Intel processors not looking so brilliant anymore. Plus AMD's debt problems get better with each passing quarter

[RejZoR]

3 hours ago, leadeater said:

Absolutely not, these researchers were specifically looking in to such vulnerabilities contuning existing research that found spectre and meltdown. None of this was 'by chance'. I don't by chance find my car in the driveway when I go out to drive it, I mean it might not be there if someone stole it but I wouldn't by chance discover that either.

Jesus, do you people even context at all? Finding of these vulnerabilities wasn't down to years and years of poking Intel processors, they literally found it by chance now as they were looking at something specific. Not being by chance would be finding some specific trait that follows Intel processors for years and find vulnerability out of it. Which just wasn't the case here. Otherwise they'd find it years ago given they have existed years ago...

[Drak3]

5 hours ago, schwellmo92 said:

A 6600K is not very capable by todays standards, I would consider it low-er end.

5 hours ago, SupaKomputa said:

6600K is comparable to R3 1200, 4 core no HT.

4 cored i5 can no longer be called "high end" cpu.

5 hours ago, Albal_156 said:

 

They won't be much slower in the real world yes. But I take the view that 4 cores, though still enough for gaming these days probably won't be sufficient to run livestreams and do content creation and many games will start taking advantage of more cores soon. The enthusiast market will probably change unless Intel comes out with something and even that's under question as 10nm is looking unlikely. Intel will be back as they have deep pockets but it won't be for a while as they might just skip straight to 7nm.

4 hours ago, RejZoR said:

It's a quad core. Even if it has some clock advantage, it just doesn't have many cores. I mean, hell, I had a quad core back in 2008/2009 (Core i7 920). A decade ago. And that wasn't even the first quad core in existence...

 

If you're not a demanding gamer, it'll still work. But that's about it.

For the vast majority of users, gamers included, quad cores are still far more than proficient.

 

Games can only be made multithreaded so much, they are, by nature, highly serialized. What few games that do take advantage of anything higher than 4c/8t only do by narrow margin. And 4c/4t is still highly capable of tasks asked by users of the mainstream platform.

 

6C+ chips are wasted on the general consumer, and will be for quite some time.

[Blze001]

So have we reached the "I should panic and replace my 8700k with a 2700X" moment yet, or can I shelter my bank account a little longer?

[GoldenLag]

10 minutes ago, RejZoR said:

Not being by chance would be finding some specific trait that follows Intel processors for years and find vulnerability out of it

Hasnt the fact that speculative execution has been a point of interest for security researchers as it open up moments where it may be possible to exploit. 

 

Its not by chance, but looking at how things function and then figure out how it could be possible exploit it

[GoldenLag]

1 minute ago, Blze001 said:

So have we reached the "I should panic and replace my 8700k with a 2700X" moment yet, or can I shelter my bank account a little longer?

Zen 2  is around the corner so doing it right now would be a bad idea.

[RejZoR]

1 minute ago, Blze001 said:

So have we reached the "I should panic and replace my 8700k with a 2700X" moment yet, or can I shelter my bank account a little longer?

Nah. 8700k is still very capable CPU. I'm on 5820K and can't really justify any new Ryzen yet.

[Drak3]

4 minutes ago, Blze001 said:

So have we reached the "I should panic and replace my 8700k with a 2700X" moment yet, or can I shelter my bank account a little longer?

Not going to happen. When Ryzen 3xxx drops, 16c/32t, suddenly every octocore will magically become worthless.

 

Just like what happened to Quadcores when Ryzen dropped, according to some in this thread.

[leadeater]

50 minutes ago, RejZoR said:

Jesus, do you people even context at all? Finding of these vulnerabilities wasn't down to years and years of poking Intel processors, they literally found it by chance now as they were looking at something specific. Not being by chance would be finding some specific trait that follows Intel processors for years and find vulnerability out of it. Which just wasn't the case here. Otherwise they'd find it years ago given they have existed years ago...

That's not how it works, they come up with a hypothesis based on the architecture documentation and years of information about how the product functions and then set out to test that hypothesis. That's in no way by chance.

 

By chance is finding a $100 note on the ground, unless you specifically went for a walk to find money on the ground.

 

Time between flaw existing and flaw being found has zero relationship to something being chance. Edit: I mean what you're saying is all science and human development has only been by chance because things have taken a long time and took us a long time to discover them. Like say when we do find a cure for cancer, we've been trying for over 30 years and you're saying when we do it's only by chance. It may actually end up being by chance, that is a possibility but odds are with how many scientists looking for the cure that it will be found not by chance but continued research. 

 

You know some things are just hard to find, or hard in general. So deep and obscure there was not enough understanding to go look for it in the way that has been until we obtained that understanding. That's how things can go years, decades without being found.

[williamcll]

Honestly I just want better single core performance on Ryzen.

[LukaH]

watch out intel you thought Zen was scary, wait until the FX CPUs catch up 

[RejZoR]

But that's not how it happened for Intel. No one found these things because they go across so many product generations. They found them now and also found they are exploitable on products going several generations back. It very much makes a difference. Also, if it goes so far back, why discovering them now when OS itself is far more reinforced and secure than years ago?

[Alan G]

With the caveat that I am a long term Intel shareholder, my take is that AMD are way over valued.  Their own guidance implies that first half revenues may be down 18% from last year.  They are getting pounded by NVIDIA in the GPU space and other than sales to gamers of Ryzen CPUs they are not make much of an impact on Intel sales.  They are limited in terms of R&D budget and don't have manufacturing capability to scale things up as does Intel.  Their own CEO, Lisa Su, said it will take 4-6 quarters to see if they are having an impact on the server market.  AMD's cash flow and balance sheet are still not what I would expect of a company that is moving forward (they have only had one year of positive cash flow in the past five).  Until they can sustain a couple of more years of increased EBITDA, I'm not terribly interested in an investment.

