Re: [opensuse] Time to look for a kernel update...
Hi, While a performance hit will likely be there, it is hard to say how big it will be. It really depends on the workload. Ciao, Marcus On Thu, Jan 04, 2018 at 08:12:19PM -0200, Jones de Andrade wrote:
Am I wrong, or does it mean that a performance hit is unavoidable?
Em 4 de jan de 2018 6:25 PM, "Marcus Meissner" <meissner@suse.de> escreveu:
On Thu, Jan 04, 2018 at 08:12:53PM +0100, Daniel Bauer wrote:
Am 04.01.2018 um 19:52 schrieb James Knott:
On 01/04/2018 01:50 PM, James Knott wrote:
You could suffer worse ... TYPICAL slowdown is supposed to be about >5%. Worst case is nearer 30% I believe ... I see there's a couple of software updates available now, kernel-firmware and microcode updates for AMD CPUs. Would these have anything to do with this problem.
It appears they do. Here's what the patch description says:
"openSUSE-2018-1 - Security update for kernel-firmware
This update for kernel-firmware fixes the following issues: - Add microcode_amd_fam17h.bin (bsc#1068032 CVE-2017-5715) This new firmware disables branch prediction on AMD family 17h processor to mitigate a attack on the branch predictor that could lead to information disclosure from e.g. kernel memory (bsc#1068032
CVE-2017-5715)."
So with intel this doesn't help and must not be installed?
For Intel the update is "ucode-intel".
it is just part of the fix, the kernel update is needed to implement it.
Ciao, Marcus
-- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
-- Marcus Meissner,SUSE LINUX GmbH; Maxfeldstrasse 5; D-90409 Nuernberg; Zi. 3.1-33,+49-911-740 53-432,,serv=loki,mail=wotan,type=real <meissner@suse.de> -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Am 05.01.2018 um 16:30 schrieb Marcus Meissner:
Hi,
While a performance hit will likely be there, it is hard to say how big it will be. It really depends on the workload.
Ciao, Marcus
On Thu, Jan 04, 2018 at 08:12:19PM -0200, Jones de Andrade wrote:
Am I wrong, or does it mean that a performance hit is unavoidable?
My first experiences on a intel-i7-computer: After updating the kernel and installing "these things" I personally in general do /not/ note a big decrease in speed of my computer. The only place I note a difference is in the raw-processing image editing program in my windows xp within a virtualbox. While win itself and the program in it start with the same speed as before, loading a 25MB-raw-photo is notably slower. Before it was there immediately after click, it now appears first pixelated and in a second step appears sharp, maybe after a quarter or half a second... Also applying adjustments in this program now has a small delay, very small, but noticeable while before adjustments appeared immediately. Of course this does not say much, I did not measure anything, it's just my feeling, and apart from win in virtual-box the "heaviest" thing I do is resizing a video from my mobile using ffmpeg. The rest is some maria-db databases, digikam with large database and thousands and thousands of images and videos, gimp, internet with many open windows, some of them loaded with scripts. In all this I don't /feel/ any difference. As Marcus said it probably depends on the workload and for sure soon there will appear serious tests, but for me with my common, manual desktop-use there is no need to be afraid of "tragic" performance loss. -- Daniel Bauer photographer Basel Barcelona https://www.patreon.com/danielbauer http://www.daniel-bauer.com -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 2018-01-05 17:18, Daniel Bauer wrote:
Am 05.01.2018 um 16:30 schrieb Marcus Meissner:
Hi,
While a performance hit will likely be there, it is hard to say how big it will be. It really depends on the workload.
Ciao, Marcus
On Thu, Jan 04, 2018 at 08:12:19PM -0200, Jones de Andrade wrote:
Am I wrong, or does it mean that a performance hit is unavoidable?
My first experiences on a intel-i7-computer:
After updating the kernel and installing "these things" I personally in general do /not/ note a big decrease in speed of my computer.
The only place I note a difference is in the raw-processing image editing program in my windows xp within a virtualbox. While win itself and the program in it start with the same speed as before, loading a 25MB-raw-photo is notably slower. Before it was there immediately after click, it now appears first pixelated and in a second step appears sharp, maybe after a quarter or half a second... Also applying adjustments in this program now has a small delay, very small, but noticeable while before adjustments appeared immediately.
Of course this does not say much, I did not measure anything, it's just my feeling, and apart from win in virtual-box the "heaviest" thing I do is resizing a video from my mobile using ffmpeg. The rest is some maria-db databases, digikam with large database and thousands and thousands of images and videos, gimp, internet with many open windows, some of them loaded with scripts. In all this I don't /feel/ any difference.
As Marcus said it probably depends on the workload and for sure soon there will appear serious tests, but for me with my common, manual desktop-use there is no need to be afraid of "tragic" performance loss.
