Why is Intel such a PCIElet?
Other urls found in this thread:
en.wikipedia.org
virusbulletin.com
scanproaudio.info
youtube.com
youtube-nocookie.com
phoronix.com
phoronix.com
healthitsecurity.com
zdnet.com
twitter.com
>what is Skylake-X
>what is Cascade Lake-X
>what is Xeon
>what is shitty AMD memory latency
>BTFO by Threadripper
>BTFO by Threadripper
>BTFO by Epyc
>moving the goalposts
cope harder
>>what is Skylake-X
dogshit
>>what is Cascade Lake-X
gold spray painted dogshit
>>what is Xeon
boomer dogshit
>>what is shitty AMD memory latency
FUD
t. retired my i7-8700K system for a 3900X one, AMA
amd fans will bag on Intel innovations like AVX512 and chipset-level wireless AX and thunderbolt until AMD copies it 4 years later
shit-brain gamers will think stuff like AVX512 is useless and think "bro just use the GPU" like the dumb asses they are
amd are the wagelet gamers that are trying to justify their anime-machine purchase
>/v/ buying cheaper computer parts with their NEET bux trying to roast Intel to justify their wagelet purchase
AMD could never
>Intel brainlets think 1/2 the cores at the same clockspeed for the same price is acceptable
Except intel now has 40+ performance crippling security mitigations and hyperthreading is SO dangerously compromised it's unironically better to just completely disable it at the hardware level.
For example spectre and meltdown version 1 LITERALLY slices SSD performance in half. Except there are now 10+ variants of EACH. To top it off new side channel attacks effectively render half of the security mitigations useless and bypass web browser ones with impunity.
LITERALLY anyone who values the BARE MINIMUM security possible (ie lowest and most stripped down humanly possible) has more money than brains if they go with intel.
You realize that all these security patches and mitigation are for data centers trying to pass security audits and not for the 99.9% of users that just browses the internet and plays games and maybe produces a video now or render a 3d environment now and then.
this is a cloud server issue, not a client user issue
also AMD is vulnerable to them as well.
en.wikipedia.org
name one real-world example of these exploits being used
EXCEPT things like RIDL combined with newer spectre/meltdown with hardened side channel attacks means hackers can LITERALLY pull passwords and banking info directly out of your RAM in mere seconds. They don't even need to be 1337 linux hackers, your average script kiddie can type out javascripts and load them on a malicous ad/redirect and intercept thousands to possibly hundreds of thousands of computers RAM content PER DAY.
Remember: such an attack would flawlessly bypass ALL spectre/meltdown browser mitigations.
What's your stack? What tools do you use professionally? Notice any funny shit happening?
how's it feel to finally be on the right side of history?
Except they're only vulnerable to 2-3 spectre meltdown variants and patches for those degrade SSD performance by like 5% at most.
That's a HUGE difference. See pic related
>retired my i7-8700K system for a 3900X one
Was that really worth it? Do you need the MT horsepower or what? I also went from a 4790K to a 3900X just 2 days ago and it was definitely a solid upgrade, though if I had a 8700K I think I might've still waited.
>>what is an AMDrone
>your average script kiddie can type out javascripts and load them on a malicous ad/redirect and intercept thousands to possibly hundreds of thousands of computers RAM content PER DAY.
you sound like a scared boomer lol
if i wrote a tool that scrapped random chunks of cached memory, i wouldn't even know what I'm looking at, or that it's even a password, or that my "malware" is running while you have your banking page open, sitting indle in another tab long enough for you to sweep through an entire 64-bit address space to grab an idle plaintext(???) password or session key.
>your average script kiddie can type out javascripts and load them on a malicous ad/redirect and intercept thousands to possibly hundreds of thousands of computers RAM content PER DAY.
there is literally no validated virus that takes advantage of any of these exploits that aren't just Proof-Of-Concept code.
virusbulletin.com
besides, icelake solves these vulnerabilities
Pretty good desu, I still remember how excited I was for my i7-8700K which I paid $400 for willingly because that's how much I wanted it. Only to find a 212 couldn't keep temps under control and then the security mitigations completely fucked my shit up.
