Question on Computational Speed...

Archive AnyScript Yahoo archive
1

Hi AnyBuddies,

I use the ‘wheel turn’ model in Repository 5 using AnyBody ‘pre 2.0’ version and
I remember the computation speed is considerably fast.

I recently run the same ‘wheel turn’ model using Repository 6 model with AnyBody
2.x version but it is somehow slow. Same situation with Repository 7 model with
AnyBody 3.0.

I do not have quantitative data to provide here, and I do not have ‘pre 2.0’
version available to run the Repository 5.0 model anymore - using a
non-quantitative comparison: I can sit in front of my PC and get the result
using Repository 5 and ‘pre 2.0’ version in minutes, but with Repository 6.0 and
ver 2.x, I can go make a cup of coffee, enjoying it and when I come back, it is
still computing.

Anyone have similar experience? Or any idea why this is happening?

P.S. is it possible to provide AnyBody ver. 1.x to let me try it out?

Thanks,
Leng-Feng

2

Hi Leng-Feng

I have checked the speeds in two different ways:

  1. Running the WheelTurn model from rep. 6 in version 2.0.1 and 3.1
    gives the same speed. In 2 min 15 steps is done.
  2. Running Wheelturn model from Rep 5 in version 2.01 does 25 steps in
    2 min so repository 5 is faster than ver. 6

So my conclusion is that the difference in speed; is due to the repository
not AnyBody.

I can not remember all the differences between the two repositories, but
most likely it is caused by the wrapping muscles where the StringMesh
variable may have increased on some the muscles, typically if there is a
wrapping problem we will increase this number which makes it stable but also
slows it down, so I guess during time this number has a tendency to
increase, but it has also improved the robustness.

For some model it may be possible to reduce the StringMesh variable but this
may cause other models to fail, so if you change this variable in principle,
it would be good to test all other models.

There is a download page where you can download all versions of the system,
please see this link http://www.anybodytech.com/index.php?id=877

unfortunately ver.1 is missing.

Best regards

Søren

From: anyscript@yahoogroups.com [mailto:anyscript@yahoogroups.com] On Behalf
Of Leng-Feng Lee
Sent: 16 February 2009 16:26
To: anyscript@yahoogroups.com
Cc: lengfenglee@yahoo.com
Subject: [AnyScript] Question on Computational Speed…

Hi AnyBuddies,

I use the ‘wheel turn’ model in Repository 5 using AnyBody ‘pre 2.0’ version
and I remember the computation speed is considerably fast.

I recently run the same ‘wheel turn’ model using Repository 6 model with
AnyBody 2.x version but it is somehow slow. Same situation with Repository 7
model with AnyBody 3.0.

I do not have quantitative data to provide here, and I do not have ‘pre 2.0’
version available to run the Repository 5.0 model anymore - using a
non-quantitative comparison: I can sit in front of my PC and get the result
using Repository 5 and ‘pre 2.0’ version in minutes, but with Repository 6.0
and ver 2.x, I can go make a cup of coffee, enjoying it and when I come
back, it is still computing.

Anyone have similar experience? Or any idea why this is happening?

P.S. is it possible to provide AnyBody ver. 1.x to let me try it out?

Thanks,
Leng-Feng

[Non-text portions of this message have been removed]

3

Hi Søren,

Thanks for the reply.

But when I ran Model from Rep. 5 using AnyBody 3.0 I ran into error - I think
the scripting is more strict in AnyBody 3.0?

Anyway, I shall use the latest Rep. in my future model.

btw, the new AnyBody website looks very neat. =)

Thanks,
Leng-Feng

— On Wed, 2/18/09, AnyScript Support <anyscriptsup@anybodytech.com> wrote:
From: AnyScript Support <anyscriptsup@anybodytech.com>
Subject: RE: [AnyScript] Question on Computational Speed…
To: anyscript@yahoogroups.com
Date: Wednesday, February 18, 2009, 10:28 AM

         Hi Leng-Feng

I have checked the speeds in two different ways:

  1. Running the WheelTurn model from rep. 6 in version 2.0.1 and 3.1

gives the same speed. In 2 min 15 steps is done.

  1. Running Wheelturn model from Rep 5 in version 2.01 does 25 steps in

2 min so repository 5 is faster than ver. 6

So my conclusion is that the difference in speed; is due to the repository

not AnyBody.

I can not remember all the differences between the two repositories, but

most likely it is caused by the wrapping muscles where the StringMesh

variable may have increased on some the muscles, typically if there is a

wrapping problem we will increase this number which makes it stable but also

slows it down, so I guess during time this number has a tendency to

increase, but it has also improved the robustness.

For some model it may be possible to reduce the StringMesh variable but this

may cause other models to fail, so if you change this variable in principle,

it would be good to test all other models.

There is a download page where you can download all versions of the system,

please see this link http://www.anybodyt ech.com/index. php?id=877

unfortunately ver.1 is missing.

