits with E: ERROR: modpost: "zero" [net/netfilter/nf_nat.ko] undefined!)

[geerlingguy]

One quick note: There is an official Debian image, but also Ubuntu 24.04 Server supports the Mars, but you have to update the bootloader via SPI first, if you want to use that.

[geerlingguy]

At boot, I'm not getting HDMI output, at least not to my typical monitor. Something that seems common with some dev boards.

I was able to SSH in, and it is a device registered to "Guangdong Starfive", with the local hostname milkv. I tried an apt update and got:

user@milkv:~$ sudo apt update
[sudo] password for user: 
Hit:1 https://snapshot.debian.org/archive/debian-ports/20221225T084846Z unstable InRelease
Err:1 https://snapshot.debian.org/archive/debian-ports/20221225T084846Z unstable InRelease
  The following signatures were invalid: EXPKEYSIG E852514F5DF312F6 Debian Ports Archive Automatic Signing Key (2022) <ftpmaster@ports-master.debian.org>
Reading package lists... Done
Building dependency tree... Done
Reading state information... Done
13 packages can be upgraded. Run 'apt list --upgradable' to see them.
W: An error occurred during the signature verification. The repository is not updated and the previous index files will be used. GPG error: https://snapshot.debian.org/archive/debian-ports/20221225T084846Z unstable InRelease: The following signatures were invalid: EXPKEYSIG E852514F5DF312F6 Debian Ports Archive Automatic Signing Key (2022) <ftpmaster@ports-master.debian.org>
W: Failed to fetch https://snapshot.debian.org/archive/debian-ports/20221225T084846Z/dists/unstable/InRelease  The following signatures were invalid: EXPKEYSIG E852514F5DF312F6 Debian Ports Archive Automatic Signing Key (2022) <ftpmaster@ports-master.debian.org>
W: Some index files failed to download. They have been ignored, or old ones used instead.
user@milkv:~$ date
Mon Jul  1 19:04:03 UTC 2024
user@milkv:~$ sudo apt upgrade
Reading package lists... Done
Building dependency tree... Done
Reading state information... Done
Calculating upgrade... Done
The following packages have been kept back:
  ffmpeg libavcodec59 libavdevice59 libavfilter8 libavformat59 libavutil57 libpostproc56 libsdl2-2.0-0 libswresample4
  libswscale6 vlc-plugin-base vlc-plugin-samba vlc-plugin-svg
0 upgraded, 0 newly installed, 0 to remove and 13 not upgraded.

[geerlingguy]

I noticed the default system partition /dev/mmcblk1p4 was only 4GB, and it ran out of space trying to install git to clone the top500-benchmark. So I ran through the expanding the boot partition guide helpfully published on the Milk-V docs site. It would be nice if that were run automatically on first boot like on the Pi and many other SBCs, but it's at least nice they include a guide :)

[geerlingguy]

Oops... trying to run some of the pts benchmarks, I run out of free space :(

    pts/phpbench-1.1.6:
        Test Installation 1 of 1
            There is not enough space at /var/lib/phoronix-test-suite/installed-tests/pts/phpbench-1.1.6/ for the test files.

Lock creation failed due to lack of disk space.

[geerlingguy]

And after finding this tweet from @mediocreDevops, I ran through the gauntlet of getting Ansible installed finally. I've documented it here: Installing Ansible on a RISC-V computer.

[geerlingguy]

I couldn't get the disk benchmarks to finish on the microSD card I had installed. Going to re-test with a larger potentially faster card...

[geerlingguy]

Another oddity: I can shut down the Mars using sudo shutdown now, but then to turn it back on, the button I assume is a power button doesn't do anything. Tried pressing it, holding for 3s, 5s, and 10s, but the only way to power back on is to unplug and re-plug the power supply.

[geerlingguy]

lspci isn't included with the default packages:

user@milkv:~$ lspci
-bash: lspci: command not found

Fixed with sudo apt install -y pciutils, and now I see this with an Inland 256 GB M.2 SSD plugged in using an A+E Key to M-key adapter:

user@milkv:~$ lspci
0000:00:00.0 PCI bridge: PLDA XpressRich-AXI Ref Design (rev 02)
0000:01:00.0 USB controller: VIA Technologies, Inc. VL805/806 xHCI USB 3.0 Controller (rev 01)
0001:00:00.0 PCI bridge: PLDA XpressRich-AXI Ref Design (rev 02)
0001:01:00.0 Non-Volatile memory controller: Silicon Motion, Inc. SM2263EN/SM2263XT SSD Controller (rev 03)

