New box, new Mint, same CPU but only half the performance?

Questions about other topics - please check if your question fits better in another category before posting here
Forum rules
Before you post read how to get help. Topics in this forum are automatically closed 6 months after creation.
Locked
User avatar
bjornmu
Level 3
Level 3
Posts: 189
Joined: Wed Dec 19, 2012 2:50 am
Location: Trondheim, Norway

New box, new Mint, same CPU but only half the performance?

Post by bjornmu »

I have a strange problem with my new desktop. It has the same CPU as the old one: a 6-core AMD FX-6300 and even the same mother board ASUS M5A97 R2.0 (both boxes shipped without OS). The old box is running Mint 14, the new is on Mint 17.

The BogoMIPS number reported at boot is almost exactly the same for both, but the new box gets only about half the performance! I am running SETI@Home and the floating point benchmark there says 2617 MIPS per CPU on the old one, only 1366 on the new. I also ran a banal test program adding one billion different square roots, it takes twice as long on the new box. :? Integer arithmetic is also affected.

I can confirm that all cores are running on the max. frequency 3.50 GHz so that's not the issue, I can even hear the fans picking up when I start SETI@Home.

I do notice one difference. cpufreq_info -c 0 -d on the old box reports 'powernow-k8' as the driver, while on the new box it's 'acpi-cpufreq'. Can that be the problem? And if so how do I fix it?

In /var/log/kern.log/1 which I think is from the initial boot, I can see a number of lines like this:

Jun 7 14:01:03 igil kernel: [ 1.883896] acpi-cpufreq: overriding BIOS provided _PSD data
Last edited by LockBot on Wed Dec 28, 2022 7:16 am, edited 1 time in total.
Reason: Topic automatically closed 6 months after creation. New replies are no longer allowed.
User avatar
bjornmu
Level 3
Level 3
Posts: 189
Joined: Wed Dec 19, 2012 2:50 am
Location: Trondheim, Norway

Re: New box, new Mint, same CPU but only half the performanc

Post by bjornmu »

Now here's an even stranger thing: on the Mint 14 box it takes 2.5 times longer to run my simple floating point test program from mate-terminal that from xterm of from an ssh session from the other box. How can that be????

With higher optimization it still takes 50% longer to run on Mint 17 than 14.
Locked

Return to “Other topics”