FastDL [message #54063] |
Thu, 17 May 2007 06:30  |
Chris W
Messages: 12 Registered: May 2007
|
Junior Member |
|
|
Hi everyone,
I would like to ask if anyone has experience with the FastDL
products. We are considering them to process sub regions of data in
parallel, with no "interaction" between the different data sets.
If anyone could describe their experiences with FastDL I would
appreciate it, or if something is available for this as well.
Thanks
Chris
|
|
|
Re: FastDL [message #54182 is a reply to message #54063] |
Wed, 23 May 2007 16:28  |
Trae
Messages: 23 Registered: May 2007
|
Junior Member |
|
|
I've been looking into this too. It seems easy to use and will add
true clustering capability to IDL jobs with very minimal effort.
However...
1. Their sales staff have been less than responsive to my inquiries.
2. Man! It is expensive. For an academic license it would be $9,000
for the full FastDl package for 64 CPUS. They count a dual core chip
as two CPUs and a quad core as 4. It is $5,500 for just the mpiDL
which is what you need if you want to do cluster operations as opposed
to grid operations.
UGH!
I'm doing plasma physics simulations and have to crunch lots of data
and make lots of movies. I might have paid the price of their sales
staff ever got back with me after my second, or third email. Now I
might use the money to buy another new computer.
Let me know if you find out anything.
Cheers,
-Trae
On May 17, 7:30 am, Chris W <cwood1...@gmail.com> wrote:
> Hi everyone,
>
> I would like to ask if anyone has experience with the FastDL
> products. We are considering them to process sub regions of data in
> parallel, with no "interaction" between the different data sets.
>
> If anyone could describe their experiences with FastDL I would
> appreciate it, or if something is available for this as well.
>
> Thanks
> Chris
|
|
|
Re: FastDL [message #54183 is a reply to message #54063] |
Wed, 23 May 2007 15:38  |
Trae
Messages: 23 Registered: May 2007
|
Junior Member |
|
|
I've been looking into this too. It seems easy to use and will add
true clustering capability to IDL jobs with very minimal effort.
However...
1. Their sales staff have been less than responsive to my inquiries.
2. Man! It is expensive. For an academic license it would be $9,000
for the full FastDl package for 64 CPUS. They count a dual core chip
as two CPUs and a quad core as 4. It is $5,500 for just the mpiDL
which is what you need if you want to do cluster operations as opposed
to grid operations. It sounds like you want grid operations only so
that would be $4,000 for up to 64 CPUs with TaskDL.
I'm doing plasma physics simulations and have to crunch lots of data
and make lots of movies. I might have paid the price if their sales
staff ever got back with me after my second, or third email. Now I
might use the money to buy another new computer.
Let me know if you find out anything. I'm interested in trying to make
a cluster (preferably a Mac cluster) that is easy for IDL's multi-
threading and IDL_IDLBridge to use.
Cheers,
-Trae
On May 17, 7:30 am, Chris W <cwood1...@gmail.com> wrote:
> Hi everyone,
>
> I would like to ask if anyone has experience with the FastDL
> products. We are considering them to process sub regions of data in
> parallel, with no "interaction" between the different data sets.
>
> If anyone could describe their experiences with FastDL I would
> appreciate it, or if something is available for this as well.
>
> Thanks
> Chris
|
|
|
Re: FastDL [message #54185 is a reply to message #54063] |
Wed, 23 May 2007 14:58  |
Trae
Messages: 23 Registered: May 2007
|
Junior Member |
|
|
I've been looking into this too. It seems easy to use and will add
true clustering capability to IDL jobs with very minimal effort.
However...
1. Their sales staff have been less than responsive to my inquiries.
2. Man! It is expensive. For an academic license it would be $9,000
for the full FastDl package for 64 CPUS. They count a dual core chip
as two CPUs and a quad core as 4. It is $5,500 for just the mpiDL
which is what you need if you want to do cluster operations as opposed
to grid operations. It sounds like you want grid operations only so
that would be $4,000 for up to 64 CPUs with TaskDL.
I'm doing plasma physics simulations and have to crunch lots of data
and make lots of movies. I might have paid the price if their sales
staff ever got back with me after my second, or third email. Now I
might use the money to buy another new computer.
Let me know if you find out anything. I'm interested in trying to make
a cluster (preferably a Mac cluster) that is easy for IDL's multi-
threading and IDL_IDLBridge to use.
Cheers,
-Trae
On May 17, 7:30 am, Chris W <cwood1...@gmail.com> wrote:
> Hi everyone,
>
> I would like to ask if anyone has experience with the FastDL
> products. We are considering them to process sub regions of data in
> parallel, with no "interaction" between the different data sets.
>
> If anyone could describe their experiences with FastDL I would
> appreciate it, or if something is available for this as well.
>
> Thanks
> Chris
|
|
|