Currently, not all benchmarks have a full set of inputs. Furthermore, even when we have all inputs, it would be great to use different inputs so that all invocations are not doing exactly the same input file. The input generation should still be deterministic.
This issue also connects with #298 and #300. For some benchmarks, batching support might be needed for a proper "large" input, e.g., see issue #296
Currently, not all benchmarks have a full set of inputs. Furthermore, even when we have all inputs, it would be great to use different inputs so that all invocations are not doing exactly the same input file. The input generation should still be deterministic.
This issue also connects with #298 and #300. For some benchmarks, batching support might be needed for a proper "large" input, e.g., see issue #296