It's so bizzare looking back, it's like I just survived the Titanic. Can't wait for zen 3 and the 14 tile thing, hopefully we get to cheap 24-32 core processors soon.
ALL FIXED IN LATEST CPUS
AYYMD HOUSEFIRES IS LOSING AGAIN AND AGAIN
I'm genuinely curious, why do you guys keep religiously defending intel. They're literally selling defective processors filled to the brim with speed hacks that you lose once you apply all the security mitigations.
On top of that newer ones just solve HALF of current spectre/meltdown CVEs. They do nothing to solve all the new RIDL, Fallout, and slew of brand spanking new sidechannel attacks. AMD gets like what, not even 10% of all that?
t. genuinely concerned intel user planning to move onto to zen 2 soon
IIRC the R20 multi is double that of my old i7-8700K. pic related
Very useful for 10-bit HEVC encoding. Can almost get 20 FPS encoding encoding 4K hevc which is bananas to me desu.
I wasn't defending intel at all, why are you so defensive? Genuinely considering ryzen, I had a fx4100 followed by fx8320. Both were fine for gaming but offered an inconsistent experience in actual work software. I couldn't give a shit about value and that seems to be ryzens number one shilling point.
>mentions Intel's current HEDT platforms
>tries to suggest that Intel win on memory latency
LOL
Shitlake-X has Ryzen 1000-tier memory latency.
Does the Intel HEDT have latency issues?
The latency of the 9900 makes it genuinely appealing even if compared to new Ryzen it is Threadlet
Yes, but you're dumb if you think that "latency" in this context has any impact on you as an end user. Memory latency is measured in NANOseconds for a start. One millisecond, the metric in which latency is measured for end users, is one million nanoseconds. You do the maths. The only relevance memory latency has is on overall system performance, and it's not something that most workloads even care about. Both Ryzen (especially Ryzen 3000) and Shitlake-X also have a ton of cache to offset their slightly higher memory latency.
Only a complete dumbass would factor it into their purchasing decision instead of looking at actual performance.
delid
It actually makes a massive difference for me and is probably the determining factor of real time audio production and recording. Many microphones and many takes getting ruined because of pops and crackles is a big deal.
All these 1200 dollar cpu vs 1199 dollar cpu threads... I'm to poor for these things :D
Any microcenter near you?
>t. audiofool pretend/hack sound engineer
Again, you're a dumbass if you think that's what memory latency is. Memory latency isn't passed on to the end user. It simply affects the CPU's overall speed. It isn't latency added on top of anything even at an OS kernel level, let alone a user level. Hell, even if it was, you're delusional if you think that a 20ns gap would matter in that context. You should actually read up on things that you don't understand very well before using them as a basis on where to spend your money.
Or don't. I don't care what you buy. Audiophiles are generally retards who buy junk anyway.
Preach.
>I couldn't give a shit about value
that makes no sense
I think he's trying to say he has more money than brains evident by his feeble attempt at human communication. I postulate he's an Apple user as well praising the recent cheese grater.
I'm so happy Intel is going the Nokia way.
there's always these fuckers that say they don't care about value, yet don't go out and get a dual-socket server-grade/workstation and custom loop
which kinda clues me in to them actually caring about value
>imagine actually thinking this
they have tons of cash on hands, or IP/assets to turn into cash
even AMD was on life-support, but wasn't allowed to die, simply because of a monopoly being bad for the market
scanproaudio.info
It is you who does not understand.
Latency is the defining bottleneck of sequential audio processing when using ASIO.
If the CPU cannot match the timings of the ASIO buffer then there will be drop outs and latency is a large cause of this, and is the reason that AMD was never used for audio until very recently.
Reducing the ASIO buffer will reduce real-time latency, but will require a more powerful CPU (and better latency). If the buffer is not effectively filled during the cycle, it will cause errors in recordings and playback.
