Spectre Mitigation Performance Impact Benchmarks On AMD Ryzen 3700X / 3900X Against Intel

[phoronix]

Phoronix: Spectre Mitigation Performance Impact Benchmarks On AMD Ryzen 3700X / 3900X Against Intel

AMD Zen 2 processors feature hardware-based mitigations for Spectre V2 and Spectre V4 SSBD while remaining immune to the likes of Meltdown and Zombieload. Here are some benchmarks looking at toggling the CPU speculative execution mitigations across various Intel and AMD processors.

http://www.phoronix.com/vr.php?view=28068

[johanb]

I wonder why the 7960X got faster WITH mitigations on hackbench...

[Michael]

I wonder why the 7960X got faster WITH mitigations on hackbench...

Likely an anomaly since it was 16 thread test and so the 7960X wasn't fully saturated and may have been some odd scheduler between HT/cores in that particular case.

[FireBurn]

Please can you fix the wall of text:

The default Intel mitigations namely come down to "l1tf: Mitigation of PTE Inversion + mds: Mitigation of Clear buffers; SMT vulnerable + meltdown: Mitigation of PTI + spec_store_bypass: Mitigation of SSB disabled via prctl and seccomp + spectre_v1: Mitigation of __user pointer sanitization + spectre_v2: Mitigation of Full generic retpoline IBPB: conditional IBRS_FW STIBP: conditional RSB filling." In the case of AMD Zen/Zen+ CPUs, the default software mitigations are "l1tf: Not affected + mds: Not affected + meltdown: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl and seccomp + spectre_v1: Mitigation of __user pointer sanitization + spectre_v2: Mitigation of Full AMD retpoline IBPB: conditional STIBP: disabled RSB filling." And then in the case of Zen 2 with Linux 5.2, "l1tf: Not affected + mds: Not affected + meltdown: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl and seccomp + spectre_v1: Mitigation of __user pointer sanitization + spectre_v2: Mitigation of Full AMD retpoline IBPB: conditional STIBP: always-on RSB filling."

[speculatrix]

I'm still concerned that we're yet to see more security problems caused by fundamental design decisions in the CPU cores.
I have noticed that when renting AWS EC2 instances, the ones powered by AMD are noticeably cheaper than Intel ones, and I've migrated pretty much all our servers over now. I wonder how much of the price benefit is due to reduced performance losses in AMD processors?

Amazon EC2 Instance Comparison

[andreano]

Funny how Intel and AMD performs much more similar with mitigations. That says something about where Intel's IPC advantage comes from, or should I say came from.

That said, desparate & necessary software workarounds probably aren't the most optimal tradeoff, so I wouldn't be surprised to see some of the loss regained if/when some of these get fixed in the hardware.

[starshipeleven]

Funny how Intel and AMD performs much more similar with mitigations. That says something about where Intel's IPC advantage comes from, or should I say came from.

That said, desparate & necessary software workarounds probably aren't the most optimal tradeoff, so I wouldn't be surprised to see some of the loss regained if/when some of these get fixed in the hardware.

I wouldn't be surprised if there is no regain.

[Jumbotron]

Just MORE proof that the advantage that Intel built for itself ALL these years was, in the words of the late, great comedian George Carlin, "Grade A Prime American Bullshit"

[Mark625]

Michael, in the intro page you list the two Threadrippers as "Ryzen 9 2950X" and "Ryzen 9 2990WX". Shouldn't those be just "Ryzen 2950X", or maybe "Ryzen Threadripper 2950X"? It took me a few seconds to figure out what you were talking about.

It would be nice if the charts listed them as "TR 2950X" etc, too.

Thanks.

[pipe13]

Just MORE proof that the advantage that Intel built for itself ALL these years was, in the words of the late, great comedian George Carlin, "Grade A Prime American Bullshit"

Mr. Carlin was not an x86_64 CPU designer and neither, one might surmise, are you.