Re: very, very slow usb data transfer speed on 3.7
Andreas Bihlmaier wrote: -- [EMAIL PROTECTED]: ~ (17:39) -- # dd if=/dev/wd0c of=/dev/null bs=819200 count=20 20+0 records in 20+0 records out 16384000 bytes transferred in 0.711 secs (23012820 bytes/sec) recall the old speed with apm0: - -- [EMAIL PROTECTED]: ~ (17:13) -- # dd if=/dev/wd0c of=/dev/null bs=819200 count=20 20+0 records in 20+0 records out 16384000 bytes transferred in 1.129 secs (14509606 bytes/sec) that's pretty harsh if other people can reproduce it. :( Incidentally usb transfers *weren't* improved by removing apm0 - -- [EMAIL PROTECTED]: ~ (17:39) -- # dd if=/dev/sd0c of=/dev/null bs=819200 count=20 20+0 records in 20+0 records out 16384000 bytes transferred in 6.017 secs (2722653 bytes/sec) so there's some other factor limiting those. I'm getting the same speed on a snapshot from 09/21 with amd64 on a brand new amd 64 3800+. Lately I was copying around 40G of data onto a usb 2.0 hard disk (yes it was attached to EHCI) and wondered why it took so long, but I didn't pursue the issue further. I also tried with different blocksizes and to eliminate the issue of a too short benchmark I ran for a couple of minutes (about count=500). Actually wait a minute ... /dev/sd0c and /dev/wd0c ? Are you SUPPOSED to read of a block device ? SHOULDN'T it be /dev/rsd0c and /dev/rwd0c ??? ^ ^ RAW DEVICE With the raw devices the speed looks QUITE different: BLOCK DEVICE: sudo dd if=/dev/wd0c of=/dev/null bs=512k count=500 500+0 records in 500+0 records out 262144000 bytes transferred in 16.957 secs (15458831 bytes/sec) # Top shows CPU usage as 28.7% system, 27.9% interrupt, 41.9% idle RAW DEVICE: sudo dd if=/dev/rwd0c of=/dev/null bs=512k count=200 200+0 records in 200+0 records out 104857600 bytes transferred in 1.787 secs (58666485 bytes/sec) # Top shows CPU usage as 4.6% system, 5.4% interrupt, 90.0% idle (same with USB device) BLOCK DEVICE: sudo dd if=/dev/sd0c of=/dev/null bs=512k count=500 500+0 records in 500+0 records out 262144000 bytes transferred in 96.561 secs (2714791 bytes/sec) # Top shows CPU usage as 4.7% system, 10.8% interrupt, 84.6% idle RAW DEVICE: sudo dd if=/dev/rsd0c of=/dev/null bs=512k count=500 500+0 records in 500+0 records out 262144000 bytes transferred in 19.015 secs (13785462 bytes/sec) # Top shows CPU usage as 1.6% system, 2.3% interrupt, 96.1% idle I'm sorry if I understood something wrong, but my understanding was/is that you only use RAW devices with dd (since it uses it's own blocks ). Please tell me if I'm wrong, since (right) knowledge is valueable! I think you're right. I've been in the FreeBSD/DragonFly world for about 11 years now and about 5 years ago FreeBSD ripped out block device support so that block devices are now really just character devices. From your observations in OpenBSD block devices are clearly still block devices ;) .. Since I've only being using it for about a week now I can still claim to be a newbie on OpenBSD. :O but not for much longer. :) Incidentally, I now get (DragonFly) - # dd if=/dev/da0 of=/dev/null bs=8192000 count=10 10+0 records in 10+0 records out 8192 bytes transferred in 5.940737 secs (13789535 bytes/sec) and on OpenBSD 3.7-stable I get - # dd if=/dev/rsd0c of=/dev/null bs=8192000 count=10 10+0 records in 10+0 records out 8192 bytes transferred in 3.798 secs (21563802 bytes/sec) which is quite impressive! Woohoo! Thanks Andreas! Andrew.
Re: very, very slow usb data transfer speed on 3.7
On Tue, 11 Oct 2005, Andrew Atrens wrote: I don't see the 'EHCI' controller in there anywhere. :( UHCI == usb1.1 EHCI == usb2.0 Top speed for usb1.1 is roughly 1MB/s. Your getting that. :| Two possibilities - your mobo doesn't do usb2.0 - or the ehci device probe isn't grokking your hardware. Andrew. Thank you Andrew and Peter for the explanation of why that drive is so slow. I inherited it at work. It's also got a firewire interface, so perhaps when firewire support comes along it will actually be usable. Jeff Jeff Ross wrote: I've got a USB external drive that is virtually unusable because it is so slow. mount dev/sd0a on / type ffs (local) /dev/sd0p on /backup type ffs (local, nodev, nosuid, softdep) /dev/sd0o on /destdir type ffs (local, nodev, nosuid, softdep) /dev/sd0d on /home type ffs (local, nodev, nosuid, softdep) /dev/sd0n on /releasedir type ffs (local, nodev, nosuid, softdep) /dev/sd0e on /tmp type ffs (local, nodev, nosuid, softdep) /dev/sd0f on /usr type ffs (local, nodev, softdep) /dev/sd0g on /usr/X11R6 type ffs (local, nodev, softdep) /dev/sd0h on /usr/local type ffs (local, nodev, softdep) /dev/sd0i on /cvs type ffs (local, nodev, nosuid, softdep) /dev/sd0j on /usr/src type ffs (local, nodev, nosuid, softdep) /dev/sd0k on /var type ffs (local, nodev, nosuid, softdep) /dev/sd0l on /var/qmail/bin type ffs (local, nodev, softdep) /dev/sd0m on /var/qmail/queue type ffs (local, nodev, nosuid) /dev/sd1a on /log type ffs (local, nodev, nosuid, softdep) /dev/sd1d on /offline type ffs (local, nodev, nosuid, softdep) /dev/sd1e on /wal type ffs (local, nodev, nosuid, softdep) mfs:13470 on /var/mfs type mfs (asynchronous, local, size=200 512-blocks) /dev/sd2a on /usb_drive type ffs (local) (write a file to the usb drive) time sudo dd if=/dev/zero of=/usb_drive/test_file count=100 100+0 records in 100+0 records out 51200 bytes transferred in 452.234 secs (1132156 bytes/sec) 7m32.69s real 0m0.51s user 0m3.88s system (write a file to the crappy 3WARE RAID1) time sudo dd if=/dev/zero of=/offline/test_file count=100 100+0 records in 100+0 records out 51200 bytes transferred in 9.298 secs (55064036 bytes/sec) 0m13.68s real 0m0.58s user 0m3.78s system dmesg: OpenBSD 3.8-current (GENERIC) #0: Wed Sep 14 22:05:15 MDT 2005 [EMAIL PROTECTED]:/usr/src/sys/arch/i386/compile/GENERIC cpu0: Intel(R) Xeon(TM) CPU 2.66GHz (GenuineIntel 686-class) 2.67 GHz cpu0: FPU,V86,DE,PSE,TSC,MSR,PAE,MCE,CX8,APIC,SEP,MTRR,PGE,MCA,CMOV,PAT,PSE36,CFLUSH,ACPI,MMX,FXSR,SSE,SSE2,SS,HTT,TM,SBF,CNXT-ID real mem = 2147000320 (2096680K) avail mem = 1953087488 (1907312K) using 4278 buffers containing 107454464 bytes (104936K) of memory mainbus0 (root) bios0 at mainbus0: AT/286+(00) BIOS, date 02/04/03, BIOS32 rev. 0 @ 0xf0010 pcibios0 at bios0: rev 2.1 @ 0xf/0x1 pcibios0: PCI IRQ Routing Table rev 1.0 @ 0xf2fb0/256 (14 entries) pcibios0: PCI Interrupt Router at 000:31:0 (Intel 82801CA LPC rev 0x00) pcibios0: PCI bus #4 is the last bus bios0: ROM list: 0xc/0x8000 0xc8000/0x1000 0xc9800/0x800 0xca000/0x1800 cpu0 at mainbus0 pci0 at mainbus0 bus 0: configuration mode 1 (no bios) pchb0 at pci0 dev 0 function 0 Intel E7501 MCH Host rev 0x01 ppb0 at pci0 dev 2 function 0 Intel E7500 MCH rev 0x01 pci1 at ppb0 bus 1 Intel 82870P2 IOxAPIC rev 0x04 at pci1 dev 28 function 0 not configured ppb1 at pci1 dev 29 function 0 Intel 82870P2 PCI-PCI rev 0x04 pci2 at ppb1 bus 2 em0 at pci2 dev 1 function 0 Intel PRO/1000MT (82545EM) rev 0x01: irq 10, address: 00:e0:81:28:e9:71 Intel 82870P2 IOxAPIC rev 0x04 at pci1 dev 30 function 0 not configured ppb2 at pci1 dev 31 function 0 Intel 82870P2 PCI-PCI rev 0x04 pci3 at ppb2 bus 3 ahc1 at pci3 dev 3 function 0 Adaptec AHA-29160 U160 rev 0x02: irq 10 scsibus0 at ahc1: 16 targets st0 at scsibus0 targ 6 lun 0: SEAGATE, DAT 9SP40-000, 910B SCSI3 1/sequential removable st0: density code 0x26, 512-byte blocks, write-enabled twe0 at pci3 dev 6 function 0 3ware Escalade IDE RAID rev 0x01: irq 10 twe0: Escalade V1.3 scsibus1 at twe0: 16 targets sd0 at scsibus1 targ 0 lun 0: 3WARE, Host drive #00, SCSI2 0/direct fixed sd0: 117799MB, 15017 cyl, 255 head, 63 sec, 512 bytes/sec, 241252672 sec total sd1 at scsibus1 targ 2 lun 0: 3WARE, Host drive #02, SCSI2 0/direct fixed sd1: 117799MB, 15017 cyl, 255 head, 63 sec, 512 bytes/sec, 241252672 sec total uhci0 at pci0 dev 29 function 0 Intel 82801CA/CAM USB rev 0x02: irq 10 usb0 at uhci0: USB revision 1.0 uhub0 at usb0 uhub0: Intel UHCI root hub, rev 1.00/1.00, addr 1 uhub0: 2 ports with 2 removable, self powered uhci1 at pci0 dev 29 function 1 Intel 82801CA/CAM USB rev 0x02: irq 9 usb1 at uhci1: USB revision 1.0 uhub1 at usb1 uhub1: Intel UHCI root hub, rev 1.00/1.00, addr 1 uhub1: 2 ports with 2 removable, self powered uhci2 at pci0 dev 29 function 2 Intel 82801CA/CAM USB rev 0x02:
very, very slow usb data transfer speed on 3.7
