*Official* NBR Desktop Overclocker's Lounge [laptop owners welcome, too]

Read it earlier today. Lots of good info as always (I read your 10th gen stuff too).

 

11th gen is a step in the right direction and at least gives us a glimpse of what's to come. I don't think Intel had any aspirations of dethroning AMD. They just wanted to slow down the blood letting while they finish working on Alder Lake.

Also for those who are interested, rBAR support is finally here! (!):

 

If they did they have to offer better than this... Not every day you see the top dog get less cores than its predecessor. Hope it will be long time to next time I see something like this.

As suspected at the performance two weeks ago, Rocket Lake-S ends up offering too little, too late. Even in 14 nm production that has been optimized over the years, eight cores are not enough, despite modern architecture, to be able to keep up with the competitor AMD at the beginning of 2021. It is not without reason that Cannon Lake-S and Ice Lake-S were originally intended to compete for customers with the same eight cores in 10 nm as early as 2019. Two years later, the compromise born of necessity fails to do this. The turbo crowbar with power consumption beyond 300 watts as an afterburner does not help either - a new production process and a new approach for more cores are needed.
https://www.computerbase.de/2021-03/intel-core-i9-11900k-i5-11600k-test/6/

And this will neither help... 4th is still better than last

 

If this was Intel's intention, they absolutely botched it. They would have been better served riding 10th-gen for another 6-10 months instead of wasting resources on this fart cloud.

 

"fart cloud"

You win today sir.

 

I found kit guru actually approached it somewhat accurately showing ABT (aka PBO) on and then off along with the huge increase in power requirements and scores. This is the most realistic scenario for the 11900k and is similar to Ryzen's approach with PBO.

One of his final assessments is that with Intel adopting AMDs boosting approach, 24/7 daily manual overclocking is basically dead. You will always have those who will still push beyond that to extract every bit of performance out of their CPU and memory, but a large chunk of that free overclock headroom that could vary from chip to chip depending on silicon quality is a thing of the past:



Then you have Luumi who tweaks it manually to achieve much of the same thing and basically ends up at 5.2 ish on all cores and comes to the same conclusion many have and that is the overclocking capabilities of the 11th gen is lower than 10th gen on average.



Then there's the forum over at overclock.net for 11th gen results which is getting pretty interesting, but as always you have many users pushing it beyond what 99% of users will ever do with this or a Ryzen chip in their system. The most realistic scenario is ABT vs PBO as kit shows above.

My tentative takeaway is if you have the cooling and power and time, you can dial in an 11th gen to make it competitive with the 5000 series core for core depending on task and sometimes it will even win simply because even though it doesn't overclock as well as 10th gen, it overclocks well enough coupled with the IPC gains to at least stand its ground. Larger die sizes are making it easier to cool, it pulls insane amounts of power in many stressful conditions depending on work load and gaming performance is game by game. Some games it wins, most it loses. Of course if you need more than 8 cores, it's useless either way. Intel needs to drop the 11900k to $450 and the 11700k to $350. It is overpriced as at best it holds its own against the 5800x sometimes and just gets decimated by the 5900x.

 

Steve is really railing on 11900K... Quite a contrast to Brother Luumi's take on the only downsides being a reduction in core count and high price tag. Not sure why Steve is seeming so biased lately.

 

Didn't know you worked as a reviewer and bug-hunter.

 

Luumi is approaching it as a hardcore overclocker. Steve is approaching it like the vast majority of consumers/buyers who aren't even going to knock on the hood let alone lift it up even a fraction of an inch. If you plan on immediately getting into the BIOS and tweaking it, Luumi makes more sense from your perspective. If you're like a large segment of YT viewers and buyers, Steve makes more sense. I tend to lean towards Luumi's POV, so I see merit in the 11900k and even 11700k (where you can tweak it just as much and ABT becomes a non issue). If you're manually optimizing your CPU, while the 11700k in theory is binned lower, you still stand that chance of getting a better binned 11700k than some 11900k's as we've already seen.

