LLamaSpeedTest2.cs

var modelParams = new ModelParams(@"C:\Users\Martin\Documents\Python\oobabooga_windows\text-generation-webui\models\llama-2-7b-chat.Q5_K_M.gguf");
using var weights = LLamaWeights.LoadFromFile(modelParams);
using var ctx = weights.CreateContext(modelParams);

var tokensArr = Enumerable.Range(1, 100).ToArray();

unsafe
{
    fixed (int* tokens = tokensArr)
    {
        var timer = new Stopwatch();
        timer.Start();
        {
            if (NativeApi.llama_eval(ctx.NativeHandle, tokens, tokensArr.Length, 0) != 0)
                throw new NotImplementedException("FAIL1");
        }
        timer.Stop();
        Console.WriteLine($"Batch ({tokensArr.Length}): {timer.ElapsedMilliseconds}ms");
    }
}