Hey guys! I'm running 3.7 and am getting really, really crappy usb throughput :( - # dd if=/dev/sd0c of=/dev/null bs=819200 count=100 100+0 records in 100+0 records out 8192 bytes transferred in 175.970 secs (465533 bytes/sec) Here's an excerpt from my dmesg - uhci0 at pci0 dev 29 function 0 Intel 82801DB USB rev 0x01: irq 11 usb0 at uhci0: USB revision 1.0 uhub0 at usb0 uhub0: Intel UHCI root hub, class 9/0, rev 1.00/1.00, addr 1 uhub0: 2 ports with 2 removable, self powered uhci1 at pci0 dev 29 function 1 Intel 82801DB USB rev 0x01: irq 10 usb1 at uhci1: USB revision 1.0 uhub1 at usb1 uhub1: Intel UHCI root hub, class 9/0, rev 1.00/1.00, addr 1 uhub1: 2 ports with 2 removable, self powered uhci2 at pci0 dev 29 function 2 Intel 82801DB USB rev 0x01: irq 9 usb2 at uhci2: USB revision 1.0 uhub2 at usb2 uhub2: Intel UHCI root hub, class 9/0, rev 1.00/1.00, addr 1 uhub2: 2 ports with 2 removable, self powered ehci0 at pci0 dev 29 function 7 Intel 82801DB USB rev 0x01: irq 3 ehci0: EHCI version 1.0 ehci0: companion controllers, 2 ports each: uhci0 uhci1 uhci2 usb3 at ehci0: USB revision 2.0 uhub3 at usb3 uhub3: Intel EHCI root hub, class 9/0, rev 2.00/1.00, addr 1 uhub3: single transaction translator uhub3: 6 ports with 6 removable, self powered ... umass0 at uhub3 port 2 configuration 1 interface 0 umass0: vendor 0x0457 USB Mass Storage Device, rev 2.00/1.00, addr 2 umass0: using SCSI over Bulk-Only scsibus1 at umass0: 2 targets sd0 at scsibus1 targ 1 lun 0: OCZ, ET1208AD, 1.0 SCSI2 0/direct removable sd0: 2000MB, 2000 cyl, 64 head, 32 sec, 512 bytes/sec, 4096000 sec total Looks like sd0 is attached to the EHCI controller. On DragonflyBSD, same h/w, I get 13MB/s when I use ehci .. -- [EMAIL PROTECTED]: /usr/src/sys/i386/conf (11:09) -- # dd if=/dev/da0 of=/dev/null bs=8192000 count=100 100+0 records in 100+0 records out 81920 bytes transferred in 59.371990 secs (13797752 bytes/sec) when I use uhci on Dragonfly, I get - -- [EMAIL PROTECTED]: /home/atrens (11:07) -- # dd if=/dev/da0 bs=8192000 of=/dev/null count=100 ^C5+0 records in 5+0 records out 4096 bytes transferred in 45.750635 secs (895288 bytes/sec) which is still double what I'm seeing on OpenBSD 3.7 Hope it's something dumb on my side (and therefore easy to fix) :( ... Andrew.
Re: very, very slow usb data transfer speed on 3.7
Yes dumb. Where are the whole dmesg? If you had sent them we could have told you if you ran into the hlt hlt bug. On Tue, Oct 11, 2005 at 12:20:45PM -0400, Andrew Atrens wrote: Hey guys! I'm running 3.7 and am getting really, really crappy usb throughput :( - # dd if=/dev/sd0c of=/dev/null bs=819200 count=100 100+0 records in 100+0 records out 8192 bytes transferred in 175.970 secs (465533 bytes/sec) Here's an excerpt from my dmesg - uhci0 at pci0 dev 29 function 0 Intel 82801DB USB rev 0x01: irq 11 usb0 at uhci0: USB revision 1.0 uhub0 at usb0 uhub0: Intel UHCI root hub, class 9/0, rev 1.00/1.00, addr 1 uhub0: 2 ports with 2 removable, self powered uhci1 at pci0 dev 29 function 1 Intel 82801DB USB rev 0x01: irq 10 usb1 at uhci1: USB revision 1.0 uhub1 at usb1 uhub1: Intel UHCI root hub, class 9/0, rev 1.00/1.00, addr 1 uhub1: 2 ports with 2 removable, self powered uhci2 at pci0 dev 29 function 2 Intel 82801DB USB rev 0x01: irq 9 usb2 at uhci2: USB revision 1.0 uhub2 at usb2 uhub2: Intel UHCI root hub, class 9/0, rev 1.00/1.00, addr 1 uhub2: 2 ports with 2 removable, self powered ehci0 at pci0 dev 29 function 7 Intel 82801DB USB rev 0x01: irq 3 ehci0: EHCI version 1.0 ehci0: companion controllers, 2 ports each: uhci0 uhci1 uhci2 usb3 at ehci0: USB revision 2.0 uhub3 at usb3 uhub3: Intel EHCI root hub, class 9/0, rev 2.00/1.00, addr 1 uhub3: single transaction translator uhub3: 6 ports with 6 removable, self powered ... umass0 at uhub3 port 2 configuration 1 interface 0 umass0: vendor 0x0457 USB Mass Storage Device, rev 2.00/1.00, addr 2 umass0: using SCSI over Bulk-Only scsibus1 at umass0: 2 targets sd0 at scsibus1 targ 1 lun 0: OCZ, ET1208AD, 1.0 SCSI2 0/direct removable sd0: 2000MB, 2000 cyl, 64 head, 32 sec, 512 bytes/sec, 4096000 sec total Looks like sd0 is attached to the EHCI controller. On DragonflyBSD, same h/w, I get 13MB/s when I use ehci .. -- [EMAIL PROTECTED]: /usr/src/sys/i386/conf (11:09) -- # dd if=/dev/da0 of=/dev/null bs=8192000 count=100 100+0 records in 100+0 records out 81920 bytes transferred in 59.371990 secs (13797752 bytes/sec) when I use uhci on Dragonfly, I get - -- [EMAIL PROTECTED]: /home/atrens (11:07) -- # dd if=/dev/da0 bs=8192000 of=/dev/null count=100 ^C5+0 records in 5+0 records out 4096 bytes transferred in 45.750635 secs (895288 bytes/sec) which is still double what I'm seeing on OpenBSD 3.7 Hope it's something dumb on my side (and therefore easy to fix) :( ... Andrew.