The problem is this weird stopgap transition really feels like the Pentium III to Pentium IV transition. A new architecture that required raw frequency and tweaking to finally break free of its predecessor. In the interim, AMD introduced the Athlon and won the performance crown overall and took the lead at lower clock speeds while Intel worked out the kinks, discarded what didn't work (shortened up that EPL), kept what worked, fine tuned everything and brought us the Core architecture which buried AMD for good....till now.

This feels just like that period in time but with a different set of conditions. Intel got caught milking an aging architecture while resting on its laurels and soaking up the profits while assuming their projected time tables would give us 10nm a few years ago and they would be well on their way to 7nm by this time. AMD slowly makes in roads then here comes the 5000 series completely outclassing 10th gen and now Intel is playing catch up again. If Alder Lake is truly coming this fall, this makes the RKL release just seem even more out of step. If Alder Lake winds up being delayed to 2022 I wouldn't be surprised.

If Intel had somehow managed to squeeze 10 cores on their i9, many of these 10th vs 11th arguments wouldn't exist. Core for core,11th gen is a better architecture than 10th gen (many of the reasons detailed by @Falkentyne 's excellent post). The problem is it is meant for 10nm so 8 is all we get without a new package design and thermals exceeding 350w in some loads.

RKL is at best a technological preview of things to come and is able to compete somewhat with the 5800x depending on use case. Clock for clock, IPC uplift over 10th gen is healthy but not up to the 5000 series. The 11600k looks very compelling both price and performance wise (we often forget the lower core segment) along with the 11400. If I was building my brother's system right now and within my return window, I would have switched him over to the 11600k. I think in the right situation, the 11900k is a decent chip. If all you need is an 8 core / 16 thread chip and you have the cooling and power in place to handle ABT or your own manual adjustments, it's definitely decent. Depending on what games/applications you run, it might even be better than the 5800x. But if you need more than 8 cores, it is a total dead end; and I still stand by the pricing issue. The 11700k and 11900k need to be priced at $350 and $450 respectively.

 

This problem already makes AMD unacceptable to me and if Intel is suddenly becoming brain-dead by adopting the AMD gamer-boy approach to boosting, here ends my interest in computers. I guess it was nice while it lasted. Time to find something that I think is awesome to replace that which used to be. I have no interest in buying fancy hardware that sucks at overclocking. If I wanted to be a consumer that runs everything stock then I would just lower the bar to the floor and reach for a smartphone.

Indeed. Since overclocking is about 98% all I care that much about I would be approaching it from a similar perspective as Luumi. I know I am severely outnumbered by people that don't care or don't even know what overclocking is. Unfortunately, that's going to really eff things up for those that do know and care.

 

If Jayz2cent's speculation is right, you'll be heading back to the X platform. I agree with his theory. Intel has conceded the HEDT market to AMD long enough. I think they will make another push there.

I think we forgot that the Z platform, for all intents and purposes, is a mainstream platform and the X platform is where we should be in as oc'ers.

 

Excellent video on new overclocking features.

 

Very possible. I wouldn't mind that necessarily, but the cost might rule out my participation if the pricing of HEDT enthusiast motherboards and CPUs remains as retarded as they have been.

Speaking of which... finally retired my 3-year old phone. It still looks and works like new. Still a really good phone, but I finally got fed up with it not being rooted, and being such a huge pain in butt to root using severe brute force (not to mention easily bricked in the process).

Rooting a OnePlus is easier than any other I have done before. It took only a few minutes and it was rooted before I made the first phone call. Literally, from the time I opened the box until the rooting was done, including reading the instructions and downloading the files to do it, was about 30 minutes... like falling off a log easy.

Old phone: https://browser.geekbench.com/v5/cpu/7088160

Spoiler: 1440 Multi-Core

New phone: https://browser.geekbench.com/v5/cpu/7160833

Spoiler: 3123 Multi-Core

 

How did your 11900k pan out? Versus your 11700k? I saw over on the other forums that SP55 11900k that was truly cringe worthy even with @Falkentyne 's adjusted down SP expectations for 11th gen.