From what I read, a program that runs within itself is not affected; for instance, most games. But a program that reads or writes to files, calling the kernel to do the ops, is affected greatly. -- Cheers / Saludos, Carlos E. R. (from 42.2 x86_64 "Malachite" at Telcontar)
On 01/05/2018 10:18 AM, Daniel Bauer wrote:
Of course this does not say much, I did not measure anything, it's just my feeling, and apart from win in virtual-box the "heaviest" thing I do is resizing a video from my mobile using ffmpeg. The rest is some maria-db databases, digikam with large database and thousands and thousands of images and videos, gimp, internet with many open windows, some of them loaded with scripts. In all this I don't /feel/ any difference.
As Marcus said it probably depends on the workload and for sure soon there will appear serious tests, but for me with my common, manual desktop-use there is no need to be afraid of "tragic" performance loss.
Just be glad you are not google or amazon or the like that purchased 4000 16/32 core servers to handle the load (based on pre-Meltdown) processor capability (as well as determining the number of routers, patch panels, etc.) No you are a big data company and find out you cannot support your current customer base because you have just been told all of the data you built your hardware requirements from were 30% inflated and, now that after the failed speculative execution reload fix, the additional registers and needed instructions for cleansing the processor cache data will leave you with a system that is too-small by 1/3 to meet your business needs. Them's the ones that are going to feel the impact the most. For you, and just about every other desktop user where your system runs at .02 - .05 utilization, you probably won't notice much. Now gamers that make a bizillion system calls per-second will be much more likely to seem scenery drag. It will be interesting to see where the impact is felt the most. -- David C. Rankin, J.D.,P.E. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On Monday, 2018-01-08 at 04:15 -0600, David C. Rankin wrote:
On 01/05/2018 10:18 AM, Daniel Bauer wrote:
Of course this does not say much, I did not measure anything, it's just my feeling, and apart from win in virtual-box the "heaviest" thing I do is resizing a video from my mobile using ffmpeg. The rest is some maria-db databases, digikam with large database and thousands and thousands of images and videos, gimp, internet with many open windows, some of them loaded with scripts. In all this I don't /feel/ any difference.
As Marcus said it probably depends on the workload and for sure soon there will appear serious tests, but for me with my common, manual desktop-use there is no need to be afraid of "tragic" performance loss.
Just be glad you are not google or amazon or the like that purchased 4000 16/32 core servers to handle the load (based on pre-Meltdown) processor capability (as well as determining the number of routers, patch panels, etc.)
No you are a big data company and find out you cannot support your current customer base because you have just been told all of the data you built your hardware requirements from were 30% inflated and, now that after the failed speculative execution reload fix, the additional registers and needed instructions for cleansing the processor cache data will leave you with a system that is too-small by 1/3 to meet your business needs. Them's the ones that are going to feel the impact the most.
For you, and just about every other desktop user where your system runs at .02 - .05 utilization, you probably won't notice much. Now gamers that make a bizillion system calls per-second will be much more likely to seem scenery drag.
Not gamers. Most games run internally, not doing system calls. They handle video directly, no?
It will be interesting to see where the impact is felt the most.
- -- Cheers, Carlos E. R. (from openSUSE 42.2 x86_64 "Malachite" at Telcontar) -----BEGIN PGP SIGNATURE----- Version: GnuPG v2 iEYEARECAAYFAlpTexgACgkQtTMYHG2NR9VDdQCeOih1kbdINOJlxJiY7RBCPJ+4 SY8AnRPcLtOKPf+6WC9IMT+n1hu0hcox =LNVa -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Hello, On Mon, 08 Jan 2018, Carlos E. R. wrote:
Most games run internally, not doing system calls. They handle video directly, no?
You've never done a 'strace' on any game then... How do you suppose games get stuff from internal data to the libs (mostly libGL) and then to the X driver (Mesa, AMD/NVIDIA drivers) and then to the kernel(sic!) module(s) and via that to the GPU... -dnh --
Blame directed at the wrong vendor tends to get reclassified as whining. No worries, I'm good at that too. -- Chris Hacking and Jay Mottern
-- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On Monday, 2018-01-08 at 20:55 +0100, David Haller wrote:
Hello,
On Mon, 08 Jan 2018, Carlos E. R. wrote:
Most games run internally, not doing system calls. They handle video directly, no?
You've never done a 'strace' on any game then... How do you suppose games get stuff from internal data to the libs (mostly libGL) and then to the X driver (Mesa, AMD/NVIDIA drivers) and then to the kernel(sic!) module(s) and via that to the GPU...
