TPU: Intel Gags Customers from Publishing Performance Impact of Microcode Updates

[csbin]

https://www.techpowerup.com/247028/...shing-performance-impact-of-microcode-updates

Much of the secret sauce that made Intel processors faster than AMD is going sour, as the cybersecurity community is finding gaping security vulnerabilities by exploiting features such as speculative execution. Intel's microcode updates that mitigate these vulnerabilities impact performance. Intel isn't too happy about public performance numbers put out by its customers, which it fears could blunt the competitive edge of its products. The company has hence updated the license terms governing the microcode update distribution to explicitly forbid its users from publishing comparative "before/after" performance numbers of patched processors.

The updated license for the microcode update has this controversial sentence (pay attention to "v"):

"You will not, and will not allow any third party to (i) use, copy, distribute, sell or offer to sell the Software or associated documentation; (ii) modify, adapt, enhance, disassemble, decompile, reverse engineer, change or create derivative works from the Software except and only to the extent as specifically required by mandatory applicable laws or any applicable third party license terms accompanying the Software; (iii) use or make the Software available for the use or benefit of third parties; or (iv) use the Software on Your products other than those that include the Intel hardware product(s), platform(s), or software identified in the Software; or (v) publish or provide any Software benchmark or comparison test results."

Some of Intel's biggest enterprise customers are cloud computing providers such as AWS, Microsoft, and Google, who have made it their duty to keep their customers informed about the performance impact of microcode updated processors, since it impacts their cost/performance when the scale is big enough. This gag is both unethical, and probably even illegal.

 

[Nothingness]

That's incredibly stupid. I have little doubt Intel will backpedal. They have to.

 

[naukkis]

Fully patched performance is so poor that Intel don't want anyone to measure it.

So, is this great opportunity for eg Anantech to do a little benchmarking of their own

 

[richaron]

wat?

If anything this makes people more interested in the numbers they are trying to hide. It's like they've never been on the internet..

 

[TheELF]

This doesn't change anything for the market,anybody(and I do mean server etc) interested in upgrading to a new CPU will buy one ,do all their testing and then decide to buy the thousands they need or not,nobody is relying on 3rd party tests for that much money anyway.
This is a bad thing only for us end users who will be left wondering a bit longer because these numbers will come out eventually.

 

[french toast]

Classic Intel, nothing new here, move along.

 

[LTC8K6]

It's totally unenforceable, and the distribution part isn't going to work either.

I think the new guy wrote it.

https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=906158#14

 

[Cerb]

Why even bother? It's 100% bad PR, and nothing more. Unless you actually sign, or otherwise specifically and knowingly agree to such a license, it may as well not exist. As a user, that needs security updates for your software, and in a business situation really needs to have them (if TSHTF via malware, and you wren't updating your systems, good luck keeping liability for that off of your back), it wouldn't be worth the bytes it is wasting, even if you did have to agree to it.

 

[Arkaign]

Wonder if the penalties are similar to what AT found with testing on a Kaby box :

https://www.anandtech.com/show/12566/analyzing-meltdown-spectre-perf-impact-on-intel-nuc7i7bnh/2

Seemed to vary between minor and moderate impact, with nVME SSD being dramatically affected.

I know big metal back-office and cloud services hardware is a sort of Oranges to Tangerines comparison, but the raw gap pre and post microcode seem like they might share a ballpark.

Well, with this PR blunder, I'm sure we'll see a real world test to give it right back to Intel lol.

 

[Accord99]

Wonder if the penalties are similar to what AT found with testing on a Kaby box :

Intel published their own test here:

https://www.intel.com/content/www/us/en/architecture-and-technology/l1tf.html

The biggest hit is when you have untrusted virtual machines and you have to turn HT off.

 

[richaron]

Intel published their own test here: * snip*

*Slow clap*

 

[dainthomas]

Lol, good luck enforcing that.

 

[whm1974]

Lol, good luck enforcing that.

I can't see how they can if no one signed an agreement.

 

[formulav8]

Updated: Intel Answers Complaints About Microcode Benchmarking Ban

UPDATE: Intel has shared its new, and very brief, licensing agreement:

Redistribution and use in binary form, without modification, are permitted, provided that the following conditions are met:

• Redistributions must reproduce the above copyright notice and the following disclaimer in the documentation and/or other materials provided with the distribution.
• Neither the name of Intel Corporation nor the names of its suppliers may be used to endorse or promote products derived from this software without specific prior written permission.
• No reverse engineering, decompilation, or disassembly of this software is permitted.

Looks like the eula portion restricting benchmarking has been removed.

 

[EXCellR8]

Seems like whoever updated the license agreement...

A. Didn't really understand the latest request changes that came from marketing

or

B. Was experimenting with how promptly backlash would occur by customers

Either option is more believable than them thinking that what they were doing was unethical and/or possibly illegal.

 

[DaveSimmons]

Some lawyers are idiots about adding nonsense restrictions they make up without consulting technical staff or anyone with a working brain.

At work we sell site licenses for our software to universities, and some of the requested license changes we get back from them are, uh, creative? (Calling them "demented" might annoy our customers )

So this could have been an overreach attempt by an Evil Corporation, or it could just have been an incompetent lawyer trying to justify his salary by making bad changes.

 

[Ottonomous]

At a certain point Intel will just start suing individual consumers

 

[LTC8K6]

https://www.guru3d.com/news_story/intel_removes_benchmarks_conditions_for_linux_from_eula.html

Was either a test or just an error.

Since it was totally unworkable in more than one way, I think error.

 

[Borealis7]

i dont understand...even with the restrictions, whats preventing any ol' Joe Schmoe on the BananaTech forums from posting benchmarks and comparisons? the user is (almost) anonymous and Intel can't subpeana the forums for his IP, they are not the police. and BananaTech is in no way liable for the content it's users are posting. so whats the deal here?

 

[DaveSimmons]

i dont understand...even with the restrictions, whats preventing any ol' Joe Schmoe on the BananaTech forums from posting benchmarks and comparisons? the user is (almost) anonymous and Intel can't subpeana the forums for his IP, they are not the police. and BananaTech is in no way liable for the content it's users are posting. so whats the deal here?

The restriction no longer exists, read the post right above yours. It was probably a stupid mistake, nothing to see here, move along.

 

[rainy]

It was probably a stupid mistake, nothing to see here, move along.

Interesting, Intel doing lately a lot of stupid mistakes.

 

[WelshBloke]

So this could have been an overreach attempt by an Evil Corporation, or it could just have been an incompetent lawyer trying to justify his salary by making bad changes.

Do large corporations let single incompetent lawyers write their EULAs and then not check them?

 

[DaveSimmons]

Do large corporations let single incompetent lawyers write their EULAs and then not check them?

Yep, they also forget to renew their domains, let idiots post embarrassing tweets, run multi-million dollar ad campaigns suggesting you'll have sex with their hamburgers ("I'd hit it!"), on and on. People do dumb things.