 

I hit an SP80 from Microcenter. From Falkentyne's post, I guess this is a pretty good sample. It's definitely a better chip. I haven't even really tried to push it too much yet. I messed around a bit yesterday but ran short of time. If I was that user and got an SP55, that chip would be going back, in no way would I accept that after paying the i9 insane tax.

 

SP80 is a damn good rating, congrats! I swear when it really matters, you got the CPU magic touch.

 

I think a big problem is that Intel has consistently demonstrated a complete and total lack of leadership in the desktop CPU market. Where is their vision? Intel is just playing catchup to AMD. AMD, now that they appear to be on top, are going to realize soon that failure to innovate will cost the company unless enough sleep-walking zombies continue to blindly prop up companies by purchasing their products which are only repackaged/rebranded/remarketed junk.

 

And, that doesn't overclock worth a damn... CPU and GPU. This has been a driving force in marketing even though the participant pool has always been small. You can't make any claims to fame if your stuff sucks at overclocking. It's just a bunch of belly-buttons and butt-holes consumer electronics fodder when it doesn't.

 

Please forgive my ignorance, but how do people derive these ratings? I know you can check a GPU's ASIC quality from GPU-Z, but how do you check the silicon quality of a CPU?

 

As far as I know it is a feature that is only resident on ASUS enthusiast grade motherboards, and only the enthusiast boards. It is not included on the gamer-boy budget junk like the ASUS Prime series.

I wish there was some kind of third-party utility available to check this. Apparently, someone at ASUS had a secret meeting with Intel to figure out how to parse that information from the CPU micro-code and display it in the BIOS.

 

I think Intel try put more eggs into notebooks. They lost Apple and try to please Microsoft. And they need to catch up against AMD for notebooks. Each new month and more and more OEMs will push for more Amd powered notebooks. See etc here and here. They Intel will go where the biggest profit is. Remember desktops chips (mainstreeam) will be from notebook chips. And not the opposite. big.Little here we come


Unreality bites: Intel Core i9-11900K looks practically obsolete in Unreal Engine comparison with Ryzen 7 5800X and i9-10900K

--------------------------------------------------------------

And Happy Easter holidays boys and girls

 

Sad times, indeed...

 

I think he is very much pissed off at Intel, maybe they really need a kick in the nuts but sadly Intel has felt the heat only from Apple booting them off, not the Fabrication problems nor the Xeon efficiency nor the HEDT nor the Mainstream Z, that's what I personally think of Intel now. Looking at the Adaptive Boost how Intel gated it behind the i9 K and KF only, they only seek maximum half measure while as we can see from Papu's post on how they removed the memory OC limits on the lower tier mobo with this new Rocket Lake launch.

Also it's so big of a corporation how this product ends up being made and the marketing materials etc, plus their x86 server marketshare complete dominance and how their deep pockets pay off to these C suite people in mega corps for the Infrastructure, however since Cloud is gaining more traction it's now perhaps hard for Intel to push that backyard deals to blow the competition off like they did to AMD long time back. And they should really worry about EPYC Milan processors. SP3 socket is total massacre for Intel, they do not have anything now, utter shame. Even big sadness is people rely on Smartphones more than at-least having a look at the HEDT TR4 socket lineup on what the computers really can do.

Also I didn't like GN's RKL approach I feel that's even worse on Steve's part, because CML had tons of benches with so many processors now it's reduced pathetically to simple bs stock setting and a handful of processors it's really boring and barebones crap, honestly speaking whoever buys a Z board and puts a K or KF processor would they run them stock PL2, okay forget PL2 but Stock TAU ?? I don't get what Steve wants to do with that type of approach. It was really fun to see how K processors esp the 7700K, 8700K and 9900K overclocked gave a good hefty boost. With CML it got reduced since we all know 14nm++ couldn't be milked more or the Skylake for that matter or AMD's approach playbook by clocking them max out of factory. GN even put high power consumption charts with those K series processors until CML, I remember they even put 9980XE X299 OCed on all cores to 5.1GHz which matched 9900KS in the gaming and their whole suite, ofc the added high power consumption charts too. Also notice they don't even put the AMD full PBO2 unleashed Ryzen parts scores, only stupid stock matters !!

