I'm sorry if this is the wrong place to ask this, but what I want to do is make a 32 core cluster computer.
I want to use one master and the other as nodes. So I want the storage to be centralized, and the operating system running on only one machine.
I want to use Windows 7/8 on this machine, and use it for rendering 3d movies and photoshop work. I also want to add a graphics card and 16 gb ram to each of the 4 motherboards, and they all should utilize load balancing, and ideally, power up only when their resources are needed.
I read about beowulf, but a beowulf cluster uses identical hardware, which I dont want to. Also, I want it to utilize as much CUDA or OpenCL as possible.
I want the computer to identify all the hardware as one single piece if possible, but I want to power up each node only when the node before it reaches 75% usage.
I haven't done any work using clusters before, so I barely have any knowledge of the field.
So can someone please tell me whether or not this is possible? And what kind of a cluster would need to be made for this? Thanks in advance.
I want to use one master and the other as nodes. So I want the storage to be centralized, and the operating system running on only one machine.
I want to use Windows 7/8 on this machine, and use it for rendering 3d movies and photoshop work. I also want to add a graphics card and 16 gb ram to each of the 4 motherboards, and they all should utilize load balancing, and ideally, power up only when their resources are needed.
I read about beowulf, but a beowulf cluster uses identical hardware, which I dont want to. Also, I want it to utilize as much CUDA or OpenCL as possible.
I want the computer to identify all the hardware as one single piece if possible, but I want to power up each node only when the node before it reaches 75% usage.
I haven't done any work using clusters before, so I barely have any knowledge of the field.
So can someone please tell me whether or not this is possible? And what kind of a cluster would need to be made for this? Thanks in advance.