The Downfall Mitigation Impact For Xeon E-2300 Series On Linux 6.5

Written by Michael Larabel in Software on 25 September 2023 at 10:40 AM EDT. Page 3 of 3. 3 Comments.
libavif avifenc benchmark with settings of Encoder Speed: 0. Vulnerable, No Microcode was the fastest.
libavif avifenc benchmark with settings of Encoder Speed: 2. Vulnerable, No Microcode was the fastest.
OSPRay Studio benchmark with settings of Camera: 1, Resolution: 4K, Samples Per Pixel: 1, Renderer: Path Tracer. mitigations=off, Latest ucode was the fastest.
OSPRay Studio benchmark with settings of Camera: 1, Resolution: 4K, Samples Per Pixel: 16, Renderer: Path Tracer. mitigations=off, Latest ucode was the fastest.
OSPRay Studio benchmark with settings of Camera: 1, Resolution: 4K, Samples Per Pixel: 32, Renderer: Path Tracer. Vulnerable, No Microcode was the fastest.
OSPRay Studio benchmark with settings of Camera: 2, Resolution: 4K, Samples Per Pixel: 32, Renderer: Path Tracer. mitigations=off, Latest ucode was the fastest.

AVX user-space workloads haven't tended to be impacted much by the other CPU security vulnerabilities of recent years, so running mitigations=off is quite similar to the "Vulnerable, No microcode" results.

Neural Magic DeepSparse benchmark with settings of Model: NLP Text Classification, BERT base uncased SST2, Sparse INT8, Scenario: Asynchronous Multi-Stream. mitigations=off, Latest ucode was the fastest.
Neural Magic DeepSparse benchmark with settings of Model: NLP Text Classification, BERT base uncased SST2, Sparse INT8, Scenario: Asynchronous Multi-Stream. mitigations=off, Latest ucode was the fastest.
Neural Magic DeepSparse benchmark with settings of Model: ResNet-50, Sparse INT8, Scenario: Asynchronous Multi-Stream. mitigations=off, Latest ucode was the fastest.
Neural Magic DeepSparse benchmark with settings of Model: ResNet-50, Sparse INT8, Scenario: Asynchronous Multi-Stream. mitigations=off, Latest ucode was the fastest.
Neural Magic DeepSparse benchmark with settings of Model: BERT-Large, NLP Question Answering, Sparse INT8, Scenario: Asynchronous Multi-Stream. Vulnerable, No Microcode was the fastest.
Neural Magic DeepSparse benchmark with settings of Model: BERT-Large, NLP Question Answering, Sparse INT8, Scenario: Asynchronous Multi-Stream. Vulnerable, No Microcode was the fastest.

For those making use of Neural Magic's DeepSparse AI software, the inference performance has been impacted by the Downfall mitigations as also seen on the Ice Lake and Cascade Lake server benchmarks from August.

Geometric Mean Of All Test Results benchmark with settings of Result Composite, Intel Downfall Xeon E. Vulnerable, No Microcode was the fastest.

With the affected workloads tested, the Xeon E-2366 server with the updated kernel and microcode was running at around 91% of its prior level of performance. At least for those very concerned about performance loss and not too concerned about the Downfall exposure, booting with mitigations disabled -- either selectively disabling Downfall or using the mitigations=off hammer -- is capable of restoring the prior level of performance in these computationally heavy workloads.

If you enjoyed this article consider joining Phoronix Premium to view this site ad-free, multi-page articles on a single page, and other benefits. PayPal or Stripe tips are also graciously accepted. Thanks for your support.


Related Articles
About The Author
Michael Larabel

Michael Larabel is the principal author of Phoronix.com and founded the site in 2004 with a focus on enriching the Linux hardware experience. Michael has written more than 20,000 articles covering the state of Linux hardware support, Linux performance, graphics drivers, and other topics. Michael is also the lead developer of the Phoronix Test Suite, Phoromatic, and OpenBenchmarking.org automated benchmarking software. He can be followed via Twitter, LinkedIn, or contacted via MichaelLarabel.com.