New UMASS driver available for testing

[Ulrich Spoerlein]

Ulrich Spoerlein wrote:
> 258998272 bytes transferred in 1771.034516 secs (146241 bytes/sec)
> dd if=/dev/da0 of=/dev/zero bs=$i  0.20s user 3.10s system 0% cpu 29:31.05 total
> 258998272 bytes transferred in 775.213552 secs (334099 bytes/sec)
> dd if=/dev/da0 of=/dev/zero bs=$i  0.01s user 0.41s system 0% cpu 12:55.23 total
> 258998272 bytes transferred in 650.920062 secs (397896 bytes/sec)
> dd if=/dev/da0 of=/dev/zero bs=$i  0.00s user 0.07s system 0% cpu 10:50.93 total
> 258998272 bytes transferred in 641.615133 secs (403666 bytes/sec)
> dd if=/dev/da0 of=/dev/zero bs=$i  0.00s user 0.03s system 0% cpu 10:41.63 total
> 258998272 bytes transferred in 641.612930 secs (403667 bytes/sec)
> dd if=/dev/da0 of=/dev/zero bs=$i  0.00s user 0.04s system 0% cpu 10:41.63 total

I tested the same USB stick on a Gentoo/Linux machine, though it is
roughly twice as fast as my own machine:


Linux DVPC03 2.6.17-gentoo-r4 #1 SMP PREEMPT Mon Aug 14 17:19:40 CEST 2006 i686 Intel(R) Celeron(R) CPU 3.06GHz GNU/Linux

Oct  5 11:19:45 localhost usb 4-2: new full speed USB device using ohci_hcd and address 2
Oct  5 11:19:45 localhost usb 4-2: configuration #1 chosen from 1 choice
Oct  5 11:19:45 localhost scsi4 : SCSI emulation for USB Mass Storage devices
Oct  5 11:19:45 localhost usb-storage: device found at 2
Oct  5 11:19:45 localhost usb-storage: waiting for device to settle before scanning
Oct  5 11:19:50 localhost Vendor:           Model: MP3 Flash Stick   Rev: 1.00
Oct  5 11:19:50 localhost Type:   Direct-Access                      ANSI SCSI revision: 00
Oct  5 11:19:50 localhost SCSI device sda: 505856 512-byte hdwr sectors (259 MB)
Oct  5 11:19:50 localhost sda: Write Protect is off
Oct  5 11:19:50 localhost sda: Mode Sense: 20 00 00 00
Oct  5 11:19:50 localhost sda: assuming drive cache: write through
Oct  5 11:19:50 localhost SCSI device sda: 505856 512-byte hdwr sectors (259 MB)
Oct  5 11:19:50 localhost sda: Write Protect is off
Oct  5 11:19:50 localhost sda: Mode Sense: 20 00 00 00
Oct  5 11:19:50 localhost sda: assuming drive cache: write through
Oct  5 11:19:50 localhost sda:
Oct  5 11:19:50 localhost sd 4:0:0:0: Attached scsi removable disk sda
Oct  5 11:19:50 localhost usb-storage: device scan complete
Oct  5 11:19:50 localhost scsi.agent[2225]: disk at /devices/pci0000:00/0000:00:03.2/usb4/4-2/4-2:1.0/host4/target4:0:0/4:0:0:0


for i in 512 2k 32k 256k 2048k; do time dd if=/dev/sda of=/dev/null bs=$i;done
258998272 bytes (259 MB) copied, 648.34 seconds, 399 kB/s
dd if=/dev/sda of=/dev/null bs=$i  0.14s user 0.84s system 0% cpu 10:48.35 total
258998272 bytes (259 MB) copied, 648.347 seconds, 399 kB/s
dd if=/dev/sda of=/dev/null bs=$i  0.04s user 0.65s system 0% cpu 10:48.35 total
258998272 bytes (259 MB) copied, 648.452 seconds, 399 kB/s
dd if=/dev/sda of=/dev/null bs=$i  0.02s user 0.65s system 0% cpu 10:48.46 total
258998272 bytes (259 MB) copied, 647.917 seconds, 400 kB/s
dd if=/dev/sda of=/dev/null bs=$i  0.00s user 0.74s system 0% cpu 10:47.92 total
258998272 bytes (259 MB) copied, 648.605 seconds, 399 kB/s
dd if=/dev/sda of=/dev/null bs=$i  0.00s user 0.78s system 0% cpu 10:48.75 total

So, the 400kB/s limit is entirely due to the crappy hardware. I wonder
though, why is Linux able to achieve 400kB/s with 512 bytes reads? Is it
cheating here?

Ulrich Spoerlein
-- 
A: Yes.
>Q: Are you sure?
> >A: Because it reverses the logical flow of conversation.
> >>Q: Why is top posting frowned upon?