So let me get this straight: Intel processors have been faster than AMD's all this time only because they've been...

So let me get this straight: Intel processors have been faster than AMD's all this time only because they've been "cheating" by doing something risky security-wise, and now their luck has caught up with them.

Is that right?

Other urls found in this thread:

techreport.com/news/8547/does-intel-compiler-cripple-amd-performance
youtu.be/SqbNWCehbmU
twitter.com/SFWRedditImages

Yes.

...

...

Yes and more

techreport.com/news/8547/does-intel-compiler-cripple-amd-performance

No. Both processors do the same thing. This has been explained many times.

When they're being fed instructions, they sometimes receive some that are conditional on another instructions. Because discarding these (possibly unnecessary, possibly not) instructions saves no time vs just executing them and waiting to see of they're needed (if they are this saves time), they just do the work and throw it out if it's not useful.

The problem with Intel is that these "speculative processes" are checked for security AFTER they execute. So they can for example access information that the program has no access to. Remember that the memory management unit in a cpu assigns memory for a program, and will return an error called a page fault when the program goes out of bounds. Intel doesn't do this until the speculative process has already read the data it should have no access to, by which point it's too late.there is nothing a program can't access pretty much trivially on the computer

AMD CPUs according to AMD, checks it's safe before running, anf won't speculate on page faults

Intel isn't "doing something risky" security wise, they're handing out keys to the kingdom and telling everyone to go wild.

Someone with bigger brain can correct me if I missed something

>Intel isn't "doing something risky" security wise
>they're handing out keys to the kingdom and telling everyone to go wild.
>which isn't risky at all
>security wise

top kek

"Doing something risky" implies they're trying to squeeze extra performance and are not just retards that fucked up. They get nothing out of doing things this way other than the satisfaction of being retards.

>Intel isn't "doing something risky" security wise
>they're handing out keys to the kingdom and telling everyone to go wild.
>which isn't risky at all
>security wise

>They get nothing out of doing things this way other than the satisfaction of being retards.
>30% performance

Not him but:
That's the best part. If they had done things properly and checked before there would be almost 0% difference in performance.
The duct tape emergency jerry rig to go around this shit is what's costing them 30%, it's not even they were smart about this they just fucked up

>AMD Airline: Security Check before boarding
>Intel Airline: Security Check after landing
>Both the same

They must have had some reason not to do the check beforehand, any engineer would have pointed out the flaw with their planned model

so intel = shit & amd = good?

do i buy amd right now?

Is there even that mch difference performance wise right now anyway? Unless you're going super high end amd and Intel compete pretty evenly, and amd uses less power now

So yeah, go amd

no.
just fucking wait till the dust has settled

yes, buy now before prices go up

>AMD Airline: Security Check before boarding
>Intel Airline: Security Check after bombing

Yes. That reason was more intelligence agency access to hardware level backdoors and in return they corner the market and monopolize it due to the increased speed the would gain on their chips.

One is a risk, the other is a certainty.

You are right, except in the part that you are saying that they are the same and Intel isn't doing anything risky.
Nice try Brian.

S-should I buy a Bristol Ridge Thinkpad?
I'm poor.

>Bristol Ridge
wtf is that? sounds like some $4 bottle of wine

not like jews would lie...

if those missing security feature are included in new cpus, then they'll need more power and more cooling for the same speed. intel housefire incoming

AM4 Bulldozer Based APUS before the glorious Raven Ridge.

whats the equivalent AMD to a i7-5930k

>amd cpu's rise in price from intel fallout
>intel's cpu's drop in price from people leaving
This is a good thing if you like getting deals on cheap ass cpu's

>6 cores
>hyperthreaded

R5 1600.

ideal to wait for the new refresh or buy now?

Can't wait for the flood of used enterprise CPUs

Will be they worth?
I don't mind pick a Xeon V4 for my next build, but really I'm not sure.

No you idiot. The software fix is what causes the drop in performance. If they could fix the hardware there would be no impact.

No, Intel has better silicon...for now. Clock speed accounts for nearly the entire difference in single core performance between something like a 7700k and a 1800X.

