chwe
Posts: 126
Joined: Tue Jul 31, 2018 1:35 pm

Re: Next raspberry ethernet?

Mon Feb 11, 2019 10:08 pm

rpdom wrote:
Mon Feb 11, 2019 5:20 pm
chwe wrote:
Mon Feb 11, 2019 4:59 pm
If you have a full HD display but 2/3 of it is always black, do you call it fullHD? Just ask for a friend..
Yes. If it is a true HD display. It may be that the device driving it isn't capable of full HD, but that is not the point. (I'll skip the overscan issue here because that isn't "2/3 of it").
Yes, it is faster than fast Ethernet on previous RPis but far away from real GbE..
The data is sent down the cable at 1000mb/S. No matter how you word it that is a fact.
no it's not about the device which is driving the display.. the display has fullHD but ~66% is black. yeah the pixels are there 'in theory' but not used.. Don't bring up overscan.. that's not what it is about..
You need a different example? On your pay slip it's written that you got 9000$ but the company takes back 6000$ so it's now your job to tell the govt. that you only got 3000$ (that's then the flow-control) otherwise taxex might hurt...
I really don't care if by 'electrical characteristics' it's fast Ethernet, GbE, or even 10 Gbps. I'm interested what I get out of it.. And from all SBCs which are by 'electrical characteristics' GbE capable the RPi is due to GbE over USB2 shared with all other USB ports just the low-performer here. That's not an issue if you don't need it and in most use-cases I have for the RPi, GbE isn't really needed.. but for other use-cases it might matter.. If you then have to deal with flow control etc.. you probably just buy an SBC which has GbE not only by 'electrical characteristics'..

Ernst
Posts: 1232
Joined: Sat Feb 04, 2017 9:39 am
Location: Germany

Re: Next raspberry ethernet?

Mon Feb 11, 2019 10:27 pm

chwe wrote:
Mon Feb 11, 2019 10:08 pm
I really don't care if by 'electrical characteristics' it's fast Ethernet, GbE, or even 10 Gbps. I'm interested what I get out of it..
What are you not getting out of it ? Is it latency, response time, transferrate ? Or is it something else ?
The road to insanity is paved with static ip addresses

jamesh
Raspberry Pi Engineer & Forum Moderator
Raspberry Pi Engineer & Forum Moderator
Posts: 23683
Joined: Sat Jul 30, 2011 7:41 pm

Re: Next raspberry ethernet?

Mon Feb 11, 2019 11:35 pm

chwe wrote:
Mon Feb 11, 2019 10:08 pm
rpdom wrote:
Mon Feb 11, 2019 5:20 pm
chwe wrote:
Mon Feb 11, 2019 4:59 pm
If you have a full HD display but 2/3 of it is always black, do you call it fullHD? Just ask for a friend..
Yes. If it is a true HD display. It may be that the device driving it isn't capable of full HD, but that is not the point. (I'll skip the overscan issue here because that isn't "2/3 of it").
Yes, it is faster than fast Ethernet on previous RPis but far away from real GbE..
The data is sent down the cable at 1000mb/S. No matter how you word it that is a fact.
no it's not about the device which is driving the display.. the display has fullHD but ~66% is black. yeah the pixels are there 'in theory' but not used.. Don't bring up overscan.. that's not what it is about..
You need a different example? On your pay slip it's written that you got 9000$ but the company takes back 6000$ so it's now your job to tell the govt. that you only got 3000$ (that's then the flow-control) otherwise taxex might hurt...
I really don't care if by 'electrical characteristics' it's fast Ethernet, GbE, or evesn 10 Gbps. I'm interested what I get out of it.. And from all SBCs which are by 'electrical characteristics' GbE capable the RPi is due to GbE over USB2 shared with all other USB ports just the low-performer here. That's not an issue if you don't need it and in most use-cases I have for the RPi, GbE isn't really needed.. but for other use-cases it might matter.. If you then have to deal with flow control etc.. you probably just buy an SBC which has GbE not only by 'electrical characteristics'..
When measured, do any SBC reach their maximum theoretical throughout? No, they don't. Yet they are all claiming GigE. Because that is the spec of the connection. Semantics? Possibly. Do you ever get maximum speed out of your broadband? Nope, yet that is sold by the maximum speed possible. Lots of examples and counter examples.
Principal Software Engineer at Raspberry Pi (Trading) Ltd.
Contrary to popular belief, humorous signatures are allowed. Here's an example...
"My grief counseller just died, luckily, he was so good, I didn't care."

chwe
Posts: 126
Joined: Tue Jul 31, 2018 1:35 pm

Re: Next raspberry ethernet?

