Intel Kernel design flaw / bug

A forum for discussing DataStage<sup>®</sup> basics. If you're not sure where your question goes, start here.

Moderators: chulett, rschirm, roy

Post Reply
thompsonp
Premium Member
Premium Member
Posts: 205
Joined: Tue Mar 01, 2005 8:41 am

Intel Kernel design flaw / bug

Post by thompsonp »

Any idea what impact the patches to Linux and Windows to address the Intel cpu flaw will likely have on typical DataStage workloads?

I've been reading that workloads with lots of network and i/o may be worst hit with up to a 30% performance drop.
qt_ky
Premium Member
Premium Member
Posts: 2895
Joined: Wed Aug 03, 2011 6:16 am
Location: USA

Post by qt_ky »

I read that yesterday also... in the 18% to 30% overhead range. That is significant!!! I'm glad we run ours on POWER8!
Choose a job you love, and you will never have to work a day in your life. - Confucius
chulett
Charter Member
Charter Member
Posts: 43085
Joined: Tue Nov 12, 2002 4:34 pm
Location: Denver, CO

Post by chulett »

Ah... hadn't heard of this yet but was just reading about their "Page Table Isolation" issue in articles like this one and this one. Thankfully, gaming performance doesn't seem to be effected but who games on a Linux box anyway? :wink:

There seems to be a lot of "depends" around this, for example from the second link:
Performance penalties from single to double digits are expected on patched kernels. The penalty depends upon how much interaction the application/workload deals with the kernel if there's a lot of context switching and other activity. If it's a simple user-space application not doing much, the x86 PTI additions shouldn't cause much of an impact. Newer Intel CPUs with PCID should also help in ensuring less of a performance impact.
-craig

"You can never have too many knives" -- Logan Nine Fingers
PaulVL
Premium Member
Premium Member
Posts: 1315
Joined: Fri Dec 17, 2010 4:36 pm

Post by PaulVL »

I just read about this topic. Nasty


pushing out patches to a raised floor area is never easy or fast. Big companies will most likely lag on this patching for their ETL environments.

But ya... it would be interesting to see the hit on DataStage.

If someone out in the wild is about to patch this... please share your results with us. Select a good meaty job to run. Not a 2 min job.

Do a pre/post patch run of course.
Application restart times logged, etc...
qt_ky
Premium Member
Premium Member
Posts: 2895
Joined: Wed Aug 03, 2011 6:16 am
Location: USA

Post by qt_ky »

Now it sounds like there are two vulnerabilities and one of them is not limited to Intel CPUs as it affects a lot of CPU types including IBM POWER...

https://www.ibm.com/blogs/psirt/potenti ... er-family/
Choose a job you love, and you will never have to work a day in your life. - Confucius
UCDI
Premium Member
Premium Member
Posts: 383
Joined: Mon Mar 21, 2016 2:00 pm

Post by UCDI »

we knew it was going to be brutal though. You can't really patch a CPU. The BIOS, but not the actual CPU chip. So they *have* to be jacking in a software solution that bypasses the hardware flaw which *has* to be a ton slower. This is on par with taking the math coprocessor off the chip for us old guys :)
asorrell
Posts: 1707
Joined: Fri Apr 04, 2003 2:00 pm
Location: Colleyville, Texas

Post by asorrell »

UCDI wrote:...This is on par with taking the math coprocessor off the chip for us old guys :)
I prefer "seasoned" or "experienced"... thank you! :-)

...and yes - this is going to be a real pain if your system is already stressed.

One of the major game platforms patched their Linux servers and said they saw a "hit in excess of 20% of performance capacity".
Andy Sorrell
Certified DataStage Consultant
IBM Analytics Champion 2009 - 2020
Post Reply