It's a refresh, so don't expect a big leap. 10% overall performance boost maybe, I don't think it's going to be a big deal. If you see a discount on one, go for it.

t. bought 1700 on release with no regrets

There's what, 5-10% gap in pure IPC? I believe Ryzen is roughly on the level of Skylake-X.

my FX 8350 is so goddamn comfy

Intel Ryzen?

kek

>tfw all AMD chips (even bulldozer) will now have a higher resale value

time to sell my old build

So let's demand a recall then.

Not entirely, Vishera/Bulldozer was rushed shit from AMD. Secure, but rushed shit with weird inefficient cores, shit performance, shit thermals. For this reason, Intel basically [BLACKED] AMD for a few years until they got their shit together and made the new "Zen" series of processors.

Then this exploit was found and at this point, some FX (bulldozer/vishera) processors may now even out perform some patched i5s and i3s.

That's literally impossible. Best we can hope for is a lawsuit.

not really. intel processors are still faster .

>tfw you own your competitor so hard that even after a crippling 30% performance hit from a serious bug you still own them

Holy shit is this real? What benchmark is that?

Dont forget, AMD did optimize microcode of lot of these faildozer processors with microcode updates and shit.
Hell, I'm considering a low power Am3 Opteron for my NAS. I bet it would be sufficient.

Actual IPC gap is kinda hard to measure, and theoretically AMD should have the IPC advantage according to certain researchers. If you compare Skylake-X and a Ryzen processor with the same number of cores and at the same clock speed performance is extremely similar, but there are per-workload variances.

However, Intel's 14nm++(++?) is definitely superior to GloFo's 14nm, which allows Intel to enjoy a significant clock speed advantage.

>today in things that didn't happen.

Absolutely disgusting. shintel fanboys will obviously laugh off this one.

This really has me worried. I don't like AMD, the acronym sounds so stupid, and I feel intelligent when I buy Intel cuz it sounds like intelligence.

Also, I don't like the color red. Yucky.

Spoofed CPU vendor ID
The benchmark is compiled with Intel's ICC compiler which optimizes compiled code for Intel processors and leaves out certain optimizations for non-Intel processors, so faking the vendor ID on an AMD chip to "GenuineIntel" gives the processor a much higher bench score when compared to using the default "AuthenticAMD" vendor ID.

No, the wine you're thinking of is Harlem Ridge

>they are still doing that shit

So fucking fake

No, for desktop users the performance difference is likely to be less than 5%. If Intel suddenly lost 30% across the board they'd be completely fucked.

[laughs in Ryzen]

youtu.be/SqbNWCehbmU
Frances Intel hub is caught red handed altering algorithmic kernels with the Dvorjak code I8SUMGR1M3
WHEN WILL THIS MADNESS END

Also BogForce5 codebreakwr schematic is still being developed and all our chips are compromised

Nothing anyone can do about it. Reviewers will continue to ignore the obvious bias in benchmarking programs either because they are "sponsored" by Intel or because they just don't care.

microcode updates? like you could flash an update to your processor or do you mean later series of processors had it?

Microcode is contained in the BIOS or UEFI firmware.

So how do I disable this security check on my AMD processor for the speed boost?

That pic triggers me

In linux is easy as put nopti in your kernel parameters.
In Windows no idea. I dont know how to pass parameters to NT kernel in windows. I bet that's a undocumented way or edit the BCD to pass the parameters to get the kernel doing what you want.

Never have I seen a faker post.

thanks for explaining this to me user. i have a 2013 AM3+ board with a FX 6300 and i've never tried to update the bios. i'll try it and see if any performance improvement come.

What game publishers are still compiling their shit with the Intel compiler?

The "Ti"s are photoshopped tho, those are regular 1080s. Maybe it'll ease your frustration.

I'll wait when amd patches the hardware issue first

Yes. Lots of errata can be fixed with those updates.
Sadly, meltdown can't be fixed in that way because is a flaw in the design of the processor.

...

...

>is generous as fuck and gives performance for great prices
Your forgot that one in Chad.

Performance. To be fair there is a check in Intel processors before that data can be accessed normally, but that relies on the assumtion that the data can only be accessed normally. That assumtion flew out of the window once someone figured out you could pry bits of information from the cache with a side channel timing attack

Noo goy! Do not buy the rotten indian pop chips, they're just glued together poo that doesn't even work. I stronly recommend Intel