A shame really, K series processor with locked Tau is really really stupid esp on a Desktop Socket / Mobo and chassis, beats me, it's even more stupid than running an M1 Mac and expecting miracles out of it, yea Intel Rocket Lake unlocked to peak run super hot but why would any K buyer run it on Stock TAU of all ?? Maybe clockspeed limits due to temps or their ITX small pet food box cases (sorry I do not really like that form factor, it's like trying to fit a Mustang V8 into a small Prius chassis) or maybe people do not want to OC.

I think they gained over 1M subs, so they want to appease normies idk. It's pathetic, even AMD also people who buy upper SKUs of 8C and 12C with SMT probably many of them definitely run all unlocked and PBO, since the damn processor costs a lot with Zen 3 Vermeer and the investment they made is for that performance not sandbag it buy running on low clocks, when you literally have **Free** performance uplift, esp nowadays everyone runs an AIO, which are cooler in temps than a Tower Coolers (which are my fav since longevity) or the niche open loops which are only enthusiast class AND good quality motherboards with fairly good VRM circuity which can handle the higher power consumption.

 

@Mr. Fox

What are the best timings you got on these Corsair 4000 sticks? I want to compare to what I am going to try to get on Ryzen.

 

I never went lower than 4000 on the memory clock because I never saw any point in bothering.

As far as timings go, 4000 CL15 and 4200 CL16 totally stable. (I could do some benching with 4000 CL14.)

 

Well, I was going to grab an example of the 4000 CL14, but it appears the damned dunderheaded dolts (yes, DDD, but not the Guy Fieri kind) at HWBOT have removed certain menu items under My Submissions and the memory read test is no longer a thing. They're gone now. It must be too hard on the children to have benchmarks that require any skill or effort. Worthless bums.

I will rifle through my Imgur pages and see if I can find it for you, bro.

 

From watching the GN video earlier today, I definitely could see Steve's frustration and disappointment with the product. My take away is that the 11900K is over-priced Meh. At least at stock settings. His video confirmed that two fewer cores will impact multi-core/thread performance. Overall, I felt this specific video wasn't complete. There's more to the story. In GN's 11700K video, he pointed out that different motherboard OEMs take liberty out of the box with setting power limits. Not that important to me since I would overclock the hell out of it anyway. But, he wanted to make Apple-Apple comparisons of performance across different mother boards.

Bearded Hardware produced a video detailing his overclocking of the 11900K. I haven't watched it, yet. But, BH does typically do a good job showing how far he can overclock.

Again, my take away from GN and even Jayz2cents is that the 11900K fails as a flagship part. Especially, if reviewers had hoped that the chip would beat or at least come close to the 5900X at that price point. The LGA1200 socket is a poor purchasing decision at this point given that it will be replaced by the 1700 later this year. There is no incentive to upgrade from the 10850K/10900K.

I have to hope that companies like Intel don't view customers as stupid idiots. That is so 2021.

 

Wait, wait, wait, huh, what? You mean we're not?

 

Haha thanks! So just tested 5.2Ghz all cores (no offsets) 43x cache and 3733Mhz 1:1 CL15 in a CB23 run. This required 1.42v LLC5 with 1.296vmin. Temps on my 3 year old peaked at 80c on a single core. Rest were low 70s. I honestly think my cold plate is clogged or something in that area since my 11700K hit the same higher temp on the same core. I'm replacing the AIO since it's 3 years old and it is my opinion that these things start to degrade/clog after that period soon followed by a pump failure. I'll repurpose the AIO as a backup for my FIL's build I made for him a year ago.

53x CB23 crashed almost immediately at 1.45v, but I haven't tried messing with any of the extreme options and LLC5 is likely not gonna cut it for higher clocks. So loss compared to my 10900K that could do the same run at 53x all cores at 1.39v. A loss of 100Mhz but again were comparing different archs. I also could easily need to up my VCCIO since I have it set pretty low around 1.050v. In fact I think I was way overvolting my 11700K VCCIO and it may have been affecting stability. On RKL there are two separate VCCIO rails. One for the memory overclocking and one specifically for everything else.