If I am recording while hosting the audio through plugins in real-time CPU latency will have a huge impact on how well it performs within the ASIO buffer.
When 20ns gaps manifest as drop outs, then yes they are perceived by the user and yes its a big deal.
also
>youtube.com
>CPU Performance vs. Real-Time Performance in Digital Audio Workstations (DAW)
2700 is shit
>Memory latency isn't passed on to the end user
It totally is. Memory latency is per operation. Even basic bitch programs like web browers make tens of thousands of memory operations, or rather I should say the OS does for them. That's why the Apple chips score so well despite being smaller than other ARM processors, extremely high performance memory and cache with very low latency, and superior memory parallelism to even Intel's Core processors
Imagine having gorillions of shekels and your 10nm is still fucking garbage.
While TSMC and Samsung 5nm is already in risk production.
where is the 2990WX in that graph? its cheaper than the 9980XE
afaik the 2990WX scores 13.5~14k multi and 410~430 single core
thats one hell of a deal, too bad Microcenters are rare as fuck even in the US
Depends what GPU you use and what RAM. A fully maxed out 2700 locked at 4.1-4.2 GHz with 3600 CL16 RAM with tighened timings is actually 90% as fast as an i7-8700K overclocked to 5GHz on all cores using the same high end tightened RAM with both systems using an RTX 2080ti. But if you say dropped down to an Rx 5700XT the 2700 suddenly becomes 95% as fast (90% as fast if using 3000MHz CL15 RAM.
Any GPU less powerful than that and it practically doesn't matter what GPU you use at 1080p and even less so at 1440p as the GPU becomes the buggest bottleneck. And of course all gaymen aside the 2700 CRUSHES the i7-8700K at multi-core.
Remember the i7-8700K + high end AIO is ~$500 WITHOUT including the motherboard.
Intel obviously bribed them to exclude it just like how they keep bribing reviewers to not include every single security mitigation to date in benchmarks. They can't save face when it comes to multi-core performance anymore.
>doesn't matter what 12 threaded CPU you use*
Another bullshit by AMD
You cant even use to combine few weak NAVI cards into something usefull
Why?
Because Crossfire was a jnoke on customers
But they do have RIS which means you can get 1080P FPS at 1440p or 1440p FPS at 4K with visually very similar if not better quality. So there's that, which is nice
NVIDIA has sharpening too, you know. And no, a sharpened upscale is not as good as a native render, not to mention that you can sharpen the native render all the same.
Also what's wrong with PCI express 4.0 SSDs?
Intel still have better PCI-E responsiveness for videocard which more important
youtube-nocookie.com
Right but nvidia's imolementation if it looks like someone took a shit on your screen and then proceeded to smear it. And again RIS turned out so well (especially with improved 19.9.x drivers) that it will in some cases look BETTER than the native rendered res
Doesn't matter, security mitigations effectively nullify that ESPECIALLY with RIDL/Fallout V2.
NVIDIA has sharpening which looks identical to AMD, probably because they implemented literally the same algorithm which is open source.
There is any REAL technology from AMD which is not another buzzword?
see
There's a reason why audioidiots are called "phools" and there's a reason that 2-3 pieces of s/w crap themselves on amd, incompetent development.
A programmer never blames the h/w, unless there's a bug in the h/w.
If a system works X way and you program it Y way, obviously it won't run.
Do you know why intel gets the props here? Because audioidiot companies have been selling to itards since ages, itards buy crap s/w in general and companies spent dev money only on that platform.
I have a friend who is 70 years old, he has produced many "cd"s, he wrote music for some olympic fest in athens2004 and he does all his work on a 2002 athlon with 512mb ram and windows XP.
Apparently nobody shat on intel for making "high precision timers" with microsoft and they were not precise at all but all drivers and s/w relied on them and everybody made workarounds until intel-ms said to stop using them.
based
>Still talking about Skylake-X when Cascade Lake-X has fixed all of these
TOP KEK, KILL YOURSELF FAGGOT
Except there's a catch: the performance hit is higher and the scalers are nowhere near the same quality essentially making it a meme on RTX cards ATM. I believe the lanczos scaler has like a 10-20% performance hit when done with floating point precision IIRC.
