Query : WORKLOAD RESULTS about coremark-pro HOT 3 CLOSED

If you look at the make flow, you will see that it calls a PERL script to compute the scores from the log files.

Please refer to these lines of code:

	'CoreMark-PRO' => {
		'factor' => 1000,
		'geomean' => {
			'cjpeg-rose7-preset' => (1 / 40.3438),
			'linear_alg-mid-100x100-sp' => (1 / 38.5624),
			'loops-all-mid-10k-sp' => (1 / 0.87959),
			'nnet_test' => (1 / 1.45853 ),
			'parser-125k' => (1 / 4.81116),
			'radix2-big-64k' => (1 / 99.6587),
			'sha-test' => (1 / 48.5201),
			'zip-test' => (1 / 21.3618),
			'core' => (10000 / 2855)
		}
	},

These reference scores were selected from an internal platform when the benchmark was written.

All scores, single and multi are computed by this script. The scaling is simply the ratio of the two scores.

Does this answer your question?

from coremark-pro.

I have attached log reports for cjpeg-rose7-preset.exe can you show me how we can calculate it,i understood for geomean and factor you are going fo follow pearl script.i want to understand the complete calculation for one binary because due to some issue i am simply running binaries on my target but not pearl script.
Thanks

from coremark-pro.

The result is the geomean of each score times the weights I provided above.

from coremark-pro.