It was written at some of the sites linked these days, that games were not affected. Maybe proprietary games? Windows games? That the fps would be the same. - -- Cheers, Carlos E. R. (from openSUSE 42.2 x86_64 "Malachite" at Telcontar) -----BEGIN PGP SIGNATURE----- Version: GnuPG v2 iEYEARECAAYFAlpUnJwACgkQtTMYHG2NR9VeVgCgjZhO5dtdTi0Ji7wdVXKw5a/w UugAn0Al7jxrlI1qzc7gF14QcjrnV72j =RZQE -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 01/05/2018 10:30 AM, Marcus Meissner wrote:
Hi,
While a performance hit will likely be there, it is hard to say how big it will be. It really depends on the workload.
Ciao, Marcus
I am still unclear. Will all this all affect performance on my new AMD Ryzen processor? Thanks Mark -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 05/01/18 18:07, Mark Hounschell wrote:
On 01/05/2018 10:30 AM, Marcus Meissner wrote:
Hi,
While a performance hit will likely be there, it is hard to say how big it will be. It really depends on the workload.
Ciao, Marcus
I am still unclear. Will all this all affect performance on my new AMD Ryzen processor?
Who knows? Probably not much. Expected slowdown is probably about 5%. The Ryzen is immune to exploit 1, so that lessens the hit even further. The big question is what sort of programs are you running? Do you run disk-intensive programs? They'll get hit hard. Do you do home banking on the web? That's a big risk. It's probably a safe bet that your new Ryzen - even with the performance hit - will be faster than what it replaces. At the end of the day, nobody really knows at present. Bear in mind this wasn't even supposed to become public knowledge until next week, so everybody's been caught on the hop. And whatever the hit is, once a working fix is out there, the kernel devs will be working to reduce it ... :-) Cheers, Wol -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 01/05/2018 10:07 AM, Mark Hounschell wrote:
On 01/05/2018 10:30 AM, Marcus Meissner wrote:
Hi,
While a performance hit will likely be there, it is hard to say how big it will be. It really depends on the workload.
Ciao, Marcus
I am still unclear. Will all this all affect performance on my new AMD Ryzen processor?
How has your Ryzen been working for you, Mark? I've been thinking of replacing my 10-year old HP desktop containing an Intel Core(TM)2 Quad CPU. I've used bunches of Opterons in decades past and wouldn't mind ditching Intel again. Are you running 42.3 on it? Any problems? Regards, Lew -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 05/01/18 20:53, Lew Wolfgang wrote:
How has your Ryzen been working for you, Mark? I've been thinking of replacing my 10-year old HP desktop containing an Intel Core(TM)2 Quad CPU. I've used bunches of Opterons in decades past and wouldn't mind ditching Intel again. Are you running 42.3 on it? Any problems?
I'm building a new Ryzen system for myself. At the moment, it won't POST, and I don't have time to debug it :-( My current system is an Athlon X-III. I've bought myself a Ryzen-3. Something I was unaware of (but it probably wouldn't make difference) is that AMD's numbering system is now similar to Intel's - a Ryzen-3 is like an i3, Ryzen-5 to i5, and Ryzen-7 to i7. And 3 is a basic cpu, 5 is a hefty desktop cpu if you use your system for rather more than just the simple things, and 7 is high-end gaming, video-processing etc. Sounds like I should have gone for a Ryzen-5, but they cost a lot more money ... maybe I'll upgrade sometime ... Cheers, Wol -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA512 Wow, that's a lot of processors.... From https://security-center.intel. com/advisory.aspx?intelid=INTEL-SA-00088&languageid=en-fr The following Intel-based platforms are impacted by this issue. Intel may modify this list at a later time. Please check with your system vendor or equipment manufacturer for more information regarding updates for your system. Intel® Core™ i3 processor (45nm and 32nm) Intel® Core™ i5 processor (45nm and 32nm) Intel® Core™ i7 processor (45nm and 32nm) Intel® Core™ M processor family (45nm and 32nm) 2nd generation Intel® Core™ processors 3rd generation Intel® Core™ processors 4th generation Intel® Core™ processors 5th generation Intel® Core™ processors 6th generation Intel® Core™ processors 7th generation Intel® Core™ processors 8th generation Intel® Core™ processors Intel® Core™ X-series Processor Family for Intel® X99 platforms Intel® Core™ X-series Processor Family for Intel® X299 platforms Intel® Xeon® processor 3400 series Intel® Xeon® processor 3600 series Intel® Xeon® processor 5500 series Intel® Xeon® processor 5600 series Intel® Xeon® processor 6500 series Intel® Xeon® processor 7500 series Intel® Xeon® Processor E3 Family Intel® Xeon® Processor E3 v2 Family Intel® Xeon® Processor E3 v3 Family Intel® Xeon® Processor E3 v4 Family Intel® Xeon® Processor E3 v5 Family Intel® Xeon® Processor E3 v6 Family Intel® Xeon® Processor E5 Family Intel® Xeon® Processor E5 v2 Family Intel® Xeon® Processor E5 v3 Family Intel® Xeon® Processor E5 v4 Family Intel® Xeon® Processor E7 Family Intel® Xeon® Processor E7 v2 Family Intel® Xeon® Processor E7 v3 Family Intel® Xeon® Processor E7 v4 Family Intel® Xeon® Processor Scalable Family Intel® Xeon Phi™ Processor 3200, 5200, 7200 Series Intel® Atom™ Processor C Series Intel® Atom™ Processor E Series Intel® Atom™ Processor A Series Intel® Atom™ Processor x3 Series Intel® Atom™ Processor Z Series Intel® Celeron® Processor J Series Intel® Celeron® Processor N Series Intel® Pentium® Processor J Series Intel® Pentium® Processor N Series -----BEGIN PGP SIGNATURE----- iQEzBAEBCgAdFiEE7GM/Dul8WSWn72odQ1nEo4DFCIUFAlpP6zAACgkQQ1nEo4DF CIUgAggAmvnxTdroXWRcxrZOUcTF3k6qYR+AZ7Pn+ZsvbouMFy4gcyLyF/GjHntn mG5iYmn14eWozWt7AB2IKOzbQHngn6El9SKox/M9+ir0lcdQHwWLeLDNfjRrNOXL oIzh8REkpotaxmGA6MDry+k0L88a8iBT5mvJ+47kcGGRnPOPt9hP7wJG0EqKyySh gyfJIR+d74Zed/gIKiy141lidPRJLffwtKOLDJrwswVsxGAEUPWsF3wLYSBzwSt/ WyUOlHT65RylWqIDOK+CEu48o5YPYDziIMbNiMMv5XV/Vb71UKCLS6+hrBxkE1er 3+bcfz4yUc/nC70RrMAaIOpNIprgPA== =AjuN -----END PGP SIGNATURE----- N�����r��y隊Z)z{.�ﮞ˛���m�)z{.��+�: �{Zr�az�'z��j)h���Ǿ� ޮ�^�ˬz��
On 01/05/2018 03:53 PM, Lew Wolfgang wrote:
On 01/05/2018 10:07 AM, Mark Hounschell wrote:
On 01/05/2018 10:30 AM, Marcus Meissner wrote:
Hi,
While a performance hit will likely be there, it is hard to say how big it will be. It really depends on the workload.
Ciao, Marcus
I am still unclear. Will all this all affect performance on my new AMD Ryzen processor?
How has your Ryzen been working for you, Mark? I've been thinking of replacing my 10-year old HP desktop containing an Intel Core(TM)2 Quad CPU. I've used bunches of Opterons in decades past and wouldn't mind ditching Intel again. Are you running 42.3 on it? Any problems?
I am running 42.3 on it. There were some install issues. Simple to get around. Be sure to disable the IOMMU in the bios, then mem=4096M on the kernel command line at install time. The dist 4.4 kernel has IOMMU bugs in it and most hardware (at least all I've used) can do DAC properly so if ya don't limit your mem to < 4GB you will have issues. Once installed and updated you "should" be able to turn the IOMMU back on and get rid of the mem=4096M cmdline option. I can't say that for sure because I don't run the DIST kernels. I run the latest from kernel.org. I just don't know if the latest SuSE kernel has the IOMMU fixes in it or not. You will need an IOMMU if running 4GB+ memory. Be warned about the Ryzens though. There is a hardware "segfault" issue they discovered by doing something like building a kernel with "make -j16". I got mine back in Nov and had to RMA it. They sent me a new one fairly quick because they KNEW about the issue. Date codes after week 21 of 2017 should be good to go. Other than that I'm really happy with it. It is a AMD Ryzen 7 1700X. Mark -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 06/01/18 05:07, Mark Hounschell wrote:
On 01/05/2018 10:30 AM, Marcus Meissner wrote:
Hi,
While a performance hit will likely be there, it is hard to say how big it will be. It really depends on the workload.
Ciao, Marcus
I am still unclear. Will all this all affect performance on my new AMD Ryzen processor?
Thanks Mark
I'm with you. I have the FX-8150 cpu and am totally confused if the ucode-amd which was installed this morning (in TW) is necessary for my AMD cpu. BC -- Always be nice to people on your way up -- you'll see the same people on your way down. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
participants (10)
-
Basil Chupin -
Carlos E. R. -
Christopher Myers -
Daniel Bauer -
David C. Rankin -
David Haller -
Lew Wolfgang -
Marcus Meissner -
Mark Hounschell -
Wol's lists