Re: very, very slow usb data transfer speed on 3.7
Just for fun I ran 'top' during the said slow transfer, and it says - load averages: 0.55, 0.20, 0.12 11:41:59 22 processes: 21 idle, 1 on processor CPU states: 0.2% user, 0.0% nice, 0.2% system, 0.8% interrupt, 98.9% idle Memory: Real: 8932K/100M act/tot Free: 904M Swap: 0K/2048M used/tot Can someone point me to the cvs commit that fixes 'hlt hlt'. I'm thinking (hoping) it could easily be applied on top of 3.7 Release. Andrew.
Re: very, very slow usb data transfer speed on 3.7
--On 11 October 2005 12:39 -0400, Andrew Atrens wrote: Can someone point me to the cvs commit that fixes 'hlt hlt'. I'm thinking (hoping) it could easily be applied on top of 3.7 Release. google hlt hlt openbsd gives this: http://marc.theaimsgroup.com/?l=openbsd-cvsm=111859519015510w=2
Re: very, very slow usb data transfer speed on 3.7
On Tue, 11 Oct 2005 12:30:30 -0400 Andrew Atrens [EMAIL PROTECTED] wrote: Marco Peereboom wrote: Yes dumb. Where are the whole dmesg? haha! If you had sent them we could have told you if you ran into the hlt hlt bug. Sure here it is - -- [EMAIL PROTECTED]: /usr/src/sys/arch/i386/conf (11:35) -- # dmesg OpenBSD 3.7 (BOOKEND) #0: Wed Oct 5 14:02:08 EST 2005 [EMAIL PROTECTED]:/usr/src/sys/arch/i386/compile/BOOKEND How about trying it with a GENRIC kernel? snip On Tue, Oct 11, 2005 at 12:20:45PM -0400, Andrew Atrens wrote: Hey guys! I'm running 3.7 and am getting really, really crappy usb throughput :( - # dd if=/dev/sd0c of=/dev/null bs=819200 count=100 100+0 records in 100+0 records out 8192 bytes transferred in 175.970 secs (465533 bytes/sec) Here's an excerpt from my dmesg - uhci0 at pci0 dev 29 function 0 Intel 82801DB USB rev 0x01: irq 11 usb0 at uhci0: USB revision 1.0 uhub0 at usb0 uhub0: Intel UHCI root hub, class 9/0, rev 1.00/1.00, addr 1 uhub0: 2 ports with 2 removable, self powered uhci1 at pci0 dev 29 function 1 Intel 82801DB USB rev 0x01: irq 10 usb1 at uhci1: USB revision 1.0 uhub1 at usb1 uhub1: Intel UHCI root hub, class 9/0, rev 1.00/1.00, addr 1 uhub1: 2 ports with 2 removable, self powered uhci2 at pci0 dev 29 function 2 Intel 82801DB USB rev 0x01: irq 9 usb2 at uhci2: USB revision 1.0 uhub2 at usb2 uhub2: Intel UHCI root hub, class 9/0, rev 1.00/1.00, addr 1 uhub2: 2 ports with 2 removable, self powered ehci0 at pci0 dev 29 function 7 Intel 82801DB USB rev 0x01: irq 3 ehci0: EHCI version 1.0 ehci0: companion controllers, 2 ports each: uhci0 uhci1 uhci2 usb3 at ehci0: USB revision 2.0 uhub3 at usb3 uhub3: Intel EHCI root hub, class 9/0, rev 2.00/1.00, addr 1 uhub3: single transaction translator uhub3: 6 ports with 6 removable, self powered ... umass0 at uhub3 port 2 configuration 1 interface 0 umass0: vendor 0x0457 USB Mass Storage Device, rev 2.00/1.00, addr 2 umass0: using SCSI over Bulk-Only scsibus1 at umass0: 2 targets sd0 at scsibus1 targ 1 lun 0: OCZ, ET1208AD, 1.0 SCSI2 0/direct removable sd0: 2000MB, 2000 cyl, 64 head, 32 sec, 512 bytes/sec, 4096000 sec total Looks like sd0 is attached to the EHCI controller. On DragonflyBSD, same h/w, I get 13MB/s when I use ehci .. -- [EMAIL PROTECTED]: /usr/src/sys/i386/conf (11:09) -- # dd if=/dev/da0 of=/dev/null bs=8192000 count=100 100+0 records in 100+0 records out 81920 bytes transferred in 59.371990 secs (13797752 bytes/sec) when I use uhci on Dragonfly, I get - -- [EMAIL PROTECTED]: /home/atrens (11:07) -- # dd if=/dev/da0 bs=8192000 of=/dev/null count=100 ^C5+0 records in 5+0 records out 4096 bytes transferred in 45.750635 secs (895288 bytes/sec) which is still double what I'm seeing on OpenBSD 3.7 Hope it's something dumb on my side (and therefore easy to fix) :( ... Andrew. -- Security is decided by quality -- Theo de Raadt
Re: very, very slow usb data transfer speed on 3.7
