John,
Since you work in F1...
The guys from Intelligent Light (http://www.ilight.com/) gave a
presentation at SC13 on how they helped both Red Bull F1 and Zipp wheels
(http://www.zipp.com) with HPC. I think both of these examples show how
easily quantifiable ROI can be in certain scenarios. It also shows how
you can use different methods to calculate ROI.
In the Red Bull F1 case, they made it easier and quicker for Red Bull to
analyze the results of their CFD simulations. Obviously, if you can
calculate the # of aerodynamics simulations they were able to analyze
per week before and after, you can figure out the value based on the
cost of labor for the aerodynamics experts, and then calculate the value
of this increase in productivity. The ultimate measure, though, would be
to measure the increase in co-sponsor dollars/prize-winnings after using
this new workflow.
For Zipp, they used HPC for the first time to design an aerodynamic
wheel, instead of just wind-tunnel testing. I have no doubt that
computer modeling is much cheaper and quicker than wind-tunnel testing.
From what I remember, Zipp used a very small cluster, or the cloud, to
do their simulations, so the investment was very small. One ROI
calculation could be based on the saving of computer simulations vs.
wind-tunnel testing. I think a better ROI calculation, though, would be
based on how much these new computer-designed wheels are adding to their
profits.
I think which ROI calculation you use in the above scenarios depends on
who you are trying to convince - your dept head, or your CEO (and
shareholders), but there's no doubt all the calculations are valid, and
the numbers would be relatively easy to obtain.
On 11/26/2013 12:57 PM, John Hearns wrote:
Regarding ROI in industry, I think that's a good topic for discussion.
Like any area HPC is applied to, I would guess that the first thing
you have to look at is the capability the equipment gives you.
Either in terms of 'we could not run this simulation, or a simulation
as detailed as we need' without it,
or 'we get the results back in time to influence our decisions', ie.
scale and time to solution.
Then you have to say what effect that will have on your business goals
- either by saying that if we do not have that HPC capacity we
will have to do the modelling on other, perhaps more expensive, ways
such as crash testing real cars or aircraft engines.
Or by saying that if we have the capability we can increase the
performance of our product, and thereby gain competitiveness.
I think the time to solution argument is an important one - certainly
is where I work! And think of obvious examples
such as weather forecasting where you need that forecast to be
produced before a certain time.
Again time to solution can be quantified int he medical diagnostic
arena - you are probably able to say that if you run (say) a radiation
dose plan within the time of a clinic it means the patient does not
have to come back for a second visit, thereby saving soem amount of money.
Apologies if i am stating the blindingly obvious.
_______________________________________________
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit
http://www.beowulf.org/mailman/listinfo/beowulf
_______________________________________________
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit
http://www.beowulf.org/mailman/listinfo/beowulf
