AMBER Archive (2008)

Subject: AMBER: Amber on Infiniband clusters: SDR vs DDR cables?

From: Sasha Buzko (obuzko_at_ucla.edu)
Date: Thu Sep 25 2008 - 18:06:44 CDT

Hi all,
we are setting up a small Infiniband cluster to run Amber, with
interface cards and switch supporting 20 Gb/s transfer rates. However,
the vendor has supplied us with 10 Gb/s (SDR) 4x cables.
How much of a performance increase could we expect from switching to
DDR-capable cables? In other words, how significant is the difference
between SDR and DDR cables in this setup?

Thank you for any comments

Sasha

Carlos Simmerling wrote:
> give us more info- it's hard to say what's going on. the protein
> moving out of the box during MD may be normal, and you need to "image"
> the trajectory with ptraj. search the archives for plenty of
> information and examples. if the box is too small before you run MD.
> then you need to increase your buffer size in the solvate command when
> you run leap. if you want solvateoct, check the archives or the ptraj
> manual, and let us know specific questions. the tutorials can also be
> helpful for this.
>
>
>
> On Wed, Sep 24, 2008 at 6:56 PM, Qiuting Hong <qiutinghong_at_gmail.com> wrote:
>
>> Dear amber users,
>>
>> I am using TIP3PBOX to solvate ubiquitin. After the explicit water
>> simulation, I reimage the water back into the box, but find out part of the
>> ubiquitin side chain is out of the water box. I guss the box is too small.
>> I am trying to re-run the simulation. But I don't know how to change the
>> size of TIP3PBOX. I want to use solvateoct.
>> Can somebody help me out?
>>
>> Qiuting Hong
>>
> -----------------------------------------------------------------------
> The AMBER Mail Reflector
> To post, send mail to amber_at_scripps.edu
> To unsubscribe, send "unsubscribe amber" (in the *body* of the email)
> to majordomo_at_scripps.edu
>
>

-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber_at_scripps.edu
To unsubscribe, send "unsubscribe amber" (in the *body* of the email)
      to majordomo_at_scripps.edu