Intel Xeon E3-1270 v5 Skylake Linux Benchmarks

[phoronix]

Phoronix: Intel Xeon E3-1270 v5 Skylake Linux Benchmarks

Our latest Intel Skylake processor to benchmark is a Xeon E3-1270 v5 processor that boasts a boost speed of 4.0GHz...

http://www.phoronix.com/scan.php?page=news_item&px=Xeon-E3-1270-v5-Benchmarks

[Kano]

Did you use ECC ram here? That lowers memory speed compared to other results.

[Passso]

Michael : I suppose you run all your tests with default frequencies for CPU, right?
(if so as I overclocked mine I have to tweak for comparison)

[Michael]

Michael : I suppose you run all your tests with default frequencies for CPU, right?
(if so as I overclocked mine I have to tweak for comparison)

yes all stock speeds.

[Michael]

Did you use ECC ram here? That lowers memory speed compared to other results.

All my Skylake Xeons were with ECC memory.

[torsionbar28]

Did you use ECC ram here? That lowers memory speed compared to other results.

How so? Every benchmark I've seen comparing identical systems with ECC vs. non-ECC showed no difference at all in performance. I've heard the theory that the overhead of the ECC calculations results in a performance hit, but real world benchmarking shows the difference to be below the level of statistical significance.

[torsionbar28]

Michael typo in the article: "top of the pack" vs "top of the pact".

[Ericg]

I'll benchmark this against F23 with my Skylake i7-6700k and post a link, just for an additional data point.

[speculatrix]

is it possible to avoid some of the effects of running a different OS by using Docker to run OpenBenchmarking with a set of libraries and user-space binaries from the same distro?
then your only significant difference should be the kernel?