Tue Feb 12, 2019 1:22 am

Ernst wrote:
Mon Feb 11, 2019 10:27 pm
What are you not getting out of it ? Is it latency, response time, transferrate ? Or is it something else ?
transfer-rate when needed.. If I don't care I go for an SBC without GbE. If I would ever buy a 3b+ I would probably cap it to full speed cause I don't want to deal with flow-control..
jamesh wrote:
Mon Feb 11, 2019 11:35 pm
When measured, do any SBC reach their maximum theoretical throughout? No, they don't. Yet they are all claiming GigE. Because that is the spec of the connection. Semantics? Possibly. Do you ever get maximum speed out of your broadband? Nope, yet that is sold by the maximum speed possible. Lots of examples and counter examples.
So let's nail down those 'semantics' a bit

RockPi 4b: 192.168.0.10
Odroid HC1: 192.168.0.11
Tinkerboard: 192.168.0.84

first round RockPi as server:
tinker:

Code: Select all

opi@tinkerboard:~$ iperf3 -c 192.168.0.10
Connecting to host 192.168.0.10, port 5201
[  4] local 192.168.0.84 port 37858 connected to 192.168.0.10 port 5201
[ ID] Interval           Transfer     Bandwidth       Retr  Cwnd
[  4]   0.00-1.00   sec   109 MBytes   913 Mbits/sec    0    363 KBytes
[  4]   1.00-2.00   sec   112 MBytes   943 Mbits/sec    0    400 KBytes
[  4]   2.00-3.00   sec   112 MBytes   941 Mbits/sec    0    400 KBytes
[  4]   3.00-4.00   sec   112 MBytes   942 Mbits/sec    0    400 KBytes
[  4]   4.00-5.00   sec   112 MBytes   941 Mbits/sec    0    400 KBytes
[  4]   5.00-6.00   sec   112 MBytes   942 Mbits/sec    0    400 KBytes
[  4]   6.00-7.00   sec   112 MBytes   941 Mbits/sec    0    400 KBytes
[  4]   7.00-8.00   sec   112 MBytes   941 Mbits/sec    0    400 KBytes
[  4]   8.00-9.00   sec   112 MBytes   942 Mbits/sec    0    400 KBytes
[  4]   9.00-10.00  sec   112 MBytes   941 Mbits/sec    0    400 KBytes
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval           Transfer     Bandwidth       Retr
[  4]   0.00-10.00  sec  1.09 GBytes   939 Mbits/sec    0             sender
[  4]   0.00-10.00  sec  1.09 GBytes   938 Mbits/sec                  receiver

iperf Done.
odroid:

Code: Select all

opi@odroidxu4:~$ iperf3 -c 192.168.0.10
Connecting to host 192.168.0.10, port 5201
[  4] local 192.168.0.11 port 44682 connected to 192.168.0.10 port 5201
[ ID] Interval           Transfer     Bandwidth       Retr  Cwnd
[  4]   0.00-1.00   sec   108 MBytes   908 Mbits/sec    0    366 KBytes
[  4]   1.00-2.00   sec   110 MBytes   926 Mbits/sec    0    366 KBytes
[  4]   2.00-3.00   sec   110 MBytes   926 Mbits/sec    0    366 KBytes
[  4]   3.00-4.00   sec   110 MBytes   926 Mbits/sec    0    366 KBytes
[  4]   4.00-5.00   sec   110 MBytes   926 Mbits/sec    0    366 KBytes
[  4]   5.00-6.00   sec   110 MBytes   926 Mbits/sec    0    366 KBytes
[  4]   6.00-7.00   sec   110 MBytes   927 Mbits/sec    0    404 KBytes
[  4]   7.00-8.00   sec   110 MBytes   926 Mbits/sec    0    404 KBytes
[  4]   8.00-9.00   sec   110 MBytes   926 Mbits/sec    0    404 KBytes
[  4]   9.00-10.00  sec   110 MBytes   926 Mbits/sec    0    404 KBytes
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval           Transfer     Bandwidth       Retr
[  4]   0.00-10.00  sec  1.08 GBytes   924 Mbits/sec    0             sender
[  4]   0.00-10.00  sec  1.08 GBytes   924 Mbits/sec                  receiver

iperf Done.
second round Odroid as a server
tinker:

