Runtime prediction framework for CPU intensive applications

1. A method comprising:

analyzing application input data for a compute-intensive application on a computer system of a customer using an application-specific module, wherein the application-specific module scans the application input data and collects statistical information relevant to calculating a computing time on a CPU farm to determine costs to run the compute-intensive application using the application input data on the CPU farm, wherein the statistical information represents a scaled-down representation of the application input data;

receiving the scaled-down representation of the application input data over a computer network;

calculating a computing requirement based on the scaled-down representation;

calculating a turn-around time and an actual cost to a customer to run the compute-intensive application with the application input data, on one or more processors, based on the calculated computing requirement; and

sending the turn-around time and the actual cost to the customer's client software.

2. The method of claim 1 wherein the compute-intensive application is to perform computer graphics rendering.

3. The method of claim 1 wherein the compute-intensive application is to perform logic simulation.

4. The method of claim 1 wherein the scaled-down representation of the application input data is generic to a class of applications.

5. The method of claim 1 wherein the scaled-down representation of the application input data includes the geometry, lights, number of triangles, textures, shading method, camera, ray-tracing, anti-aliasing, and motion-blur of an underlying scene.

6. The method of claim 1 wherein the turn-around time and actual cost are transmitted over an internet to the customer's client software.

7. The method of claim 1 wherein the compute-intensive application is to perform logic simulation and the actual cost is provided to the customer in terms of cost per logic gate.

8. The method of claim 1 wherein the compute-intensive application is to perform computer graphics rendering and the actual cost is provided to the customer in terms of cost per image frame.

9. A system having one or more processors for generating a rendering cost estimate, the system comprising:

an application-specific module to scan one or more input data files to a compute-intensive application and to collect statistical information relevant to calculating an actual computing time on a CPU farm to determine computing costs to run the compute-intensive application on the CPU farm;

a heuristic modeler module coupled to the output of the application-specific module, to calculate a computing requirement; and

a run-time calculator module coupled to the output of the heuristic modeler module, to compute a turn-around time and an actual cost to run the application using the one or more input data files on the one or more processors, wherein the run-time calculator module provides the turn-around time and the actual cost to the customer's client software.

10. The system of claim 9 wherein the modules are to communicate with each other over an internet.

11. The system of claim 9 wherein the statistical information comprises a scaled-down representation of the input data files to include the geometry, lights, number of triangles, textures, shading method, camera, ray-tracing, anti-aliasing, and motion-blur of an underlying scene.

12. An article of manufacture comprising:

a machine readable medium containing instructions which, when executed by a processor, cause a machine to perform operations comprising:

analyzing application input data for a compute-intensive application on a computer system of a customer using an application-specific module, wherein the application-specific module scans the application input data and collects statistical information relevant to calculating a computing time on a CPU farm to determine costs to run the compute-intensive application using the application input data on the CPU farm, wherein the statistical information represents a scaled-down representation of the application input data;

calculating a computing requirement based on the scaled-down representation of application input data to the compute-intensive application to determine costs to run the compute-intensive application using the application input data;

calculating a turn-around time and an actual cost to the customer to run the compute-intensive application with the application input data, on one or more processors, based on the calculated computing requirement; and

providing the turn-around time and the actual cost to the customer's client software.

13. The article of manufacture of claim 12 wherein the medium includes further instructions to create the scaled-down representation of the application input data as being generic to a class of applications.

14. The article of manufacture of claim 12 wherein the medium includes further instructions to create the scaled-down representation of the application input data as having the geometry, lights, number of triangles, textures, shading method, camera, ray-tracing, anti-aliasing, and motion-blur of an underlying scene.

15. The article of manufacture of claim 12 wherein the medium includes further instructions to enable the scaled-down representation of the application input data to be received over an internet from the client software.

16. The article of manufacture of claim 12 wherein the medium includes further instructions to enable the turn-around time and actual cost to be transmitted over the internet to the customer's client software.

17. The article of manufacture of claim 12 wherein the medium includes further instructions to calculate the actual cost in terms of cost per logic gate.

18. The article of manufacture of claim 12 wherein the medium includes further instructions to calculate the actual cost in terms of cost per image frame.