New CacheOut Speculative Execution Vulnerability Hits Intel

just an fyi i have not suffered this problem on linux, performance is largely the same maybe 10% loss max. using a penryn quad.

 

It’s only a matter of time before Intel CPUs have to be ran with more security holes than Swiss cheese to stay “fast”.

 

2500K user checks in.

Old Intel CPU gang rejoice!

 

You should.
How many rogue VM's do you think is trying to exploit these vulnerabilities?
Simply to gain access to data from other VM's in the same system ?

 

If you actually read the article, this CacheOut problem only applies to Skylake CPU's and up.
Core 6xxx series and up.
not down.

Check this list :
https://software.intel.com/security...hts/processors-affected-l1d-eviction-sampling

Ivy Bridge are NOT affected.

 

AMD was beating the 8700K even with the 2700X. Zen2 is eating the 8700K in almost any way.

 

No, 2700x are completelly destroyed even the 3900x hardly you can see beating the 8700k

https://www.techpowerup.com/review/amd-ryzen-9-3900x/15.html

 

AMD must love this free advertising

 

Oh hell....... Here we go again!.!.!

WTF Intel!?! All of these performance gains over the years are being negated by all of these hardware level issues.

Sad as hell. Just crazy that AMD is untouched by any and all of these issues..... Conspiracy!?! I hope not....

 

It's fortunate for us that AMD is making really good CPUs these days.

 

Not one attack or malware in the wild that has exploited any of the vulnerabilities since Spectre/Meltdown. The latest CA microcodes for my 8700K were kind of disastrous, they reduced the Uncore multi to 43x at defaults and dropped about 4-5% performance in a few benchmarks, both CB versions included. First microcode to have such a dramatic effect.
Planning to just return to the previous BIOS and stop patching these things. It's just getting ridiculous, especially if you don't really have super-secret info you could lose.

 

Maybe you should have read that paper ...

Well, that bold part actually translates to 10% difference. If that means "completely destroyed", what should someone with 10% higher IQ than you say about you?

Linking gaming category and ignoring all other uses. Same applies to 3900X.
Except that 3900X is not even gaming CPU. It is productivity CPU. And there it comes cheaper than i9-9900K.

Tell me which is better?
i9-9900K because it delivers 5% higher fps than 3900X @3,8GHz in average?
Or 3900X that finishes all your productivity through week quite faster and lets you play 6 hours more?

And mind that today 3900X all of 12C/24T boost to 4,175~4,2 GHz under full load. And that you can actually game while you run quite some stuff in background. Therefore your actual time when you can game is not just few hours more than with i9-9900K.
Then TPU test for Zen2 used 3200MHz memory => 1600MHz IF instead of today's quite easily achievable 1800MHz+. (I got there with 4 memory sticks.)

 

Ignorance is bliss? I remember when NCIX went out of business and read an article several years later about how customer data was being sold off. Then there was the Newegg hack which stole payment information from customers as they were checking out. My credit card was implicated in both and I immediately cancelled it - nobody had tried to use it yet, but I felt it was better to get rid of it anyways since I knew it had been compromised.

Those who leave a known exploit unpatched are basically deciding not to cancel a compromised credit card. It doesn't matter if someone's tried to use it or not, if you know it's been compromised then you should do something about it (and if the cost is a bit of performance loss then so be it). The argument of "I don't have anything important on my computer" is also moot - if you do online banking or have website login information stored on your PC (as anyone who uses their PC frequently undoubtedly will) then you are putting that information at risk.

 

Many of the vulnerabilities require admin access to the PC. I am the only one accessing it. It's been 15 years+ since my info has not been compromised in any way, because I understand how the web works. My plan is to keep the Meltdown/Spectre/L1TF/MDS patches in places, because they have basically 0 impact on performance for my PC. I'm done with patching what's next though for sure, I'll try the new microcode for CacheOut when it's released, and if the performance hit is still there, I'm revering back to B4 microcode.

It will do fine for 1-2 years until I switch the 8700K, and in the end, it's a very, very small risk I am willing to assume.

 

Many of them can be executed by script in web page you load. In worst case scenario, one of those advertisement banners you see on this very site will have payload that will get your banking certificate details. Or credit card info from other tab. Or anything else they may target in your system.

To put you into reality of your situation. You may be only one who sits in front of your computer. But you have no control and knowledge over what is executed on your CPU.
Don't get defensive just yet, none of us have that control. Not even on linux, that's nature of using web.

 

What you are doing is basing your decision on the performance impact, and then rationalizing it afterwards - which is completely backwards. You should think of security regardless of performance.

My guess is that if there was a critical exploit that allowed anyone to easily gain access to your PC, but the fix resulted in a 10% performance hit to games, then you'd reject it.

 

not really that surprised anymore how ppl "weigh" things.
working at a parts shop i had ppl buying stuff like 15$ brake pads below oem/stock,
because of cost, when proper ones (around 30$) would not only brake better and last longer,
but probably save your life, by stopping the car before you hit something.

and even when i explained they would easily spend more money, over the same time frame,
just by buying 0.3$/gal more expensive gas.

 

sorry but i don´t use productivity tools, i don´t play cinebench or winrar benchmark, i using my pc to play games on fullhd 144hz, and 3900x are slower than old i7 8700k, i7 9700k and i9 9900k

 

running 1080p.
once you take that a way and play a 1440p/144Hz, they arent.
so basically anyone that looks at everything else, will be better off, which is a much bigger market.

i havent seen someone i know (personally) spending money on a i7 or higher, a decent gpu, and then still runs 1080p (not matter if fps or not, no matter if larger screen and/or res),
as your wasting a lot of HP.

 

Then explain why are you bringing productivity CPU into discussion instead of something like 3700X? Maybe because that's quite cheaper and is still just some 5% slower than i9-9900K?

5%, contemplate that.

Then again, you brought into discussion 2700X. Even new one costs less than half of what 9700K costs. You are justifying more than double price by 10% higher achievable performance in games when fps is uncapped.
And guess what? In 7 out of 10 games used there, all do get you quite more than 144fps therefore all that bonus you may get is capped by your screen.

And again, if any of those Zen2 chips were configured properly and with newer BIOSes, they would win in some of those games.