Code: Select all

opi@tinkerboard:~$ iperf3 -c 192.168.0.11
Connecting to host 192.168.0.11, port 5201
[  4] local 192.168.0.84 port 37874 connected to 192.168.0.11 port 5201
[ ID] Interval           Transfer     Bandwidth       Retr  Cwnd
[  4]   0.00-1.00   sec   112 MBytes   940 Mbits/sec    0    392 KBytes
[  4]   1.00-2.00   sec   112 MBytes   941 Mbits/sec    0    392 KBytes
[  4]   2.00-3.00   sec   112 MBytes   941 Mbits/sec    0    392 KBytes
[  4]   3.00-4.00   sec   112 MBytes   941 Mbits/sec    0    392 KBytes
[  4]   4.00-5.00   sec   112 MBytes   941 Mbits/sec    0    392 KBytes
[  4]   5.00-6.00   sec   112 MBytes   942 Mbits/sec    0    392 KBytes
[  4]   6.00-7.00   sec   112 MBytes   941 Mbits/sec    0    392 KBytes
[  4]   7.00-8.00   sec   112 MBytes   941 Mbits/sec    0    392 KBytes
[  4]   8.00-9.00   sec   112 MBytes   941 Mbits/sec    0    392 KBytes
[  4]   9.00-10.00  sec   112 MBytes   941 Mbits/sec    1    352 KBytes
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval           Transfer     Bandwidth       Retr
[  4]   0.00-10.00  sec  1.10 GBytes   941 Mbits/sec    1             sender
[  4]   0.00-10.00  sec  1.09 GBytes   940 Mbits/sec                  receiver

iperf Done.
rockpi:

Code: Select all

opi@rockpi:~$ iperf3 -c 192.168.0.11
Connecting to host 192.168.0.11, port 5201
[  4] local 192.168.0.10 port 56214 connected to 192.168.0.11 port 5201
[ ID] Interval           Transfer     Bandwidth       Retr  Cwnd
[  4]   0.00-1.00   sec   113 MBytes   951 Mbits/sec    0    554 KBytes
[  4]   1.00-2.00   sec   112 MBytes   941 Mbits/sec    0    554 KBytes
[  4]   2.00-3.00   sec   112 MBytes   942 Mbits/sec    0    554 KBytes
[  4]   3.00-4.00   sec   112 MBytes   941 Mbits/sec    0    554 KBytes
[  4]   4.00-5.00   sec   112 MBytes   941 Mbits/sec    0    554 KBytes
[  4]   5.00-6.00   sec   112 MBytes   941 Mbits/sec    0    554 KBytes
[  4]   6.00-7.00   sec   112 MBytes   941 Mbits/sec    0    554 KBytes
[  4]   7.00-8.00   sec   112 MBytes   942 Mbits/sec    0    554 KBytes
[  4]   8.00-9.00   sec   112 MBytes   941 Mbits/sec    0    554 KBytes
[  4]   9.00-10.00  sec   112 MBytes   941 Mbits/sec    0    554 KBytes
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval           Transfer     Bandwidth       Retr
[  4]   0.00-10.00  sec  1.10 GBytes   942 Mbits/sec    0             sender
[  4]   0.00-10.00  sec  1.09 GBytes   940 Mbits/sec                  receiver

iperf Done.
last round tinker as a server:
rockpi:

Code: Select all

opi@rockpi:~$ iperf3 -c 192.168.0.84
Connecting to host 192.168.0.84, port 5201
[  4] local 192.168.0.10 port 36040 connected to 192.168.0.84 port 5201
[ ID] Interval           Transfer     Bandwidth       Retr  Cwnd
[  4]   0.00-1.00   sec   113 MBytes   945 Mbits/sec    0    424 KBytes
[  4]   1.00-2.00   sec   112 MBytes   942 Mbits/sec    0    424 KBytes
[  4]   2.00-3.00   sec   112 MBytes   942 Mbits/sec    0    424 KBytes
[  4]   3.00-4.00   sec   112 MBytes   941 Mbits/sec    0    424 KBytes
[  4]   4.00-5.00   sec   112 MBytes   941 Mbits/sec    0    424 KBytes
[  4]   5.00-6.00   sec   112 MBytes   942 Mbits/sec    0    424 KBytes
[  4]   6.00-7.00   sec   112 MBytes   941 Mbits/sec    0    424 KBytes
[  4]   7.00-8.00   sec   112 MBytes   941 Mbits/sec    0    424 KBytes
[  4]   8.00-9.00   sec   112 MBytes   942 Mbits/sec    0    424 KBytes
[  4]   9.00-10.00  sec   112 MBytes   941 Mbits/sec    0    424 KBytes
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval           Transfer     Bandwidth       Retr
[  4]   0.00-10.00  sec  1.10 GBytes   942 Mbits/sec    0             sender
[  4]   0.00-10.00  sec  1.09 GBytes   940 Mbits/sec                  receiver

iperf Done.
Odroid:

