sorry for the delay, got buried under a ton of work :( I haven't applied any patches to it (well except a couple of filters that I put into the default filters file).
I switched to file based cache, results are below, sadly no great improvement. It may help to know that I'm running openload with the default number of clients which is 5. load averages were before: 0.64, 1.16, 1.07 after: 2.60, 1.57, 1.21 openload output: MaTps 0.36, Tps 0.36, Resp Time 2.742, Err 0%, Count 1 MaTps 0.50, Tps 1.74, Resp Time 3.070, Err 0%, Count 6 MaTps 0.63, Tps 1.78, Resp Time 2.927, Err 0%, Count 11 MaTps 0.95, Tps 3.83, Resp Time 1.066, Err 0%, Count 25 MaTps 3.58, Tps 27.26, Resp Time 0.477, Err 0%, Count 57 MaTps 21.59, Tps 183.71, Resp Time 0.005, Err 0%, Count 242 MaTps 27.46, Tps 80.27, Resp Time 0.012, Err 0%, Count 351 MaTps 38.32, Tps 136.00, Resp Time 0.007, Err 0%, Count 487 MaTps 46.12, Tps 116.36, Resp Time 0.049, Err 0%, Count 620 MaTps 61.69, Tps 201.80, Resp Time 0.098, Err 0%, Count 822 MaTps 77.30, Tps 217.78, Resp Time 0.018, Err 0%, Count 1040 MaTps 86.47, Tps 169.00, Resp Time 0.023, Err 0%, Count 1209 On 11/22/05, Eugene Lazutkin <[EMAIL PROTECTED]> wrote: > > Could you just for kicks switch to "file:" cache and run tests again? I > wonder if there is a significant difference in response times. It may help > us to isolate the problem. > > One more question: did you apply MySQL backend patch > (http://code.djangoproject.com/ticket/463)? I don't think it should make a > difference under mod_python, but checking just in case.