It's still a step in the right direction though, Nvidia will hopefully join hands with AMD and develop an actually good neural network like scaler that won't be an absolute disaster like DSLL was.
see
tl;dr intel's entire i3/i5/i7/i9 are forever borked as newer spectre/meltdown/RIDL/fallout/zombieload
Data centers are literally chewing off their own foot to stay alive these days. Half are so despondent they've begun migrating to threadripper systems with tears streaming down their face.
you didn't address the fact that none of this effects consumers
Except it does, all these mitigations have an impact on core perf/latency and I/O throughput/latency. Because of the branch prediction speed hacks intel's entire core series consists of, every mitigations based on that can have anywhere from 1-10% performance degradation. When you stack said mitigations it can well surpass that (I don't notice anything) 10% threashold.
What makes all this horrifying is the new side channel attacks rendering half of all current mitigations useless (especially the browser jitter thing). But you can't remove them because then that would open a door to the current exploit vectors of old security mitigations for exploiting again.
It's all very tiresome.jpg
again
it doesn't effect consumer usage, as in i dont even need to install the mitigations
you sound like a scared boomer jesus christ
I don't doubt that at all. But it depends on what you're doing and regardless of the reasoning behind it, as I described is the impact on the user.
>he does all his work on a 2002 athlon with 512mb ram
I'm not that surprised but he couldn't do what I do with those specs and vice versa probably as well. I imagine he uses a hardware mixer or just incredibly small libraries, I mean there is no way you could do contemporary Kontakt work with that little RAM and there's no way you could really get into creative design with such a weak CPU.
There is a huge variety of workflows and tasks that different people may or may not do, not everyone is recording a few guitars and not everyone is using 1000 channels of orchestral scoring.
An old athlon may well be suitable for one but certainly not for the latter.
What does Intel have to do with audiofools?
What do you want people to do, code their own programs or just simply stop using bad ones?
Its all good and well for you to tell me that ASIO is not good enough but it doesn't change the reality that its the one that I have to use and that interacts with (for whatever reason) the developers like you describe.
ALL are fixed, you're retarded period
Everything is fixed and there's nothing you can do about it
phoronix.com
>The time to perform a context switch on EPYC Rome processors is lower and comparable to Cascadelake, which itself is a huge improvement over Skylake given the hardware mitigations in place that had really drove up its context switching time.
KILL YOURSELF FAGGOT
trips of truth
intel is a fully kiked company.
I just told you, IT DOES. Not each one individually but when you stack them on (like what windows update will do by default) the cummulative performance degradation will be felt in one way or another.
The newer core series processor only have HW fixes for old spectre/meltdownd 1st versions of RIDL/zombieload/fallout. What they don't tell you is the HW spectre/metdown mitigations in place have been successfully bypassed by the new sidechannel attacks so you still get hit with half of the performance degredation of what kabylake/coffeelake sustains. It's NOT AS BAD but still bad.
gas yourself. i had a 8700k running on a 212 black just fucking fine. its not my fault you were to retarded to disable enhance multi-core in your bios or to retarded to make sure tdp turbo limit wasn't set to unlimited on auto settings because motherboard vendors are assholes.
>pic related
my 8700k with a 212 evo black.
Enjoy your MASSIVE frame stuttering. Hope that's a high end Z motherboard too.
doesn't affect gaming performance or 90% of the workloads people do.
phoronix.com
linux jew phoronix did testing against zen 2. yeah intel isn't great, but amd and intel in stuff that are affected by it are generally in the same ballpark. worst case for intel is amd level of performance. thats not great, but its not garbage. people act like it makes intel core 2 duo level of performance. pretty retarded.
frame stuttering? are you fucking retarded? and high end? what for you fucking idiot? i hope you say that about the 3600x you fucking kike. unlike you i also owned a 1800x and a 2700x. pic related my 1800x.
other pic related my 2700x.