Code: Select all

opi@odroidxu4:~$ iperf3 -c 192.168.0.84
Connecting to host 192.168.0.84, port 5201
[  4] local 192.168.0.11 port 47646 connected to 192.168.0.84 port 5201
[ ID] Interval           Transfer     Bandwidth       Retr  Cwnd
[  4]   0.00-1.00   sec   109 MBytes   914 Mbits/sec    0    355 KBytes
[  4]   1.00-2.00   sec   110 MBytes   923 Mbits/sec    0    355 KBytes
[  4]   2.00-3.00   sec   110 MBytes   923 Mbits/sec    0    355 KBytes
[  4]   3.00-4.00   sec   110 MBytes   924 Mbits/sec    0    400 KBytes
[  4]   4.00-5.00   sec   110 MBytes   923 Mbits/sec    0    400 KBytes
[  4]   5.00-6.00   sec   110 MBytes   924 Mbits/sec    0    400 KBytes
[  4]   6.00-7.00   sec   110 MBytes   924 Mbits/sec    0    400 KBytes
[  4]   7.00-8.00   sec   110 MBytes   924 Mbits/sec    0    400 KBytes
[  4]   8.00-9.00   sec   110 MBytes   923 Mbits/sec    0    400 KBytes
[  4]   9.00-10.00  sec   110 MBytes   923 Mbits/sec    0    400 KBytes
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval           Transfer     Bandwidth       Retr
[  4]   0.00-10.00  sec  1.07 GBytes   923 Mbits/sec    0             sender
[  4]   0.00-10.00  sec  1.07 GBytes   922 Mbits/sec                  receiver

iperf Done.
semantics may matter here.. :lol:

sorry don't have a Pi3b+ to get it run under the same conditions... But we can take the numbers provided by @MuntyScruntfundle
MuntyScruntfundle wrote:
Sat Feb 09, 2019 9:01 pm
A simple iperf test one at a time to a central Windows PC running the Windows 10 unix install returns Bandwidth of 293 for the Rasp, 949 for the Tink.
Running both boards together is even more surprising, Rasp 87, Tink 910.
All those boards running an debian stretch based Linux with a mainline based kernel.. No tweaks just stock OS. RockPi and Odroid should be capable of saturating the link (HC1 due to SATA over USB3, RockPi either via USB3 or NVme. The tinker not fully (only USB2) but at least the SDIO interface for its SD card is a speedy one, so with a decent card you might get near to saturation:

Code: Select all

                                                              random    random     bkwd    record    stride                               
              kB  reclen    write  rewrite    read    reread    read     write     read   rewrite      read   fwrite frewrite    fread  freread
          102400       4     3494     3573    14684    14646    13265     5467
          102400      16     9693    10659    33987    33997    32392    14184
          102400     512    28791    43623    63744    63746    63246    49403
          102400    1024    59659    59934    64568    64577    64484    50461
          102400   16384    58269    60555    68443    68358    68443    45531

Ernst
Posts: 1232
Joined: Sat Feb 04, 2017 9:39 am
Location: Germany

Re: Next raspberry ethernet?

Tue Feb 12, 2019 7:41 am

chwe wrote:
Tue Feb 12, 2019 1:22 am

All those boards running an debian stretch based Linux with a mainline based kernel.. No tweaks just stock OS. RockPi and Odroid should be capable of saturating the link (HC1 due to SATA over USB3, RockPi either via USB3 or NVme. The tinker not fully (only USB2) but at least the SDIO interface for its SD card is a speedy one, so with a decent card you might get near to saturation:
Can you explain why the results of these tests are so important to you ? What is the significance of getting near to saturation ?
The road to insanity is paved with static ip addresses

jamesh
Raspberry Pi Engineer & Forum Moderator
Raspberry Pi Engineer & Forum Moderator
Posts: 23683
Joined: Sat Jul 30, 2011 7:41 pm

Re: Next raspberry ethernet?

