intel

[stj]

time to give them a again!

this one looks big.
https://www.theinquirer.net/inquirer...threading-flaw

so if i'm reading it right,
the fake multi-core system is bugged,
and the "fix" is to disable it.
so your "super" i99 cpu operates as a single-core cpu.

LOL

[Topcat]

Re: intel

I read something about that the other day.....even worse than the 5v pentium 75, which ran so hot it would melt the ZIF socket.

[Stefan Payne]

Re: intel

Something like that they have right now with the Skylake X - they run really really hot, go straight up to ~100°C before they throttle and even with a 240 radiator you have no chance to getr the heat from the core because of the garbage heatsink compount Intel likes to use. Now on 1000$ CPUs...

[dmill89]

Re: intel

so your "super" i99 cpu operates as a single-core cpu.

You still can use all the cores on the CPU with hyper-threading disabled, but you can only use one thread per core, so an I7 basically becomes an I5, and an I3 basically becomes a Pentium. Still sucks that people paid for capabilities that cant be reliably used though, but not quite as bad as only being able to use a single core.

[stj]

Re: intel

if the product does not perform as advertised then customers should ask for a replacement or a refund - because it's defective.

you wouldnt buy a car and find it kept jumping out of 5th gear, only to be told by the maker that the "fix" was a new gate-plate that blocked the 5th gear from being used!!!

[dmill89]

Re: intel

if the product does not perform as advertised then customers should ask for a replacement or a refund - because it's defective.

you wouldnt buy a car and find it kept jumping out of 5th gear, only to be told by the maker that the "fix" was a new gate-plate that blocked the 5th gear from being used!!!

Agreed intel should either replace the defective CPUs or refund the purchase price, though interestingly with the car analogy auto makers have done similar things (though not something as major as blocking off a gear), one example was several years ago when GMs recall fix for defective washer fluid heaters (with heated windshield washer fluid being available as an option in several "Upmarket/Luxury" models) that could cause a fire was to disconnect the heaters and thus disable a feature that people had paid extra for (or at the very least was bundled with a more expensive trim level/package group), and NHTSA approved this "fix" stating that heated washer fluid was not a "required" feature and thus GM was free to disable it as part of the recall. They did give owners a "token" $100 compensation for disabling the feature, but this was generally far less than the package/trim originally cost.

[m4a2t0t]

Re: intel

Unfortunately its going to take a class action lawsuit before Intel gives anyone a dime back. They will just say to wait for a microcode update.
It's the same with the car, you would never get a refund, they would tell you to wait for a fix, if it took a month and you couldn't use your car it would take a class action suit before they compensated you for the month you had to rent a car.

I recently spent over $1K on a new paint job for my 3 year old car. The Manufacturer told me they would make a "good will" gesture and pay 75% of the repaint cost, leaving me with the $1K bill ($5K for the entire job). It was somehow 25% my fault...I guess for being stupid enough to buy a Honda. A few months later I get a Technical Service bulliten for the faulty paint. The manufacturer was, at that time, defending against a class action suit for the faulty paint jobs. Took me another month to get my money refunded.

I wrote all this an then see STJ is from Europe, EU protections seem to have actual teeth. In the US you can go fly a kite unless you have the money to hire a lawyer.

[stj]

Re: intel

yes, i remember many years ago seeing adds in imported dc comics for electrinic goods with a 90day warranty.
and thought WTF!

[goontron]

Re: intel

Well, lookie here, the first CPUs to not support Windows 7 and they are fucked worse than Windows 10! I had a feeling something ironic would happen there. To quote a site:

you will see more CPU bugs from Intel than you have in the past from the post-FDIV-bug era until recently.

Why?

Let me set the scene: It’s late in 2013. Intel is frantic about losing the mobile CPU wars to ARM. Meetings with all the validation groups. Head honcho in charge of Validation says something to the effect of: “We need to move faster. Validation at Intel is taking much longer than it does for our competition. We need to do whatever we can to reduce those times… we can’t live forever in the shadow of the early 90’s FDIV bug, we need to move on. Our competition is moving much faster than we are” - I’m paraphrasing. Many of the engineers in the room could remember the FDIV bug and the ensuing problems caused for Intel 20 years prior. Many of us were aghast that someone highly placed would suggest we needed to cut corners in validation - that wasn’t explicitly said, of course, but that was the implicit message. That meeting there in late 2013 signaled a sea change at Intel to many of us who were there. And it didn’t seem like it was going to be a good kind of sea change. Some of us chose to get out while the getting was good. As someone who worked in an Intel Validation group for SOCs until mid-2014 or so I can tell you, yes, you will see more CPU bugs from Intel than you have in the past from the post-FDIV-bug era until recently.
~ https://danluu.com/cpu-bugs/

And what do you know!
Atom bug https://www.servethehome.com/intel-a...ies-bug-quiet/
Baytrail/Broadwell bug: https://forum.manjaro.org/t/intel-ba...ux-kernel/1931

AMD's doing it too! https://lists.debian.org/debian-secu.../msg00084.html Hmm, VM to HM DoS. Sounds good!
And AMD's FDIV-esk bug:https://community.amd.com/message/2796982

Anyway, i think i've thrown enough bytes at this post. Have a nice mull over. Im goign back to bitbanging at an ARM bassed machine.

