USB1 & USB2 - speeds, debate, iBurst UTD, etc...

S

slimothy

Banned
Joined
Jan 14, 2005
Messages
4,808
Reaction score
0
USB runs at 12mbit? are you high? USB 2.0 is closer to 340 and btw ethernet these days is 100mbit
 
slimothy said:
USB runs at 12mbit? are you high? USB 2.0 is closer to 340 and btw ethernet these days is 100mbit
Click to expand...

ethernet these days is closer to a gigabit - just about all new boxen come with this standard. btw - does iBurst UTD use USB 1.1 or 2.0? I see no reason to use 2.0.
 
its 2.0 and my network card isn't exactly old and she pumps a good 100mbit both ways
 
USB 1.0 is for all intents and purposes obsolete, but supported as backwards compatibility. I know this, I design USB stuff at work :-D I use USB 2.0 in all new products as a rule
 
480Mbps in windows time slices yeah, not continuous. That's why the USB performs worse than the ethernet port
 
Look here, I am tired of people questioning my knowledge on forums and ****.
Have you ever developed a USB function?
Have you ever written a device driver for Windows using USB?

I know USB, very well and I can tell you now what you can and cannot do with it. There are products in the South African market which I designed, yes, I designed them. What have any of you designed eh?

If you want proof I will set up the experiment again, when I can get hold of a digital oscilloscope. I only have an analog scope at home. Once the experiment has been set up, I will give you a venue and a time and date and then you can see for yourself that the USB is serviced in time intervals and not continuously as some of you may think. We discovered with this setup the actual granularity of the schedulers in Windows 98 vs Windows XP

I bet no-one here knows that USB is not a network, and that it's a bus system of one master and many slaves and that units can only speak when spoken to, effectively, altho they can pull the attention of the MASTER
 
well actually I was joking, BTW though taking the attitude "OMG I DESIGNED LIEK SOME USB THINGY WHAT DID YOU IDIOTS DESIGN" is not too smart, I'm just saying there are lots of IT professionals here and if you really want to see where your productivity in your career has you on the myADSL foodchain I'd say it's nowhere near the top, just chill and lose the condesending attitude.

And I have wrote a device driver actually because um not sure if you know this BUT YOU DO IT IN SOFTWARE ENGINEERING AT UNI IN YEAR 1
 
Hey Kei

Hey Kei,

Does designing whole Computers, Operating systems, Printer Drivers, Network Operating Systems for the operating System and Database Management Systems count in your books, cause that is what I was doing in the middle to late 1980's. These things were actually sold and worked because I saw one still in operation about a year ago, over 15 years since date of manufacture. Of course I don't dabble too much in the technical side of things too much nowadays.
Reason, try Sanctions and see what innovative solutions one has to come up with.

Damn I miss those days when a 11 o'clock appointment had to be confirmed as AM or PM. Nowadays its 16:30 and everyone is off home.

@ Slimothy
Right on with your remarks re the attitude. :D
 
Kei said:
Have you ever developed a USB function?
Have you ever written a device driver for Windows using USB?
Click to expand...

I have, for Linux, for the Sucktech USB modems, and various other devices.

Kei said:
I bet no-one here knows that USB is not a network, and that it's a bus system of one master and many slaves and that units can only speak when spoken to, effectively, altho they can pull the attention of the MASTER
Click to expand...

I do, and I can agree. Althought, trying to show Slimothy "proof" of this, would be like trying to explain Einstein's relativity theory to a monkey. Pointless.
 
Last edited:
slimothy said:
well actually I was joking
Click to expand...

Dude... you weren't joking... Just a cover up when you realised that Kei actually knows what he is talking about.

TheRoDent said:
I do, and I can agree. Althought, trying to show Slimothy "proof" of this, would be like trying to explain to Einstein's relativity theory to a monkey. Pointless.
Click to expand...
Heh heh heh. A bit harsh though.
 
Kei said:
I bet no-one here knows that USB is not a network, and that it's a bus system of one master and many slaves and that units can only speak when spoken to, effectively, altho they can pull the attention of the MASTER
Click to expand...

yes, i did know that
 
I am not a software person, I am a hardware designer and I write embedded code. Nonetheless, for what we were trying to achieve at the time, USB1.1 failed miserably. Went to USB 2.0 and the problem was solved. I developed the solution from the ground up over 9 months working alongside my brother who is a software guru. He wrote and developed the driver and API and all those things.

We developed a USB -> RS232 adapter for a special application. The best full-duplex data rate achievable was 1200bps on USB1.1. The parts available to us had two small endpoints (IN and OUT) and even then we'd lose one or two characters with continuous streams on the RS232.

We migrated to USB2.0 @12mbps and the problem is solved mostly, except now with fast baudrates such as 115200bps on the RS232 the character loss is 1 in 20000. I have tried many complex buffer schemes, but in general it appears that the data at high baudrates fills the buffer faster than the USB is serviced. I am using isochronous and bulk endpoint types. (these failed miserably on USB1.1- don't know why, explains somewhat why USB1.1 was superseded rather quickly) The USB spec is rather ambiguous at best.

I am now working with a USB 2.0 root hub chip and the manufacturer states that the chip is to be serviced in intervals of 1mS (this also implies the device, the USB root hub chip, will assert an interrupt every 1mS). This means that, effectively you are transacting with your USB device in chunks every 1mS. Therefore, the more endpoints you have, the more data you can get across in those time slices. The more endpoints, the more complex the code becomes to handle all of that, so it's a tradeoff between processing speed in your USB function and the throughput you can manage on the USB
 
You must log in or register to reply here.
Back
Top
Sign up to the MyBroadband newsletter
X