Tue Feb 12, 2019 8:04 am

chwe wrote:
Tue Feb 12, 2019 1:22 am
Ernst wrote:
Mon Feb 11, 2019 10:27 pm
What are you not getting out of it ? Is it latency, response time, transferrate ? Or is it something else ?
transfer-rate when needed.. If I don't care I go for an SBC without GbE. If I would ever buy a 3b+ I would probably cap it to full speed cause I don't want to deal with flow-control..
jamesh wrote:
Mon Feb 11, 2019 11:35 pm
When measured, do any SBC reach their maximum theoretical throughout? No, they don't. Yet they are all claiming GigE. Because that is the spec of the connection. Semantics? Possibly. Do you ever get maximum speed out of your broadband? Nope, yet that is sold by the maximum speed possible. Lots of examples and counter examples.
So let's nail down those 'semantics' a bit

RockPi 4b: 192.168.0.10
Odroid HC1: 192.168.0.11
Tinkerboard: 192.168.0.84

first round RockPi as server:
tinker:

Code: Select all

opi@tinkerboard:~$ iperf3 -c 192.168.0.10
Connecting to host 192.168.0.10, port 5201
[  4] local 192.168.0.84 port 37858 connected to 192.168.0.10 port 5201
[ ID] Interval           Transfer     Bandwidth       Retr  Cwnd
[  4]   0.00-1.00   sec   109 MBytes   913 Mbits/sec    0    363 KBytes
[  4]   1.00-2.00   sec   112 MBytes   943 Mbits/sec    0    400 KBytes
[  4]   2.00-3.00   sec   112 MBytes   941 Mbits/sec    0    400 KBytes
[  4]   3.00-4.00   sec   112 MBytes   942 Mbits/sec    0    400 KBytes
[  4]   4.00-5.00   sec   112 MBytes   941 Mbits/sec    0    400 KBytes
[  4]   5.00-6.00   sec   112 MBytes   942 Mbits/sec    0    400 KBytes
[  4]   6.00-7.00   sec   112 MBytes   941 Mbits/sec    0    400 KBytes
[  4]   7.00-8.00   sec   112 MBytes   941 Mbits/sec    0    400 KBytes
[  4]   8.00-9.00   sec   112 MBytes   942 Mbits/sec    0    400 KBytes
[  4]   9.00-10.00  sec   112 MBytes   941 Mbits/sec    0    400 KBytes
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval           Transfer     Bandwidth       Retr
[  4]   0.00-10.00  sec  1.09 GBytes   939 Mbits/sec    0             sender
[  4]   0.00-10.00  sec  1.09 GBytes   938 Mbits/sec                  receiver

iperf Done.
odroid:

Code: Select all

opi@odroidxu4:~$ iperf3 -c 192.168.0.10
Connecting to host 192.168.0.10, port 5201
[  4] local 192.168.0.11 port 44682 connected to 192.168.0.10 port 5201
[ ID] Interval           Transfer     Bandwidth       Retr  Cwnd
[  4]   0.00-1.00   sec   108 MBytes   908 Mbits/sec    0    366 KBytes
[  4]   1.00-2.00   sec   110 MBytes   926 Mbits/sec    0    366 KBytes
[  4]   2.00-3.00   sec   110 MBytes   926 Mbits/sec    0    366 KBytes
[  4]   3.00-4.00   sec   110 MBytes   926 Mbits/sec    0    366 KBytes
[  4]   4.00-5.00   sec   110 MBytes   926 Mbits/sec    0    366 KBytes
[  4]   5.00-6.00   sec   110 MBytes   926 Mbits/sec    0    366 KBytes
[  4]   6.00-7.00   sec   110 MBytes   927 Mbits/sec    0    404 KBytes
[  4]   7.00-8.00   sec   110 MBytes   926 Mbits/sec    0    404 KBytes
[  4]   8.00-9.00   sec   110 MBytes   926 Mbits/sec    0    404 KBytes
[  4]   9.00-10.00  sec   110 MBytes   926 Mbits/sec    0    404 KBytes
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval           Transfer     Bandwidth       Retr
[  4]   0.00-10.00  sec  1.08 GBytes   924 Mbits/sec    0             sender
[  4]   0.00-10.00  sec  1.08 GBytes   924 Mbits/sec                  receiver

iperf Done.
second round Odroid as a server
tinker:

