It seems our demands for parallel and real-time queries in Splunk are increasing exponentially the more business users see it.
Now this is creating a resource problem on our servers. We're currently running about 20 queries in parallel at all times (ignoring real time queries) and we keep having to add more CPU cores in order to run more searches and correlations.
So to this end, does Splunk support CUDA? I'd rather to run more queries slowly in parallel than to have queries cancelling due to lack of resources.
Splunk does not support CUDA. And, typically, Splunk environments run out of I/O bandwidth well before CPU. (Yours may not be typical if you are running out of CPU before starving the CPU due to insufficient I/O bandwidth)
In your shoes, I'd make sure I'm actually running out of CPU and not just hitting an artificial limit. Then, probably, install more indexers ...