Best regards

Søren

From: anyscript@yahoogrou ps.com [mailto:anyscript@yahoogrou ps.com] On Behalf

Of Leng-Feng Lee

Sent: 16 February 2009 16:26

To: anyscript@yahoogrou ps.com

Cc: lengfenglee@ yahoo.com

Subject: [AnyScript] Question on Computational Speed…

Hi AnyBuddies,

I use the ‘wheel turn’ model in Repository 5 using AnyBody ‘pre 2.0’ version

and I remember the computation speed is considerably fast.

I recently run the same ‘wheel turn’ model using Repository 6 model with

AnyBody 2.x version but it is somehow slow. Same situation with Repository 7

model with AnyBody 3.0.

I do not have quantitative data to provide here, and I do not have ‘pre 2.0’

version available to run the Repository 5.0 model anymore - using a

non-quantitative comparison: I can sit in front of my PC and get the result

using Repository 5 and ‘pre 2.0’ version in minutes, but with Repository 6.0

and ver 2.x, I can go make a cup of coffee, enjoying it and when I come

back, it is still computing.

Anyone have similar experience? Or any idea why this is happening?

P.S. is it possible to provide AnyBody ver. 1.x to let me try it out?

Thanks,

Leng-Feng

[Non-text portions of this message have been removed]

[Non-text portions of this message have been removed]

4

Hi Leng-Feng

You are right about the parser, ver. 3.0 has a more strict type check when
it parses the script than ver. 2, so I had to a few changes in the model to
make rep 5 run on ver. 3 :wink:

Best regards

Søren

From: anyscript@yahoogroups.com [mailto:anyscript@yahoogroups.com] On Behalf
Of Leng-Feng Lee
Sent: 24 February 2009 22:56
To: anyscript@yahoogroups.com
Subject: RE: [AnyScript] Question on Computational Speed…

Hi Søren,

Thanks for the reply.

But when I ran Model from Rep. 5 using AnyBody 3.0 I ran into error - I
think the scripting is more strict in AnyBody 3.0?

Anyway, I shall use the latest Rep. in my future model.

btw, the new AnyBody website looks very neat. =)

Thanks,
Leng-Feng

— On Wed, 2/18/09, AnyScript Support <anyscriptsup@
<mailto:anyscriptsup%40anybodytech.com> anybodytech.com> wrote:
From: AnyScript Support <anyscriptsup@
<mailto:anyscriptsup%40anybodytech.com> anybodytech.com>
Subject: RE: [AnyScript] Question on Computational Speed…
To: anyscript@yahoogrou <mailto:anyscript%40yahoogroups.com> ps.com
Date: Wednesday, February 18, 2009, 10:28 AM

Hi Leng-Feng

I have checked the speeds in two different ways:

  1. Running the WheelTurn model from rep. 6 in version 2.0.1 and 3.1

gives the same speed. In 2 min 15 steps is done.

  1. Running Wheelturn model from Rep 5 in version 2.01 does 25 steps in

2 min so repository 5 is faster than ver. 6

So my conclusion is that the difference in speed; is due to the repository

not AnyBody.

I can not remember all the differences between the two repositories, but

most likely it is caused by the wrapping muscles where the StringMesh

variable may have increased on some the muscles, typically if there is a

wrapping problem we will increase this number which makes it stable but also

slows it down, so I guess during time this number has a tendency to

increase, but it has also improved the robustness.

For some model it may be possible to reduce the StringMesh variable but this

may cause other models to fail, so if you change this variable in principle,

it would be good to test all other models.

There is a download page where you can download all versions of the system,

please see this link http://www.anybodyt ech.com/index. php?id=877

unfortunately ver.1 is missing.

Best regards

Søren

From: anyscript@yahoogrou ps.com [mailto:anyscript@yahoogrou ps.com] On
Behalf

Of Leng-Feng Lee

Sent: 16 February 2009 16:26

To: anyscript@yahoogrou ps.com

Cc: lengfenglee@ yahoo.com

Subject: [AnyScript] Question on Computational Speed…

Hi AnyBuddies,

I use the ‘wheel turn’ model in Repository 5 using AnyBody ‘pre 2.0’ version

and I remember the computation speed is considerably fast.

I recently run the same ‘wheel turn’ model using Repository 6 model with

AnyBody 2.x version but it is somehow slow. Same situation with Repository 7

model with AnyBody 3.0.

I do not have quantitative data to provide here, and I do not have ‘pre 2.0’

version available to run the Repository 5.0 model anymore - using a

non-quantitative comparison: I can sit in front of my PC and get the result

using Repository 5 and ‘pre 2.0’ version in minutes, but with Repository 6.0

and ver 2.x, I can go make a cup of coffee, enjoying it and when I come

back, it is still computing.

Anyone have similar experience? Or any idea why this is happening?

P.S. is it possible to provide AnyBody ver. 1.x to let me try it out?

Thanks,

Leng-Feng

[Non-text portions of this message have been removed]

[Non-text portions of this message have been removed]

[Non-text portions of this message have been removed]