Code: Select all

opi@tinkerboard:~$ iperf3 -c 192.168.0.11
Connecting to host 192.168.0.11, port 5201
[  4] local 192.168.0.84 port 37874 connected to 192.168.0.11 port 5201
[ ID] Interval           Transfer     Bandwidth       Retr  Cwnd
[  4]   0.00-1.00   sec   112 MBytes   940 Mbits/sec    0    392 KBytes
[  4]   1.00-2.00   sec   112 MBytes   941 Mbits/sec    0    392 KBytes
[  4]   2.00-3.00   sec   112 MBytes   941 Mbits/sec    0    392 KBytes
[  4]   3.00-4.00   sec   112 MBytes   941 Mbits/sec    0    392 KBytes
[  4]   4.00-5.00   sec   112 MBytes   941 Mbits/sec    0    392 KBytes
[  4]   5.00-6.00   sec   112 MBytes   942 Mbits/sec    0    392 KBytes
[  4]   6.00-7.00   sec   112 MBytes   941 Mbits/sec    0    392 KBytes
[  4]   7.00-8.00   sec   112 MBytes   941 Mbits/sec    0    392 KBytes
[  4]   8.00-9.00   sec   112 MBytes   941 Mbits/sec    0    392 KBytes
[  4]   9.00-10.00  sec   112 MBytes   941 Mbits/sec    1    352 KBytes
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval           Transfer     Bandwidth       Retr
[  4]   0.00-10.00  sec  1.10 GBytes   941 Mbits/sec    1             sender
[  4]   0.00-10.00  sec  1.09 GBytes   940 Mbits/sec                  receiver

iperf Done.
rockpi:

Code: Select all

opi@rockpi:~$ iperf3 -c 192.168.0.11
Connecting to host 192.168.0.11, port 5201
[  4] local 192.168.0.10 port 56214 connected to 192.168.0.11 port 5201
[ ID] Interval           Transfer     Bandwidth       Retr  Cwnd
[  4]   0.00-1.00   sec   113 MBytes   951 Mbits/sec    0    554 KBytes
[  4]   1.00-2.00   sec   112 MBytes   941 Mbits/sec    0    554 KBytes
[  4]   2.00-3.00   sec   112 MBytes   942 Mbits/sec    0    554 KBytes
[  4]   3.00-4.00   sec   112 MBytes   941 Mbits/sec    0    554 KBytes
[  4]   4.00-5.00   sec   112 MBytes   941 Mbits/sec    0    554 KBytes
[  4]   5.00-6.00   sec   112 MBytes   941 Mbits/sec    0    554 KBytes
[  4]   6.00-7.00   sec   112 MBytes   941 Mbits/sec    0    554 KBytes
[  4]   7.00-8.00   sec   112 MBytes   942 Mbits/sec    0    554 KBytes
[  4]   8.00-9.00   sec   112 MBytes   941 Mbits/sec    0    554 KBytes
[  4]   9.00-10.00  sec   112 MBytes   941 Mbits/sec    0    554 KBytes
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval           Transfer     Bandwidth       Retr
[  4]   0.00-10.00  sec  1.10 GBytes   942 Mbits/sec    0             sender
[  4]   0.00-10.00  sec  1.09 GBytes   940 Mbits/sec                  receiver

iperf Done.
last round tinker as a server:
rockpi:

Code: Select all

opi@rockpi:~$ iperf3 -c 192.168.0.84
Connecting to host 192.168.0.84, port 5201
[  4] local 192.168.0.10 port 36040 connected to 192.168.0.84 port 5201
[ ID] Interval           Transfer     Bandwidth       Retr  Cwnd
[  4]   0.00-1.00   sec   113 MBytes   945 Mbits/sec    0    424 KBytes
[  4]   1.00-2.00   sec   112 MBytes   942 Mbits/sec    0    424 KBytes
[  4]   2.00-3.00   sec   112 MBytes   942 Mbits/sec    0    424 KBytes
[  4]   3.00-4.00   sec   112 MBytes   941 Mbits/sec    0    424 KBytes
[  4]   4.00-5.00   sec   112 MBytes   941 Mbits/sec    0    424 KBytes
[  4]   5.00-6.00   sec   112 MBytes   942 Mbits/sec    0    424 KBytes
[  4]   6.00-7.00   sec   112 MBytes   941 Mbits/sec    0    424 KBytes
[  4]   7.00-8.00   sec   112 MBytes   941 Mbits/sec    0    424 KBytes
[  4]   8.00-9.00   sec   112 MBytes   942 Mbits/sec    0    424 KBytes
[  4]   9.00-10.00  sec   112 MBytes   941 Mbits/sec    0    424 KBytes
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval           Transfer     Bandwidth       Retr
[  4]   0.00-10.00  sec  1.10 GBytes   942 Mbits/sec    0             sender
[  4]   0.00-10.00  sec  1.09 GBytes   940 Mbits/sec                  receiver

iperf Done.
Odroid:

Code: Select all

opi@odroidxu4:~$ iperf3 -c 192.168.0.84
Connecting to host 192.168.0.84, port 5201
[  4] local 192.168.0.11 port 47646 connected to 192.168.0.84 port 5201
[ ID] Interval           Transfer     Bandwidth       Retr  Cwnd
[  4]   0.00-1.00   sec   109 MBytes   914 Mbits/sec    0    355 KBytes
[  4]   1.00-2.00   sec   110 MBytes   923 Mbits/sec    0    355 KBytes
[  4]   2.00-3.00   sec   110 MBytes   923 Mbits/sec    0    355 KBytes
[  4]   3.00-4.00   sec   110 MBytes   924 Mbits/sec    0    400 KBytes
[  4]   4.00-5.00   sec   110 MBytes   923 Mbits/sec    0    400 KBytes
[  4]   5.00-6.00   sec   110 MBytes   924 Mbits/sec    0    400 KBytes
[  4]   6.00-7.00   sec   110 MBytes   924 Mbits/sec    0    400 KBytes
[  4]   7.00-8.00   sec   110 MBytes   924 Mbits/sec    0    400 KBytes
[  4]   8.00-9.00   sec   110 MBytes   923 Mbits/sec    0    400 KBytes
[  4]   9.00-10.00  sec   110 MBytes   923 Mbits/sec    0    400 KBytes
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval           Transfer     Bandwidth       Retr
[  4]   0.00-10.00  sec  1.07 GBytes   923 Mbits/sec    0             sender
[  4]   0.00-10.00  sec  1.07 GBytes   922 Mbits/sec                  receiver

iperf Done.
semantics may matter here.. :lol:

sorry don't have a Pi3b+ to get it run under the same conditions... But we can take the numbers provided by @MuntyScruntfundle
MuntyScruntfundle wrote:
Sat Feb 09, 2019 9:01 pm
A simple iperf test one at a time to a central Windows PC running the Windows 10 unix install returns Bandwidth of 293 for the Rasp, 949 for the Tink.
Running both boards together is even more surprising, Rasp 87, Tink 910.
All those boards running an debian stretch based Linux with a mainline based kernel.. No tweaks just stock OS. RockPi and Odroid should be capable of saturating the link (HC1 due to SATA over USB3, RockPi either via USB3 or NVme. The tinker not fully (only USB2) but at least the SDIO interface for its SD card is a speedy one, so with a decent card you might get near to saturation:

Code: Select all

                                                              random    random     bkwd    record    stride                               
              kB  reclen    write  rewrite    read    reread    read     write     read   rewrite      read   fwrite frewrite    fread  freread
          102400       4     3494     3573    14684    14646    13265     5467
          102400      16     9693    10659    33987    33997    32392    14184
          102400     512    28791    43623    63744    63746    63246    49403
          102400    1024    59659    59934    64568    64577    64484    50461
          102400   16384    58269    60555    68443    68358    68443    45531
So, none of them hit 1000mbits/s then. As I said.

Everyone knows that the Pi is limited to its USB speed and that these devices are faster. That's not the point. The point is that all these devices are GigE, including the Pi, independent of the throughput.
Principal Software Engineer at Raspberry Pi (Trading) Ltd.
Contrary to popular belief, humorous signatures are allowed. Here's an example...
"My grief counseller just died, luckily, he was so good, I didn't care."

jahboater
Posts: 4686
Joined: Wed Feb 04, 2015 6:38 pm

Re: Next raspberry ethernet?

Tue Feb 12, 2019 8:24 am

jamesh wrote:
Tue Feb 12, 2019 8:04 am
So, none of them hit 1000mbits/s then. As I said.
There is always protocol overhead and gaps between packets.

Analogous to the way you never get 1GB of free disk space on a 1GB disk.

Incidentally, the Pi's "feel" faster when I use them over the network compared to the GigE Odroid C2 (and Gentoo64 on the Pi is fractionally faster than Raspian).
Unscientific test ...

