Intel has backtracked on the license for its latest microcode update that mitigates security vulnerabilities in its processors – after the previous wording outlawed public benchmarking of the chips.
With the Linux 4.19 kernel that just kicked off development this month has been continued churn in the Spectre/Meltdown space, just not for x86_64 but also for POWER/s390/ARM where applicable. For getting an overall look at the performance impact of these mitigation techniques I tested three Intel Xeon systems and two AMD EPYC systems as well as a virtual machine on each side for seeing how the default Linux 4.19 kernel performance -- with relevant mitigations applied -- to that of an unmitigated kernel.
At the BlackHat conference last month, Christopher Domas demonstrated an attack against an x86 CPU using MSR's and an embedded RISC core to bypass ring protections. The full presentation "GOD MODE UNLOCKED - Hardware Backdoors in x86 CPUs" is viewable on YouTube. Is it time for CPU vendors to rethink security?
(Score: -1, Troll) by Anonymous Coward on Monday September 03 2018, @12:14AM (7 children)
Where is the summary of the results? An honest article would put them up front, or in the conclusion, but probably both. Since its missing I dont trust a word of this.
(Score: 0) by Anonymous Coward on Monday September 03 2018, @12:19AM (5 children)
Results for what? The ucode slowdowns for various simulated workloads are in the phoronix article (second link), there's no summary but you can clearly estimate percentages from the test results.
(Score: 0) by Anonymous Coward on Monday September 03 2018, @12:24AM (4 children)
There should be a summary of the results. Ie, "Intel Xeon performance is a% -b% worse depending on workload, AMD Epyc is c%-d% worse." There could even be another sentence about "mitigations had biggest impact (b/d%) for x type of tasks". Then people who are intrigued will read the details.
(Score: 2, Insightful) by Anonymous Coward on Monday September 03 2018, @12:53AM (3 children)
Sure but phoronix is a tech site, these are server tests and you can skim through the synthetic benchmarks yourself and see which is most applicable to your workloads. A summary based on synthetic benchmarks would be misleading as much as it may satisfy clickbait culture.
(Score: 0) by Anonymous Coward on Monday September 03 2018, @01:15AM (2 children)
Why did they split the article up into 5 pages? I think including a summary would be less clickbaity.
(Score: 0) by Anonymous Coward on Monday September 03 2018, @03:25PM
It's for the sake of loading more adds, I guess.
(Score: 0) by Anonymous Coward on Monday September 03 2018, @04:01PM
yes, they have a free with ads/paid without business model. pay up or shut your pie hole.
(Score: 5, Informative) by Anonymous Coward on Monday September 03 2018, @01:14AM
Did you look at all of the graphs of the performance differences for the multiple Intel and AMD chips, as well as two VMs? There were more than a dozen graphs, each representing real world tasks.
There is no single measure of the performance hit because it varies depending on what the system is doing.