HT dead? Are you sure, Intel?

[igor_kavinski]

Intel® Core™ Ultra 7 Processor 255U (12M Cache, up to 5.20 GHz) - Product Specifications | Intel

Intel® Core™ Ultra 7 Processor 255U (12M Cache, up to 5.20 GHz) quick reference with specifications, features, and technologies.

www.intel.com

The P-cores having HT is the only way those threads can total up to 14.

So what's going on here? I thought they said HT was detrimental to their core performance yet here it is in all its glory.

I can't help shaking off this suspicion that Pat ordered the engineers to lie, lie, lie! Disabling HT saved them from validating it and enabled them to launch Arrow Lake early and that still didn't save Pat's neck. What a __________ (insert your favorite choice word).

 

[LazyGamer]

Who knows.

Given the security implications though, might be wise to discontinue their use and put efforts elsewhere now that thread counts are so high.

 

[igor_kavinski]

Given the security implications though, might be wise to discontinue their use and put efforts elsewhere now that thread counts are so high.

I personally think the security implications are blown out of proportion. If HT were a wide open vector for malware, AMD would be turning it off too. No matter how high the core counts, HT increases throughput and reduces idle resource time inside each core. That's worth turning it on or leaving it on.

 

[igor_kavinski]

Hmm...someone is saying that 255U is actually Meteor Lake on Intel 3. That would explain the HT availability.

 

[LazyGamer]

I personally think the security implications are blown out of proportion. If HT were a wide open vector for malware, AMD would be turning it off too.

I can't say for certain. We do know that HT as implemented in Skylake was a disaster, and an unintentional one at that - both that the design was a proof of concept not meant (by the designer) to be directly implemented, and that Intel never intended to be on 14nm producing Skylake derivatives for so long. So long that the entire world basically ran on Skylake when the vulnerabilities hit the public sphere...

But what we can say is that HT opens up unprivileged memory access, and that is an attack vector. So long as it's used, it will be targeted. We just don't and can't know how vulnerable AMDs architectures are to attacks that target HT.

No matter how high the core counts, HT increases throughput and reduces idle resource time inside each core. That's worth turning it on or leaving it on.

I don't disagree - but we do have Arrow Lake putting out solid compute numbers without it. So we have proof of both approaches being applicable IMO.