Code: Select all

$ time scp pi@pi:makefile /tmp
makefile                                                          100%   13KB  13.0KB/s   00:00    

real	0m0.406s
user	0m0.016s
sys	0m0.000s
$ time scp odroid@c2:makefile /tmp
makefile                                                          100%   13KB  13.0KB/s   00:00    

real	0m1.262s
user	0m0.016s
sys	0m0.000s
$ time scp jah@pi64:makefile /tmp
makefile                                                          100%   13KB  13.0KB/s   00:00    

real	0m0.327s
user	0m0.016s
sys	0m0.000s
$ 
Both Pi's are 3B+ models. The Odroid C2 has a faster eMMC disk and more CPU power for encryption yet is 3x slower.
The C2 runs 64-bit Ubuntu.

chwe
Posts: 126
Joined: Tue Jul 31, 2018 1:35 pm

Re: Next raspberry ethernet?

Tue Feb 12, 2019 2:20 pm

jamesh wrote:
Tue Feb 12, 2019 8:04 am
So, none of them hit 1000mbits/s then. As I said.

Everyone knows that the Pi is limited to its USB speed and that these devices are faster. That's not the point. The point is that all these devices are GigE, including the Pi, independent of the throughput.
C'mon.. there's a significant difference between roughly 5% less and 66% less than theoretical bandwith.. 5% means that even without flow-control packets drop usually don't happen.. 66% means that without flow-control you better limit it to fast ethernet..
jahboater wrote:
Tue Feb 12, 2019 8:24 am
Incidentally, the Pi's "feel" faster when I use them over the network compared to the GigE Odroid C2 (and Gentoo64 on the Pi is fractionally
'feelings' are for your private time with your wife, girlfriend, boyfriend, husband whatsoever.. For computer related stuff there are normally benchmarks (hopefully done by people knowing what they're doing).
jahboater wrote:
Tue Feb 12, 2019 8:24 am
more CPU power for encryption yet is 3x slower.
The C2 runs 64-bit Ubuntu.
Out of all SoCs on SBCs I've in mind you went for the only two boring ones when it comes to crypto.. Odroid and RPi3 are the only 64-bit SoCs without dedicated cryptoengine.. simply the wrong tool for the job.
Ernst wrote:
Tue Feb 12, 2019 7:41 am
Can you explain why the results of these tests are so important to you ? What is the significance of getting near to saturation ?
first I don't have to deal with flow-control due to near to max.. Second there's a huge difference if my NAS is able to delvier and accept ~80-90MB/s (with some overhead etc.) or ~20MB/s at absolute maximum (which you likely never reach with an RPi). I have my own little git-server with rather larger repositories on it (e.g. a full kernel repo etc.). It wouldn't make sense to host my own git-server on a RPi (git-hub delivers similar fast than an RPi could deliver and my internet connection is fast enough), whereas the tinker is really sweet here.. I opted for a notebook with enough ram and good crunching power rather than a big SSD in it.. Windows is a must due to a few programs which don't work flawlessly on ubuntu and linux on it is a second must - seriously nobody ever wants to deal with windows when it comes to stuff like python or so.. Keeping music and all my repos the whole time on my notebook isn't an option unsing external drives neither (the tinker will sync those repos automatically every day.. whereas the external drive wont... some of my machines gets a weekly backup (including my notebook).. I have absolutely no patience for network as a bottleneck.. especially if there are solutions out there.. The HC1 was roughly 60$ including shipping and it does a great job as a NAS.. Tinkers where really cheap here after 2-3 months (likely cause software support was similar to software support of the RPi in the beginnings)... So I got one for ~40$ compared to the ~35$ for the RPi, obviously the tinker was the better bet..

jamesh
Raspberry Pi Engineer & Forum Moderator
Raspberry Pi Engineer & Forum Moderator
Posts: 23683
Joined: Sat Jul 30, 2011 7:41 pm

Re: Next raspberry ethernet?

Tue Feb 12, 2019 3:01 pm

We know the device ethernet is limited by the USB bus, and of course there is a difference bertween the THROUGHPUT of the Pi and that of other devices. But that does not negate the FACT that they are GigE connections.

I'm not sure why you cannot understand that, and are continuing to flog a dead horse, and since this thread is going nowhere, and you have started getting personal (girlfriends? Really?), I'm closing it.
Principal Software Engineer at Raspberry Pi (Trading) Ltd.
Contrary to popular belief, humorous signatures are allowed. Here's an example...
"My grief counseller just died, luckily, he was so good, I didn't care."

Return to “General discussion”