Making, drinking tea and reading an opus magnum from Andrew Atrens: [Charset ISO-8859-1 unsupported, filtering to ASCII...] Okay, I've upgraded to a kernel with tag=OPENBSD_3_7 which looks to be the -stable or 'patch' tag, and while the situation has improved, performance is still off by a factor of 5. speaking about your ide benchmarks ... you cann seriously oncsider any measurment that too ~1sec try running it for at least 10 seconds (I added the wd test below because with the 3.7 kernel it (interestingly) matched that of the flash stick. Now with the -stable kernel wd is performing better, as is the flash stick, but both are still too slow. ) -- [EMAIL PROTECTED]: ~ (17:13) -- # dd if=/dev/wd0c of=/dev/null bs=819200 count=20 20+0 records in 20+0 records out 16384000 bytes transferred in 1.129 secs (14509606 bytes/sec) -- [EMAIL PROTECTED]: ~ (17:14) -- # dd if=/dev/sd0c of=/dev/null bs=819200 count=20 20+0 records in 20+0 records out 16384000 bytes transferred in 6.017 secs (2722767 bytes/sec) On DragonFly-Stable for comparisons, the flash stick is *fast* - -- [EMAIL PROTECTED]: /usr/src/sys/compile/AB-MOBILE-FAST_IPSEC (18:11) -- # dd if=/dev/da0 of=/dev/null bs=819200 count=20 20+0 records in 20+0 records out 16384000 bytes transferred in 1.223731 secs (13388563 bytes/sec) I have 3 identical boxen here (they're Dell GX240's). For fun I just bounced onto the third box (FreeBSD 4.9) and tried the ata test - -- [EMAIL PROTECTED]: /home/atrens (18:24) -- # dd if=/dev/ad0 of=/dev/null bs=819200 count=20 20+0 records in 20+0 records out 16384000 bytes transferred in 0.336737 secs (48655194 bytes/sec) I'm not sure what this means, because we're now comparing completely different ata subsystems - but the FreeBSD ata subsystem looks to be 3x faster than 3.7. This is the ata disk in question, btw - wd0 at pciide0 channel 0 drive 0: Maxtor 6Y080L0 wd0: 16-sector PIO, LBA, 78167MB, 160086528 sectors wd0(pciide0:0:0): using PIO mode 4, Ultra-DMA mode 5 I'm curious what performance other people are getting with this disk... :) since both subsystems (scsi/usb and ide) have both linearly improved, but, let's be a bit fuzzy and say, are both still off where they should be by *roughly* a factor of 4. Andrew. -- paranoic mickey (my employers have changed but, the name has remained)
Re: very, very slow usb data transfer speed on 3.7
Michael Shalayeff wrote: Making, drinking tea and reading an opus magnum from Andrew Atrens: [Charset ISO-8859-1 unsupported, filtering to ASCII...] Okay, I've upgraded to a kernel with tag=OPENBSD_3_7 which looks to be the -stable or 'patch' tag, and while the situation has improved, performance is still off by a factor of 5. speaking about your ide benchmarks ... you cann seriously oncsider any measurment that too ~1sec try running it for at least 10 seconds Agreed. But I was lazy. Here ya go, kernel without apm0: - -- [EMAIL PROTECTED]: ~ (17:56) -- # dd if=/dev/wd0c of=/dev/null bs=819200 count=400 400+0 records in 400+0 records out 32768 bytes transferred in 14.244 secs (23004549 bytes/sec) Note the consistency with the other numbers. I ran these a number of times, btw. :) :) :) Andrew.
Re: very, very slow usb data transfer speed on 3.7
I've got a USB external drive that is virtually unusable because it is so slow. mount dev/sd0a on / type ffs (local) /dev/sd0p on /backup type ffs (local, nodev, nosuid, softdep) /dev/sd0o on /destdir type ffs (local, nodev, nosuid, softdep) /dev/sd0d on /home type ffs (local, nodev, nosuid, softdep) /dev/sd0n on /releasedir type ffs (local, nodev, nosuid, softdep) /dev/sd0e on /tmp type ffs (local, nodev, nosuid, softdep) /dev/sd0f on /usr type ffs (local, nodev, softdep) /dev/sd0g on /usr/X11R6 type ffs (local, nodev, softdep) /dev/sd0h on /usr/local type ffs (local, nodev, softdep) /dev/sd0i on /cvs type ffs (local, nodev, nosuid, softdep) /dev/sd0j on /usr/src type ffs (local, nodev, nosuid, softdep) /dev/sd0k on /var type ffs (local, nodev, nosuid, softdep) /dev/sd0l on /var/qmail/bin type ffs (local, nodev, softdep) /dev/sd0m on /var/qmail/queue type ffs (local, nodev, nosuid) /dev/sd1a on /log type ffs (local, nodev, nosuid, softdep) /dev/sd1d on /offline type ffs (local, nodev, nosuid, softdep) /dev/sd1e on /wal type ffs (local, nodev, nosuid, softdep) mfs:13470 on /var/mfs type mfs (asynchronous, local, size=200 512-blocks) /dev/sd2a on /usb_drive type ffs (local) (write a file to the usb drive) time sudo dd if=/dev/zero of=/usb_drive/test_file count=100 100+0 records in 100+0 records out 51200 bytes transferred in 452.234 