USB 3.0 Speeds Dropped 1/4, from 20 to 4.3 mbps

[dugaduga]

AC66u B1 as 68u USB 3.0 SanDisk Cruzer Fit CZ33 16GB Mini Nano USB Flash Pen Drive Memory Thumb Stick; USB 3.0 port. EXT2 formatted with Minitool Partition Manager 8 @ Default cluster size, 1024?. So 4.9 mpbs read / 2.5 mpbs write random, sub usb2.0 speeds. This is what I get with sustained writes/reads in winscp. I originally got between 15-20 mbps copying files over winscp, not sure whats slowed it down so much. Then USB 3.0 failed to detect the partitions so I switched to USB 2.0 until yesterday. I reformatted and successfully switched to USB 3.0, with zero improvement over USB 2.0 and ~1/4 speeds of what I had originally. Avg Sustained write benches online also show the same 15-20mbps. I did switch to EXT2 from EXT3 this time around citing suggestions on the forum. Drive is running (usb3.0) 5000

Command line used: /jffs/bin/iozone -e -I -a -s 20M -r 4k -r 16k -r 512k -r 1024k -r 16384k -i 0 -i 1 -i 2
   Output is in kBytes/sec
   Time Resolution = 0.000001 seconds.
   Processor cache size set to 1024 kBytes.
   Processor cache line size set to 32 bytes.
   File stride size set to 17 * record size.
                                                              random    random     bkwd    record    stride
              kB  reclen    write  rewrite    read    reread    read     write     read   rewrite      read   fwrite frewrite    fread  freread
           20480       4     2574     2793    11280    11128     4927     2502
           20480      16     7380     7499    29968    30604    16727     7064
           20480     512    29891    30415    75050    75729    75953    29677
           20480    1024    29846    30845    77264    78043    77517    31046
           20480   16384    26786    32590    74559    73010    57322    27655

 

[ApexRon]

AC66u B1 as 68u USB 3.0 SanDisk Cruzer Fit CZ33 16GB Mini Nano USB Flash Pen Drive Memory Thumb Stick; USB 3.0 port. EXT2 formatted with Minitool Partition Manager 8 @ Default cluster size, 1024?. So 4.9 mpbs read / 2.5 mpbs write random, sub usb2.0 speeds. This is what I get with sustained writes/reads in winscp. I originally got between 15-20 mbps copying files over winscp, not sure whats slowed it down so much. Then USB 3.0 failed to detect the partitions so I switched to USB 2.0 until yesterday. I reformatted and successfully switched to USB 3.0, with zero improvement over USB 2.0 and ~1/4 speeds of what I had originally. Avg Sustained write benches online also show the same 15-20mbps. I did switch to EXT2 from EXT3 this time around citing suggestions on the forum. Drive is running (usb3.0) 5000

Command line used: /jffs/bin/iozone -e -I -a -s 20M -r 4k -r 16k -r 512k -r 1024k -r 16384k -i 0 -i 1 -i 2
   Output is in kBytes/sec
   Time Resolution = 0.000001 seconds.
   Processor cache size set to 1024 kBytes.
   Processor cache line size set to 32 bytes.
   File stride size set to 17 * record size.
                                                              random    random     bkwd    record    stride
              kB  reclen    write  rewrite    read    reread    read     write     read   rewrite      read   fwrite frewrite    fread  freread
           20480       4     2574     2793    11280    11128     4927     2502
           20480      16     7380     7499    29968    30604    16727     7064
           20480     512    29891    30415    75050    75729    75953    29677
           20480    1024    29846    30845    77264    78043    77517    31046
           20480   16384    26786    32590    74559    73010    57322    27655

This is just a comment:

In the design of a microprocessor driven widget that has multithread capability consideration should be made to what the design goals of the widget are.

In the case of a laptop I would think fast I/O would be important; memory being number 1, then say hard drives number 2, and so on down the line. The lower the I/O is in importance the more buffering could be used thus performance and speed may suffer.

In the case of a router I would think WAN, LAN, and WiFi would be the most important, so data in and out of those interfaces would be number 1. I would expect USB ports to play second fiddle to anything else going on in the router, so data in and out of USB port could be highly buffered and measured transfer rates be erratic and at less than maximum speed.

Just my opinion​

 

[dugaduga]

To sum up, "In the case of a router I would think WAN, LAN, and WiFi would be the most important" so "I would expect USB ports" "be erratic and at less than maximum speed."

Ok then. Personally I would prefer if everything operated at spec and at maximum potential, as it once did. Well anybody else know why these speeds are running at 1/4?