Hello All,
We have done a performance test on USB2 hard disk with the following
configuration:
(1) HDD is directly connected to one HBA;
(2) HDD is connected to one USB2 Hub, which is connected to the HBA.
From test results, we found about 10% performance drop in second case
comparing with the first one. Further, after we added more layer hubs to
do more tests,
we found that every layer hub causes one more about 10% performance drop
on USB2 HDD.
For example, we saw about 20% performance drop in two layer hubs and 30%
drop in three
layer hubs. We observed these results under both windows 2K and windows XP
with
MS PDK EHCI driver.
Does anyone know which part causes the above issue, hardware or OS or
drivers?
Thank you in advance for your help!
Regards,
Shunnian
Have you compared traces on the USB bus to see if there are turnaround /
timeout problems or if there are larger than normal interpacket delays? I’ve
seen problems with certain hubs that seem to inject additional interpacket
delays before transmission. (Noted by taking the trace before and after the
hub in the chain with a CATC)
-----Original Message-----
From: xxxxx@adaptec.com [mailto:xxxxx@adaptec.com]
Sent: Monday, March 04, 2002 11:03 PM
To: NT Developers Interest List
Subject: [ntdev] USB2 HDD performance drop if connected behind Hub
Hello All,
We have done a performance test on USB2 hard disk with the following
configuration:
(1) HDD is directly connected to one HBA;
(2) HDD is connected to one USB2 Hub, which is connected to the HBA.
From test results, we found about 10% performance drop in second case
comparing with the first one. Further, after we added more layer hubs to
do more tests,
we found that every layer hub causes one more about 10% performance drop
on USB2 HDD.
For example, we saw about 20% performance drop in two layer hubs and 30%
drop in three
layer hubs. We observed these results under both windows 2K and windows XP
with
MS PDK EHCI driver.
Does anyone know which part causes the above issue, hardware or OS or
drivers?
Thank you in advance for your help!
Regards,
Shunnian
You are currently subscribed to ntdev as: xxxxx@nvidia.com
To unsubscribe send a blank email to %%email.unsub%%
> (1) HDD is directly connected to one HBA;
(2) HDD is connected to one USB2 Hub, which is connected to the HBA.
From test results, we found about 10% performance drop in second case
comparing with the first one. Further, after we added more layer hubs to
Do you have an USB hardware-level analyzer?
Max
Hi, Mark,
Thanks for your reply. Since I am new to use USB analyzer, could you tell
me more on how to do that with CATC USB2 Analyzer? I will appreciate your
help!
Regards,
Shunnian
Hi, Max,
We just have a USB2 analyzer in the lab. Could you tell me how it can help
me on this issue?
Thanks in advance!
Shunnian
Date Mar 6, 2002 @ 10:08:27
Author “Maxim S. Shatskih”
Subject Re: USB2 HDD performance drop if connected behind Hub
Body > (1) HDD is directly connected to one HBA;
> (2) HDD is connected to one USB2 Hub, which is connected to the HBA.
>
> From test results, we found about 10% performance drop in second case
> comparing with the first one. Further, after we added more layer hubs to
Do you have an USB hardware-level analyzer?
Max
Measure all delays and timings of USB traffic during the disk operation.
What of USB-specified delays is incremented when the hub is present?
Max
----- Original Message -----
From:
To: “NT Developers Interest List”
Sent: Monday, March 11, 2002 12:52 PM
Subject: [ntdev] Re: USB2 HDD performance drop if connected behind Hub
> Hi, Max,
>
> We just have a USB2 analyzer in the lab. Could you tell me how it can help
> me on this issue?
>
> Thanks in advance!
> Shunnian
>
> Date Mar 6, 2002 @ 10:08:27
>
> Author “Maxim S. Shatskih”
>
> Subject Re: USB2 HDD performance drop if connected behind Hub
>
> Body > (1) HDD is directly connected to one HBA;
> > (2) HDD is connected to one USB2 Hub, which is connected to the HBA.
> >
> > From test results, we found about 10% performance drop in second case
> > comparing with the first one. Further, after we added more layer hubs to
>
> Do you have an USB hardware-level analyzer?
>
> Max
>
> —
> You are currently subscribed to ntdev as: xxxxx@storagecraft.com
> To unsubscribe send a blank email to %%email.unsub%%
>
Hi, Max,
So you guess it is hub hardware that causes the problem. Am I right? I
will do what you suggest when I get free time and let you know the result.
Thanks again!
Shunnian
Mar 11, 2002 @ 17:15:04
Author “Maxim S. Shatskih”
Subject Re: USB2 HDD performance drop if connected behind Hub
Body Measure all delays and timings of USB traffic during the disk
operation.
What of USB-specified delays is incremented when the hub is present?
Max
----- Original Message -----
From:
To: “NT Developers Interest List”
Sent: Monday, March 11, 2002 12:52 PM
Subject: [ntdev] Re: USB2 HDD performance drop if connected behind Hub
> Hi, Max,
>
> We just have a USB2 analyzer in the lab. Could you tell me how it can help
> me on this issue?
>
> Thanks in advance!
> Shunnian
>
> Date Mar 6, 2002 @ 10:08:27
>
> Author “Maxim S. Shatskih”
>
> Subject Re: USB2 HDD performance drop if connected behind Hub
>
> Body > (1) HDD is directly connected to one HBA;
> > (2) HDD is connected to one USB2 Hub, which is connected to the HBA.
> >
> > From test results, we found about 10% performance drop in second case
> > comparing with the first one. Further, after we added more layer hubs to
>
> Do you have an USB hardware-level analyzer?
>
> Max
>
> —
> You are currently subscribed to ntdev as: xxxxx@storagecraft.com
> To unsubscribe send a blank email to %%email.unsub%%
>