secs (1132156 bytes/sec) 7m32.69s real 0m0.51s user 0m3.88s system (write a file to the crappy 3WARE RAID1) time sudo dd if=/dev/zero of=/offline/test_file count=100 100+0 records in 100+0 records out 51200 bytes transferred in 9.298 secs (55064036 bytes/sec) 0m13.68s real 0m0.58s user 0m3.78s system dmesg: OpenBSD 3.8-current (GENERIC) #0: Wed Sep 14 22:05:15 MDT 2005 [EMAIL PROTECTED]:/usr/src/sys/arch/i386/compile/GENERIC cpu0: Intel(R) Xeon(TM) CPU 2.66GHz (GenuineIntel 686-class) 2.67 GHz cpu0: FPU,V86,DE,PSE,TSC,MSR,PAE,MCE,CX8,APIC,SEP,MTRR,PGE,MCA,CMOV,PAT,PSE36,CFLUSH,ACPI,MMX,FXSR,SSE,SSE2,SS,HTT,TM,SBF,CNXT-ID real mem = 2147000320 (2096680K) avail mem = 1953087488 (1907312K) using 4278 buffers containing 107454464 bytes (104936K) of memory mainbus0 (root) bios0 at mainbus0: AT/286+(00) BIOS, date 02/04/03, BIOS32 rev. 0 @ 0xf0010 pcibios0 at bios0: rev 2.1 @ 0xf/0x1 pcibios0: PCI IRQ Routing Table rev 1.0 @ 0xf2fb0/256 (14 entries) pcibios0: PCI Interrupt Router at 000:31:0 (Intel 82801CA LPC rev 0x00) pcibios0: PCI bus #4 is the last bus bios0: ROM list: 0xc/0x8000 0xc8000/0x1000 0xc9800/0x800 0xca000/0x1800 cpu0 at mainbus0 pci0 at mainbus0 bus 0: configuration mode 1 (no bios) pchb0 at pci0 dev 0 function 0 Intel E7501 MCH Host rev 0x01 ppb0 at pci0 dev 2 function 0 Intel E7500 MCH rev 0x01 pci1 at ppb0 bus 1 Intel 82870P2 IOxAPIC rev 0x04 at pci1 dev 28 function 0 not configured ppb1 at pci1 dev 29 function 0 Intel 82870P2 PCI-PCI rev 0x04 pci2 at ppb1 bus 2 em0 at pci2 dev 1 function 0 Intel PRO/1000MT (82545EM) rev 0x01: irq 10, address: 00:e0:81:28:e9:71 Intel 82870P2 IOxAPIC rev 0x04 at pci1 dev 30 function 0 not configured ppb2 at pci1 dev 31 function 0 Intel 82870P2 PCI-PCI rev 0x04 pci3 at ppb2 bus 3 ahc1 at pci3 dev 3 function 0 Adaptec AHA-29160 U160 rev 0x02: irq 10 scsibus0 at ahc1: 16 targets st0 at scsibus0 targ 6 lun 0: SEAGATE, DAT 9SP40-000, 910B SCSI3 1/sequential removable st0: density code 0x26, 512-byte blocks, write-enabled twe0 at pci3 dev 6 function 0 3ware Escalade IDE RAID rev 0x01: irq 10 twe0: Escalade V1.3 scsibus1 at twe0: 16 targets sd0 at scsibus1 targ 0 lun 0: 3WARE, Host drive #00, SCSI2 0/direct fixed sd0: 117799MB, 15017 cyl, 255 head, 63 sec, 512 bytes/sec, 241252672 sec total sd1 at scsibus1 targ 2 lun 0: 3WARE, Host drive #02, SCSI2 0/direct fixed sd1: 117799MB, 15017 cyl, 255 head, 63 sec, 512 bytes/sec, 241252672 sec total uhci0 at pci0 dev 29 function 0 Intel 82801CA/CAM USB rev 0x02: irq 10 usb0 at uhci0: USB revision 1.0 uhub0 at usb0 uhub0: Intel UHCI root hub, rev 1.00/1.00, addr 1 uhub0: 2 ports with 2 removable, self powered uhci1 at pci0 dev 29 function 1 Intel 82801CA/CAM USB rev 0x02: irq 9 usb1 at uhci1: USB revision 1.0 uhub1 at usb1 uhub1: Intel UHCI root hub, rev 1.00/1.00, addr 1 uhub1: 2 ports with 2 removable, self powered uhci2 at pci0 dev 29 function 2 Intel 82801CA/CAM USB rev 0x02: irq 11 usb2 at uhci2: USB revision 1.0 uhub2 at usb2 uhub2: Intel UHCI root hub, rev 1.00/1.00, addr 1 uhub2: 2 ports with 2 removable, self powered ppb3 at pci0 dev 30 function 0 Intel 82801BA AGP rev 0x42 pci4 at ppb3 bus 4 fxp0 at pci4 dev 1 function 0 Intel 82557 rev 0x10, i82551: irq 5, address 00:e0:81:28:e9:70 inphy0 at fxp0 phy 1: i82555 10/100 PHY, rev. 4 vga1 at pci4 dev 2 function 0 ATI Rage XL rev 0x27 wsdisplay0 at vga1 mux 1: console (80x25, vt100 emulation) wsdisplay0: screen 1-5 added (80x25, vt100 emulation) fxp1 at pci4 dev 3 function 0 Intel 82557 rev 0x05, i82558: irq 11, address 00:90:27:2a:33:a6 inphy1 at fxp1
Re: very, very slow usb data transfer speed on 3.7