[Topcat]

Re: intel

^
All the more reason to keep that older hardware running as much as the software vendors try to force it into obsolescence. Even Mozilla is in on this. Has anyone noticed how poorly FF50+ performs, even on high-end hardware? There's no excuse why web browsing should bring a 12-core Westmere system with 192GB of RAM to it's knees loading a fucking browser. Intel forcing windows 10 progression is just as despicable as M$ and GWx. A processor has always been a 'gender neutral' device, allowing you to run whatever OS on it that you so choose....and it should always remain that way by the processor maker. Just wait till Intel processors won't even run *nix anymore....slave to M$...

[goontron]

Re: intel

^
All the more reason to keep that older hardware running as much as the software vendors try to force it into obsolescence. Even Mozilla is in on this. Has anyone noticed how poorly FF50+ performs, even on high-end hardware? There's no excuse why web browsing should bring a 12-core Westmere system with 192GB of RAM to it's knees loading a fucking browser. Intel forcing windows 10 progression is just as despicable as M$ and GWx. A processor has always been a 'gender neutral' device, allowing you to run whatever OS on it that you so choose....and it should always remain that way by the processor maker. Just wait till Intel processors won't even run *nix anymore....slave to M$...

I honestly don't think Intel is THAT stupid. They would lose the Cluster/Supercomputing market right then and there. Windows licensing for supercomputing would double, if not triple, the project costs.

[Topcat]

Re: intel

I honestly don't think Intel is THAT stupid. They would lose the Cluster/Supercomputing market right then and there. Windows licensing for supercomputing would double, if not triple, the project costs.

Don't be too sure.... It would just be a continuation of their norm, just with more restrictions.... Hypothetically, Processors targeted for the consumer market would be clipped/restricted to specific OS's, where as processors targeted for the IT/commercial market would have these restrictions removed, and of course carry a far heftier pricetag for basically the same damn processor. HMMM....I believe they called them Xeon & Itanium versus pentium & celeron in the earlier days...

[goontron]

Re: intel

From that point of view, yeah, i could see that. But they are also trying to get into the mobile market. The mobile market consists damn near 99% of Unix these days. iOS is BSD/Darwin, Android and may others (Sailfish, etc) are Linux based, and BBOS is QNX, another Unix clone.

It would be the exact opposite to what they are trying for. Then again, after seeing Microsoft do what they did, maybe im far off base thinking they wouldn't shoot themselves in the foot.

[Stefan Payne]

Re: intel

I honestly don't think Intel is THAT stupid.

With all those Intel Fanboys behind them, they can pretty much do whatever they want without any consequences.

Why the fuck would you save on production cost on a 1000$ CPU?
Because you can! And you don't have to fear any consequences.

And that's what you see on the new Skylake-X CPUs. Those are just garbage. Consume way more power than anything before while delivering worse 'IPC' and saving cost for the Solder for the Lid.

The Xeons on the 3600 whatever pin socket are soldered...
And Ryzen also is soldered...

[RJARRRPCGP]

Re: intel

saving cost for the Solder for the Lid.

Reminds me of Ivy Bridge, which was released in 2012 and Ivy Bridge wasn't better than Vishera, (FX) per-dollar, IIRC!

Yes, you heard that right, overall no better than an FX! (performance-wise)
(except for single-threaded performance, which is the Achilles' heel of FX)

Haswell was much better!

[stj]

Re: intel

soldered?
you mean the metal cap?

[dmill89]

Re: intel

Reminds me of Ivy Bridge, which was released in 2012 and Ivy Bridge wasn't better than Vishera, (FX) per-dollar, IIRC!

Yes, you heard that right, overall no better than an FX! (performance-wise)
(except for single-threaded performance, which is the Achilles' heel of FX)

Which of course was a much bigger deal in 2012/2013 than it is today, since multi-threaded support was much more limited back then. Worked out bad for AMD, FX may have been able to match/beat Ivy-Bridge with multi-threaded apps, but by the time multi-threaded app support caught up to the CPU architecture Haswell was out and FX was just seen as outdated and inefficient by then (which it was by that time).

[RJARRRPCGP]

Re: intel

Which of course was a much bigger deal in 2012/2013 than it is today, since multi-threaded support was much more limited back then. Worked out bad for AMD, FX may have been able to match/beat Ivy-Bridge with multi-threaded apps, but by the time multi-threaded app support caught up to the CPU architecture Haswell was out and FX was just seen as outdated and inefficient by then (which it was by that time).

Well, yep when gaming, sadly. Especially Halo CE... I particularly said "WTF" with Ivy Bridge, because of paste on the IHS instead of solder...