I've noticed something unusual regarding the new Passmark V10 scoring for at least the 12-core Xeon E5-2697 V2. Somehow it has a substancially lower score than the 10-core E5-2680 v2 of the same architecture.
12669 multi-thread / 1822 single-thread on the 10-core E5-2680 v2
9338 multi-thread / 1742 single-thread on the 12-core E5-2697 v2
The single thread scores seem about right since the frequencies are slightly lower on the 12-core but that doesn't seem to account for the difference.
The older V9 test scores seemed much more accurate:
15792 multi / 1831 single on the 10-core E5-2680 v2
17021 multi / 1777 single on the 12-core E5-2697 v2
Any idea what's going on with this? The "Last 5 Baselines" scores seem all over the place.
12669 multi-thread / 1822 single-thread on the 10-core E5-2680 v2
9338 multi-thread / 1742 single-thread on the 12-core E5-2697 v2
The single thread scores seem about right since the frequencies are slightly lower on the 12-core but that doesn't seem to account for the difference.
The older V9 test scores seemed much more accurate:
15792 multi / 1831 single on the 10-core E5-2680 v2
17021 multi / 1777 single on the 12-core E5-2697 v2
Any idea what's going on with this? The "Last 5 Baselines" scores seem all over the place.
Comment