On Tue, 11 Oct 2005 17:27:27 -0600 (MDT) Jeff Ross [EMAIL PROTECTED] wrote: : I've got a USB external drive that is virtually unusable because it : is so slow. : [snip] : (write a file to the usb drive) : : time sudo dd if=/dev/zero of=/usb_drive/test_file count=100 : : 100+0 records in : 100+0 records out : 51200 bytes transferred in 452.234 secs (1132156 bytes/sec) : 7m32.69s real 0m0.51s user 0m3.88s system Over 8Mbits/sec, keep this number in mind. : dmesg: : : OpenBSD 3.8-current (GENERIC) #0: Wed Sep 14 22:05:15 MDT 2005 : [EMAIL PROTECTED]:/usr/src/sys/arch/i386/compile/GENERIC ... : uhci0 at pci0 dev 29 function 0 Intel 82801CA/CAM USB rev 0x02: irq : 10 usb0 at uhci0: USB revision 1.0 : uhub0 at usb0 : uhub0: Intel UHCI root hub, rev 1.00/1.00, addr 1 : uhub0: 2 ports with 2 removable, self powered ... : umass0 at uhub0 port 2 configuration 1 interface 0 : umass0: Prolific Technology Inc. Mass Storage Device, rev 2.00/1.00, : umass0: addr 2 using SCSI over Bulk-Only : scsibus3 at umass0: 2 targets : sd2 at scsibus3 targ 1 lun 0: WDC WD20, 00JB-00GVA0, 08.0 SCSI0 0/ : direct fixed : sd2: 190782MB, 190782 cyl, 64 head, 32 sec, 512 bytes/sec, 390721969 : sec total uhci is USB1, which theoretically tops out at 11Mbits/sec. You won't get much faster access unless you get USB2, which tops out at 480Mbits/sec. -- It is not enough to succeed. Others must fail. -- Gore Vidal
Re: very, very slow usb data transfer speed on 3.7
I don't see the 'EHCI' controller in there anywhere. :( UHCI == usb1.1 EHCI == usb2.0 Top speed for usb1.1 is roughly 1MB/s. Your getting that. :| Two possibilities - your mobo doesn't do usb2.0 - or the ehci device probe isn't grokking your hardware. Andrew. Jeff Ross wrote: I've got a USB external drive that is virtually unusable because it is so slow. mount dev/sd0a on / type ffs (local) /dev/sd0p on /backup type ffs (local, nodev, nosuid, softdep) /dev/sd0o on /destdir type ffs (local, nodev, nosuid, softdep) /dev/sd0d on /home type ffs (local, nodev, nosuid, softdep) /dev/sd0n on /releasedir type ffs (local, nodev, nosuid, softdep) /dev/sd0e on /tmp type ffs (local, nodev, nosuid, softdep) /dev/sd0f on /usr type ffs (local, nodev, softdep) /dev/sd0g on /usr/X11R6 type ffs (local, nodev, softdep) /dev/sd0h on /usr/local type ffs (local, nodev, softdep) /dev/sd0i on /cvs type ffs (local, nodev, nosuid, softdep) /dev/sd0j on /usr/src type ffs (local, nodev, nosuid, softdep) /dev/sd0k on /var type ffs (local, nodev, nosuid, softdep) /dev/sd0l on /var/qmail/bin type ffs (local, nodev, softdep) /dev/sd0m on /var/qmail/queue type ffs (local, nodev, nosuid) /dev/sd1a on /log type ffs (local, nodev, nosuid, softdep) /dev/sd1d on /offline type ffs (local, nodev, nosuid, softdep) /dev/sd1e on /wal type ffs (local, nodev, nosuid, softdep) mfs:13470 on /var/mfs type mfs (asynchronous, local, size=200 512-blocks) /dev/sd2a on /usb_drive type ffs (local) (write a file to the usb drive) time sudo dd if=/dev/zero of=/usb_drive/test_file count=100 100+0 records in 100+0 records out 51200 bytes transferred in 452.234 secs (1132156 bytes/sec) 7m32.69s real 0m0.51s user 0m3.88s system (write a file to the crappy 3WARE RAID1) time sudo dd if=/dev/zero of=/offline/test_file count=100 100+0 records in 100+0 records out 51200 bytes transferred in 9.298 secs (55064036 bytes/sec) 0m13.68s real 0m0.58s user 0m3.78s system dmesg: OpenBSD 3.8-current (GENERIC) #0: Wed Sep 14 22:05:15 MDT 2005 [EMAIL PROTECTED]:/usr/src/sys/arch/i386/compile/GENERIC cpu0: Intel(R) Xeon(TM) CPU 2.66GHz (GenuineIntel 686-class) 2.67 GHz cpu0: FPU,V86,DE,PSE,TSC,MSR,PAE,MCE,CX8,APIC,SEP,MTRR,PGE,MCA,CMOV,PAT,PSE36,CFLUSH,ACPI,MMX,FXSR,SSE,SSE2,SS,HTT,TM,SBF,CNXT-ID real mem = 2147000320 (2096680K) avail mem = 1953087488 (1907312K) using 4278 buffers containing 107454464 bytes (104936K) of memory mainbus0 (root) bios0 at mainbus0: AT/286+(00) BIOS, date 02/04/03, BIOS32 rev. 0 @ 0xf0010 pcibios0 at bios0: rev 2.1 @ 0xf/0x1 pcibios0: PCI IRQ Routing Table rev 1.0 @ 0xf2fb0/256 (14 entries) pcibios0: PCI Interrupt Router at 000:31:0 (Intel 82801CA LPC rev 0x00) pcibios0: PCI bus #4 is the last bus bios0: ROM list: 0xc/0x8000 0xc8000/0x1000 0xc9800/0x800 0xca000/0x1800 cpu0 at mainbus0 pci0 at mainbus0 bus 0: configuration mode 1 (no bios) pchb0 at pci0 dev 0 function 0 Intel E7501 MCH Host rev 0x01 ppb0 at pci0 dev 2 function 0 Intel E7500 MCH rev 0x01 pci1 at ppb0 bus 1 Intel 82870P2 IOxAPIC rev 0x04 at pci1 dev 28 function 0 not configured ppb1 at pci1 dev 29 function 0 Intel 82870P2 PCI-PCI rev 0x04 pci2 at ppb1 bus 2 em0 at pci2 dev 1 function 0 Intel PRO/1000MT (82545EM) rev 0x01: irq 10, address: 00:e0:81:28:e9:71 Intel 82870P2 IOxAPIC rev 0x04 at pci1 dev 30 function 