>spectre
Okay now show me the performance hit of all 10+ spectre AND meltdown variants combined. See Also just for teh lulz you don't even have to show all the performance degradation from RIDL/fallout/zombieload/ and brand new side channel attacks stacked on top of that.
and now my 9900k. its how i know you are full of shit. because you are a garbage piece of shit. i hope you choke on lisa's cock you fanboy piece of shit. all fan boys need to be gassed.
you really need to read the phoronix article because his tests included meltdown patches enabled. and the stuff that affects spectre generally affect meltdown. and his tests where also down with fallout and zombie patched and enabled. you really need to kill yourself.
>AMD Zen 2 processors feature hardware-based mitigations for Spectre V2 and Spectre V4 SSBD while remaining immune to the likes of Meltdown and Zombieload. Here are some benchmarks looking at toggling the CPU speculative execution mitigations across various Intel and AMD processors.
i7-8700K doesn't have a soldered IHS, this results on one of the cores getting hotter than the other ones. This translates yo wildy fluctuating FPS that can dip up to 50% of the average FPS.
You DON'T have an AIO which guarantees this will happen even at stock. You also removed the TDP limit like a dumb fuck, correct?
RIGHT THERE IN THE FUCKING FIRST PARAGRAPH OF THE ARTICLE. FUCKING HELL CAN YOU FANBOYS PLEASE KILL YOURSELVES.
THERE ARE 10+ WORKING SPECTRE AND MELTDOWN VULNERABILITIES YOU DUMB FUCK
wow you really are full of shit. please kill yourself. just because you play on a technically doesn't mean its true. please, take out a gun, and shoot yourself. or go play in traffic. just do something to end your suffering your are causing on this earth you insufferable piece of shit.
>AIO You DON'T have an AIO which guarantees this will happen even at stock.
>You also removed the TDP limit like a dumb fuck, correct?
you do know op in his rant did call you an idiot for probably keeping tdp on unlimited. he says it right here:
>retarded to make sure tdp turbo limit wasn't set to unlimited on auto settings
so then you accuse him of it? op is right, you are a retard
do humanity a favor and engage in reading compression by reading the article. actually ill start pasting it here.
AMD Zen 2 processors feature hardware-based mitigations for Spectre V2 and Spectre V4 SSBD while remaining immune to the likes of Meltdown and Zombieload. Here are some benchmarks looking at toggling the CPU speculative execution mitigations across various Intel and AMD processors.
For this round of testing are some mitigation comparison tests on the Core i7 8700K, Core i9 9900K, Core i9 7960X, Ryzen 7 2700X, Ryzen 9 2950X, Ryzen 9 2990WX, Ryzen 7 3700X, and Ryzen 9 3900X. On each processor, the tests were done when booting the Linux 5.2 kernel with the default/out-of-the-box mitigations for Spectre/Meltdown/Foreshadow/Zombieload (all CPU speculative execution mitigations to date) and then again when making use of the "mitigations=off" kernel parameter for disabling these run-time-toggleable mitigations. Basically the tests are the equivalent of mitigations=off vs. mitigations=auto (default) comparison.
All systems were tested with a Crucial MX500 SATA 3.0 SSD with Ubuntu 18.04 on the Linux 5.2 Git kenrel at the time and all systems had memory at their optimal rated frequencies and satisfying the maximum number of supported memory channels. These tests aren't being done for comparing the raw performance between the systems but looking at the relative mitigation costs in different workloads impacted by these different mitigations. See our benchmarks from last week (and more on the way) if interested in the raw Intel/AMD Linux CPU performance.
This pozzes the i7-8700K. No amount of seething copening will change that. I'm done with this thread.
t. computer that has to deal with BS like this as a job
The Zen 2 results come in as a bit of a surprise. With Zen 2 featuring hardware-based Spectre and Speculative Store Bypass mitigations, one might reasonably think that the relative performance impact is less than with the original Zen/Zen+ processors, but at least in our tests under Linux that wasn't exactly the case. There is the possibility that the current software mitigations are being over-applied to Zen 2 CPUs, but that is the current experience. I reached out to AMD last week with my initial findings but I have yet to hear back what they recommend as far as the software mitigations go if the existing defaults are what they expect or if some kernel updates will be needed either for checking some MSRs or CPU models for relaxing some of the kernel mitigations when running on these new AMD zen 2 CPUs.
