On Tue, Jan 07, 2014 at 07:41:27PM +0800, Conor O'Gorman wrote: > Hi, > > I'm seeing peak rates of about 50 Mbps write speeds, and 150 Mbps read. > Setup is host AMD SB700/SB800/Hudson-2/3 and Atheros SoC, using testusb > tool with usbtest module and g_zero gadget module on TI AM335x. > > Is this to be expected? > Is this a limitation of the controllers or the test setup? Most likely should be limitation of driver. > Can anyone point me at speed numbers for testusb/usbtest/g_zero? I can provide you figures with dwc3 in high speed mode. Bulk In - 257 Mbps (testusb -a -c 10 -t 2 -s 512000 completes in 0.159 sec) Bulk out - 130 Mbps (testusb -a -c 10 -t 1 -s 512000 completes in 0.315 sec) pattern=2 option does not verify data.If you use pattern=2 with both g_zero and usbtest module, you can achieve Bulk Out also as 257 Mbps, while for Bulk In you can reach to 290 Mbps. But if I remember well, there was some minor bug with pattern=2 implementation. I had fixed it locally and probably forgot to upstream. Using bigger buflen with g_zero (may be buflen=409600) you can reach to 344 Mbps while Bulk In and a slight improvement with Bulk Out (271). Regards Pratyush > > This is using tests 5 and 6 from testusb, which read/write various > chained block sizes for various counts. The speeds mentioned are for the > largest 32KB block size, on both the AMD and Atheros. > > Thanks, > Conor > -- > To unsubscribe from this list: send the line "unsubscribe linux-usb" in > the body of a message to majordomo@xxxxxxxxxxxxxxx > More majordomo info at http://vger.kernel.org/majordomo-info.html -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