0 not configured ppb2 at pci1 dev 31 function 0 Intel 82870P2 PCI-PCI rev 0x04 pci3 at ppb2 bus 3 ahc1 at pci3 dev 3 function 0 Adaptec AHA-29160 U160 rev 0x02: irq 10 scsibus0 at ahc1: 16 targets st0 at scsibus0 targ 6 lun 0: SEAGATE, DAT 9SP40-000, 910B SCSI3 1/sequential removable st0: density code 0x26, 512-byte blocks, write-enabled twe0 at pci3 dev 6 function 0 3ware Escalade IDE RAID rev 0x01: irq 10 twe0: Escalade V1.3 scsibus1 at twe0: 16 targets sd0 at scsibus1 targ 0 lun 0: 3WARE, Host drive #00, SCSI2 0/direct fixed sd0: 117799MB, 15017 cyl, 255 head, 63 sec, 512 bytes/sec, 241252672 sec total sd1 at scsibus1 targ 2 lun 0: 3WARE, Host drive #02, SCSI2 0/direct fixed sd1: 117799MB, 15017 cyl, 255 head, 63 sec, 512 bytes/sec, 241252672 sec total uhci0 at pci0 dev 29 function 0 Intel 82801CA/CAM USB rev 0x02: irq 10 usb0 at uhci0: USB revision 1.0 uhub0 at usb0 uhub0: Intel UHCI root hub, rev 1.00/1.00, addr 1 uhub0: 2 ports with 2 removable, self powered uhci1 at pci0 dev 29 function 1 Intel 82801CA/CAM USB rev 0x02: irq 9 usb1 at uhci1: USB revision 1.0 uhub1 at usb1 uhub1: Intel UHCI root hub, rev 1.00/1.00, addr 1 uhub1: 2 ports with 2 removable, self powered uhci2 at pci0 dev 29 function 2 Intel 82801CA/CAM USB rev 0x02: irq 11 usb2 at uhci2: USB revision 1.0 uhub2 at usb2 uhub2: Intel UHCI root hub, rev 1.00/1.00, addr 1 uhub2: 2 ports with 2 removable, self powered ppb3 at pci0 dev 30 function 0 Intel 82801BA AGP rev 0x42 pci4 at ppb3 bus 4 fxp0 at pci4 dev 1 function 0 Intel 82557 rev 0x10, i82551: irq 5, address 00:e0:81:28:e9:70 inphy0 at fxp0 phy 1:
Re: very, very slow usb data transfer speed on 3.7
-- [EMAIL PROTECTED]: ~ (17:39) -- # dd if=/dev/wd0c of=/dev/null bs=819200 count=20 20+0 records in 20+0 records out 16384000 bytes transferred in 0.711 secs (23012820 bytes/sec) recall the old speed with apm0: - -- [EMAIL PROTECTED]: ~ (17:13) -- # dd if=/dev/wd0c of=/dev/null bs=819200 count=20 20+0 records in 20+0 records out 16384000 bytes transferred in 1.129 secs (14509606 bytes/sec) that's pretty harsh if other people can reproduce it. :( Incidentally usb transfers *weren't* improved by removing apm0 - -- [EMAIL PROTECTED]: ~ (17:39) -- # dd if=/dev/sd0c of=/dev/null bs=819200 count=20 20+0 records in 20+0 records out 16384000 bytes transferred in 6.017 secs (2722653 bytes/sec) so there's some other factor limiting those. I'm getting the same speed on a snapshot from 09/21 with amd64 on a brand new amd 64 3800+. Lately I was copying around 40G of data onto a usb 2.0 hard disk (yes it was attached to EHCI) and wondered why it took so long, but I didn't pursue the issue further. I also tried with different blocksizes and to eliminate the issue of a too short benchmark I ran for a couple of minutes (about count=500). Actually wait a minute ... /dev/sd0c and /dev/wd0c ? Are you SUPPOSED to read of a block device ? SHOULDN'T it be /dev/rsd0c and /dev/rwd0c ??? ^ ^ RAW DEVICE With the raw devices the speed looks QUITE different: BLOCK DEVICE: sudo dd if=/dev/wd0c of=/dev/null bs=512k count=500 500+0 records in 500+0 records out 262144000 bytes transferred in 16.957 secs (15458831 bytes/sec) # Top shows CPU usage as 28.7% system, 27.9% interrupt, 41.9% idle RAW DEVICE: sudo dd if=/dev/rwd0c of=/dev/null bs=512k count=200 200+0 records in 200+0 records out 104857600 bytes transferred in 1.787 secs (58666485 bytes/sec) # Top shows CPU usage as 4.6% system, 5.4% interrupt, 90.0% idle (same with USB device) BLOCK DEVICE: sudo dd if=/dev/sd0c of=/dev/null bs=512k count=500 500+0 records in 500+0 records out 262144000 bytes transferred in 96.561 secs (2714791 bytes/sec) # Top shows CPU usage as 4.7% system, 10.8% interrupt, 84.6% idle RAW DEVICE: sudo dd if=/dev/rsd0c of=/dev/null bs=512k count=500 500+0 records in 500+0 records out 262144000 bytes transferred in 19.015 secs (13785462 bytes/sec) # Top shows CPU usage as 1.6% system, 2.3% interrupt, 96.1% idle I'm sorry if I understood something wrong, but my understanding was/is that you only use RAW devices with dd (since it uses it's own blocks ). Please tell me if I'm wrong, since (right) knowledge is valueable! Regards, ahb