The default Intel mitigations namely come down to "l1tf: Mitigation of PTE Inversion + mds: Mitigation of Clear buffers; SMT vulnerable + meltdown: Mitigation of PTI + spec_store_bypass: Mitigation of SSB disabled via prctl and seccomp + spectre_v1: Mitigation of __user pointer sanitization + spectre_v2: Mitigation of Full generic retpoline IBPB: conditional IBRS_FW STIBP: conditional RSB filling." In the case of AMD Zen/Zen+ CPUs, the default software mitigations are "l1tf: Not affected + mds: Not affected + meltdown: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl and seccomp + spectre_v1: Mitigation of __user pointer sanitization + spectre_v2: Mitigation of Full AMD retpoline IBPB: conditional STIBP: disabled RSB filling." And then in the case of Zen 2 with Linux 5.2, "l1tf: Not affected + mds: Not affected + meltdown: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl and seccomp + spectre_v1: Mitigation of __user pointer sanitization + spectre_v2: Mitigation of Full AMD retpoline IBPB: conditional STIBP: always-on RSB filling."
its ok. you will always be a retard who's full of shit. remember to eat your tenndies while choking to death on lisa's cock you fanboy.
>to date
>july 15
There weren't 10+ discovered spectre and meltdown mitigations then user.
Between Zen+ and Zen 2 on Linux right now, all of the software mitigations are indicative of still active and beyond that for RSB filling goes from disabled to always-on with Zen 2. That RSB filling is even more strict with "always-on" than Intel CPUs currently relying upon "conditional" filling. RSB (Return Stack Buffer) filling is part of the Retpoline mitigations for ensuring malicious user-space code isn't exectued speculatively when RSB under-fills happen.
Here's a look at how toggling the Intel/AMD CPU mitigations compare using the Linux 5.2 kernel on these different processors. Tests via the Phoronix Test Suite.
With the Sockperf test stressing the throughput performance of the kernel's socket interfaces, the mitigations caused a 20% hit to the Core i9 9900K while the Ryzen 7 3700X saw a 13% hit and the Ryzen 9 3900X a 10% hit in the default state. The Core i9 9900K still delivered better performance for this synthetic benchmark than the AMD CPUs even with Intel's greater mitigations. Interestingly though and what I alluded to in the introduction, the Ryzen 7 2700X saw a 5.5% hit and the Ryzen 9 2990WX a 6.2% hit.
In workloads like compiling the LLVM stack, there was a small but measurable difference on the tested systems.
With the Hackbench Linux kernel scheduler benchmark, the Intel CPUs have a much greater impact (sans the 7960X) and the unmitigated Core i9 9900K with no mitigations was as fast as the Ryzen 7 3700X only to become much slower with mitigations.
it
post the news and the benchmarks showing their impact then. unlike you whos just shoving fud down people's throats expecting people to take your word for it, at least i'm trying to show benchmarks and results.
healthitsecurity.com
>August 12, 2019
>Bitdefender worked with Intel and Microsoft to protect systems from attack through a working patch offered in July.
>BitDefender said only Intel chips are affected, while Red Hat and Microsoft say other makes of chips are vulnerable. Microsoft says the new vulnerability is a variant of the earlier Spectre flaw, which affected Arm, AMD and Intel chips, while Red Hat says the SWAPGSAttack affects Intel and AMD chips. SWAPGSAttack also doesn't affect systems running Linux-based operating systems.
zdnet.com
>SWAPGSAttack affects Intel and AMD chips
> SWAPGSAttack also doesn't affect systems running Linux-based operating systems.
intel truly brfo and bankrupt. toss your intel cpus into the trash now. the end is near. this is truly the year of the amd