user@milkv:~$ sudo lspci -vvv
...
0001:01:00.0 Non-Volatile memory controller: Silicon Motion, Inc. SM2263EN/SM2263XT SSD Controller (rev 03) (prog-if 02 [NVM Express])
	Subsystem: Silicon Motion, Inc. SM2263EN/SM2263XT SSD Controller
	Control: I/O- Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx+
	Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=fast >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
	Latency: 0
	Interrupt: pin A routed to IRQ 56
	Region 0: Memory at 38000000 (64-bit, non-prefetchable) [size=16K]
	Capabilities: [40] Power Management version 3
		Flags: PMEClk- DSI- D1- D2- AuxCurrent=0mA PME(D0-,D1-,D2-,D3hot-,D3cold-)
		Status: D0 NoSoftRst+ PME-Enable- DSel=0 DScale=0 PME-
	Capabilities: [50] MSI: Enable- Count=1/8 Maskable+ 64bit+
		Address: 0000000000000000  Data: 0000
		Masking: 00000000  Pending: 00000000
	Capabilities: [70] Express (v2) Endpoint, MSI 00
		DevCap:	MaxPayload 256 bytes, PhantFunc 0, Latency L0s unlimited, L1 unlimited
			ExtTag- AttnBtn- AttnInd- PwrInd- RBE+ FLReset+ SlotPowerLimit 0W
		DevCtl:	CorrErr- NonFatalErr- FatalErr- UnsupReq-
			RlxdOrd+ ExtTag- PhantFunc- AuxPwr- NoSnoop- FLReset-
			MaxPayload 128 bytes, MaxReadReq 512 bytes
		DevSta:	CorrErr- NonFatalErr- FatalErr- UnsupReq- AuxPwr+ TransPend-
		LnkCap:	Port #0, Speed 8GT/s, Width x4, ASPM L1, Exit Latency L1 <8us
			ClockPM+ Surprise- LLActRep- BwNot- ASPMOptComp+
		LnkCtl:	ASPM Disabled; RCB 64 bytes, Disabled- CommClk-
			ExtSynch- ClockPM- AutWidDis- BWInt- AutBWInt-
		LnkSta:	Speed 5GT/s (downgraded), Width x1 (downgraded)
			TrErr- Train- SlotClk+ DLActive- BWMgmt- ABWMgmt-
		DevCap2: Completion Timeout: Range ABCD, TimeoutDis+ NROPrPrP- LTR+
			 10BitTagComp- 10BitTagReq- OBFF Not Supported, ExtFmt- EETLPPrefix-
			 EmergencyPowerReduction Not Supported, EmergencyPowerReductionInit-
			 FRS- TPHComp- ExtTPHComp-
			 AtomicOpsCap: 32bit- 64bit- 128bitCAS-
		DevCtl2: Completion Timeout: 50us to 50ms, TimeoutDis- LTR- 10BitTagReq- OBFF Disabled,
			 AtomicOpsCtl: ReqEn-
		LnkCap2: Supported Link Speeds: 2.5-8GT/s, Crosslink- Retimer- 2Retimers- DRS-
		LnkCtl2: Target Link Speed: 8GT/s, EnterCompliance- SpeedDis-
			 Transmit Margin: Normal Operating Range, EnterModifiedCompliance- ComplianceSOS-
			 Compliance Preset/De-emphasis: -6dB de-emphasis, 0dB preshoot
		LnkSta2: Current De-emphasis Level: -3.5dB, EqualizationComplete- EqualizationPhase1-
			 EqualizationPhase2- EqualizationPhase3- LinkEqualizationRequest-
			 Retimer- 2Retimers- CrosslinkRes: unsupported
	Capabilities: [b0] MSI-X: Enable+ Count=16 Masked-
		Vector table: BAR=0 offset=00002000
		PBA: BAR=0 offset=00002100
	Capabilities: [100 v2] Advanced Error Reporting
		UESta:	DLP- SDES- TLP- FCP- CmpltTO- CmpltAbrt- UnxCmplt- RxOF- MalfTLP- ECRC- UnsupReq- ACSViol-
		UEMsk:	DLP- SDES- TLP- FCP- CmpltTO- CmpltAbrt- UnxCmplt- RxOF- MalfTLP- ECRC- UnsupReq- ACSViol-
		UESvrt:	DLP+ SDES+ TLP- FCP+ CmpltTO- CmpltAbrt- UnxCmplt- RxOF+ MalfTLP+ ECRC- UnsupReq- ACSViol-
		CESta:	RxErr- BadTLP- BadDLLP- Rollover- Timeout- AdvNonFatalErr-
		CEMsk:	RxErr- BadTLP- BadDLLP- Rollover- Timeout- AdvNonFatalErr+
		AERCap:	First Error Pointer: 00, ECRCGenCap+ ECRCGenEn- ECRCChkCap+ ECRCChkEn-
			MultHdrRecCap- MultHdrRecEn- TLPPfxPres- HdrLogCap-
		HeaderLog: 00000000 00000000 00000000 00000000
	Capabilities: [158 v1] Secondary PCI Express
		LnkCtl3: LnkEquIntrruptEn- PerformEqu-
		LaneErrStat: 0
	Capabilities: [178 v1] Latency Tolerance Reporting
		Max snoop latency: 0ns
		Max no snoop latency: 0ns
	Capabilities: [180 v1] L1 PM Substates
		L1SubCap: PCI-PM_L1.2+ PCI-PM_L1.1+ ASPM_L1.2+ ASPM_L1.1+ L1_PM_Substates+
			  PortCommonModeRestoreTime=10us PortTPowerOnTime=10us
		L1SubCtl1: PCI-PM_L1.2- PCI-PM_L1.1- ASPM_L1.2- ASPM_L1.1-
			   T_CommonMode=0us LTR1.2_Threshold=0ns
		L1SubCtl2: T_PwrOn=10us
	Kernel driver in use: nvme

[geerlingguy]

For some reason with an NVMe SSD, I can't get random write speeds to register with iozone. Very weird.

I've also notice in regular use, I'll randomly see:

user@milkv:~$ 
Broadcast message from Debian-gdm@milkv on tty1 (Wed 2024-07-03 15:14:31 UTC):

The system will suspend now!

This has happened during benchmarks, and sometimes when I'm doing nothing at all, like right now.

[geerlingguy]

Ah, that looks like it's related to the image's hibernation settings for the desktop OS, see https://forums.debian.net/viewtopic.php?t=156005

[eshattow]

If you can tolerate not having proprietary drivers (graphics monitor output) then please see #22 (comment) for a guide that applies to VisionFive2/Star64/Mars/MarsCM with upstream U-Boot firmware and Linux kernel with an example of official Debian GNU/Linux installation.