[leadeater]

4 minutes ago, RejZoR said:

But that's not how it happened for Intel. No one found these things because they go across so many product generations. They found them now and also found they are exploitable on products going several generations back. It very much makes a difference. Also, if it goes so far back, why discovering them now when OS itself is far more reinforced and secure than years ago?

How does that make it by chance? it doesn't.

[RejZoR]

1 minute ago, leadeater said:

How does that make it by chance? it doesn't.

Because it wasn't found by a specific pattern. They did the reverse. Found the bug and then discovered its trail/pattern goes way back. That's what I'm saying the entire bloody time and what I'd consider as "by chance" and not because of design that lead them through generations till latest ones.

[leadeater]

11 minutes ago, RejZoR said:

Because it wasn't found by a specific pattern. They did the reverse. Found the bug and then discovered its trail/pattern goes way back. That's what I'm saying the entire bloody time and what I'd consider as "by chance" and not because of design that lead them through generations till latest ones.

No it was found by a specific pattern, that's literally the research track that was being done that found these flaws. They were specifically looking at CPU data caches and how they act to see if they can find security flaws in that area. The way this has worked has been like that for years but as I mentioned earlier is very complicated and we first must obtain the understanding to find the flaws before we can discover them. They could have been found sooner but they weren't, that doesn't make it by chance unless it was actually discovered by chance which is not the case.

 

These researchers were specifically looking in to this specific area, they specifically found these flaws. That's very specific to be chance. It's almost like they had some prior reasoning to be looking in that area.

[MeatFeastMan]

6 minutes ago, VegetableStu said:

hey uhh quick question: is there any hardware expected from intel during computex? apparently they have a keynote on tuesday o_o

(all I can gather so far are Cascade Lake X and a few laptops, but not much else...)

14nm+++++++++++++++++++ is probably ready and waiting for release. Comet Lake has been rumored with a 10 core cpu. I mean, if they don't have anything to show us then they are in trouble because the zen2 steamtrain is about to arrive at the station.

[Guest]

7 hours ago, Albal_156 said:

Well looks like Intel's chips have been hit hard first by Spectre and Meltdown vulnerabilities and now by MDS vulnerabilities as well. AMD has been hit by their mitigations for Spectre but since Intel has had now 5 vulnerabilities to patch it seems like they might not be the default CPU in the market anymore particularly now AMD has a process node lead over them too. Though it seems 8th gen and 9th gen processors are not affected as much by this. But they are still significantly more than AMD's offering and they may now be more similar in performance

The way things are going its just time before hackers move from intel to amd anyways.

[CTR640]

When I'll have to ugrade the CPU, it will be AMD. I can't be arsed to feed the Intel's greed. And their shitty tactis also prevents me to buy Intel.

[Master Delta Chief]

5 hours ago, schwellmo92 said:

Depending on your game, some games you’ll get noticeable frame stuttering in heavy scenes which isn’t acceptable for some of us. If that guy wants to upgrade from a 6600K that is a perfectly acceptable reason to upgrade, it is a barely capable processor by today’s standards.

The i3-8350K outperforms the 6600K to some extent actually according to some benchmark sites. They also have similar specs. If you're just a casual gamer that's on a budget, you might wanna go for the 8350K if you're sticking with Intel. 

[maartendc]

7 hours ago, GoldenLag said:

You have a 6600k. Any upgrade to current AMD hardware isnt worth it. 

I beg to differ. I had a 6600, non K, and it held me back significantly in games. (Battlefield V in 64 player MP). A 6600 is a 4-core, 4-thread CPU. That just doesnt cut it anymore these days I'm afraid. If we were talking 6700K, maybe. The Hyperthreading helps a bit. But 4 threads... really getting long in the tooth.

 

Went to a Ryzen 2700X, best money I ever spent.

[thorhammerz]

1 hour ago, Mr. horse said:

I would pay good money for a CPU that is built for single core performance and does not care about multi core performance.

Like a dual or maybe a quad but most likely a dual core with lots of low latency l2 and a very high clock speed.

Overclocked i3-7350k/8350k?

[RejZoR]

15 minutes ago, Mr. horse said:

Yeah I was thinking of getting a system built around one of these. But It would be cool if such a CPU was on something like x299.

There are people out there like me that need faster CPUs for single threaded apps.

Why the hell would you stick an i3 on a X299 ?!

[Drak3]

14 minutes ago, Mr. horse said:

Not necessary an i3 but a single threaded optimized CPU with some PCIe lanes to back it up. The lower power draw would alow for much higher OCing with a better x299 board and quad channel ram support would be nice too.

Yeah no. Unless you're running a silicon lottery loser HCC chip, overclocking is roughly the same between mainstream and HEDT, save for 100MHz past the point of deminished returns.

Edited May 20, 2019 by Drak3

[Drak3]

Just now, Mr. horse said:

Would you like to explain how a top end board with better VRMs and witch is likely more geared for overclocking would be no better then a mid range board with weaker VRM and lesser quality parts?

Simple. The "mid range" boards already exceed what the CPU can handle.

[thorhammerz]

44 minutes ago, Mr. horse said:

Yeah I was thinking of getting a system built around one of these. But It would be cool if such a CPU was on something like x299.

i5-7640x or i7-7740x?