I applied 1.27v out of habit carried over from Skylake history trying to get my memory to go 3733Mhz 1:1. After watching SkatterBencher video I realized my error and tuned that down. My 11900K can do 3733Mhz CL15 1:1 tuned no issues at all with 1.25v SA/VCCIO MEM and 1.050 VCCIO. RKL seems to really love to be memory tuned since out of box latency is a regression from SKL.

 

I tried to ask you a question about your memory overclock on Z490 over on OCN. But I don't know if you saw it.

 

https://www.pcworld.com/article/361...e-cpu-secrets-in-a-reddit-ama.html#tk.rss_all

Interesting article. Note this:

"So much of the criticism of Rocket Lake seems centered on the regression in core count. Some have wondered why Intel bothered to “waste” valuable die space on the graphics core. If it had simply dedicated that to two more cores, wouldn’t that have been more prudent considering it’s actually priced against AMD’s 12-core Ryzen 9 5900X?

“A 10-core die without integrated graphics would have been a second die (design effort) as integrated graphics is a hard requirement for commercial customers,” Rouse responded. “Not developing the design for a second die was a business decision.”

By business decision, Rouse speaks to the business of trying to make one thing or very slight variations of it to sell to many people. While the vast majority of gamers and enthusiasts all run discrete graphics, Intel uses the same die for both gamers who rarely need that IGP, to boring business customers who almost always use integrated graphics."

Also:

"Intel isn’t the only company or business that sells different products based off one design or parts bin. Everything from cars to shoes is made with some reuse of the same building blocks. McGavock acknowledged that Intel is partially handcuffed by its current monolithic design, which can force a high-performance gamer to take the same product features that may appeal more for consumer PCs.

That, however, will change, McGavock said.

“We are working on solutions, which include more modular designs, but for now we have the monolithic, large die that has to fit a variety of usages,” he said."

My thought: All about money. Intel is feeling the market pressure and caving to shareholders.

 

Intel's 11th gen desktop chips is on level with this joke. Enjoy


https://www.pcmag.com/news/pc-builder-creates-an-nvidia-rtx-4090-for-april-fools-day-and-its-massive

@Mr. Fox Want one for mining? Yo'll need a open bench for it, HaHa


-----------------------------------------------------------

To compete in the future, Intel will most likly change the CPU Naming Scheme. 10nm doesn't look great vs. 7-6-5 or even 4nm process. Releasing several new gen chips foorwards on same process node doesn't show Intel in a good light. Neigther will the backport as Rocket. They seems to finall have learned from + 5 years with 14nm+++++++++++++++++

Forget 10nm? Intel May Change CPU Naming Scheme tomshardware.com

 

I have two open benches, but I think I will wait for the 4090 Ti XOC Nitrous edition.
It should be a better value at only $4,999.00 USD.

 

Decided to go the frugal route to play with some 11th gen stuff:



With BB combo specials, elite offers and card bonuses this month, I was able to get both for well under $600 total delivered.

Hopefully be able to get 5ghz all core, but testing it against my 5800x will be 4.7-4.8 all core like my 5800x runs.

 

I can't tell from the image... what mobo?

 

I went with the MSI Z590 Pro as soon as it was back in stock @ $189.99. Want to see how much bang:buck I can extract from a sub $600 pairing. Best buy also is running $20 off combo's along with 10% back and a $25 e-Gift coupon for a single purchase over $250.00 bringing total cost down to $570ish for both shipped.


At one point I had paired the 11900k and an Asus M13 and just like that I was over $1k so common sense had to kick in real quick. Once you get sub $300 on Asus, a lot of the good goodies go away so then I decided to go ultra frugal.

Hopefully it is a decent chip, but I won't have the luxury of Asus die sense and SP/VF curve to lay some ground work. I'll end up doing it the old school way in my spare case (where my 5800x started its life before replacing my 9900k) w/ the 360 AIO. At least the 11700k has on board video.

I have no complaints about my 5800x at all except the USB issue rears its ugly head whenever I do long, sustained Xfers even running the newest BIOS. It does drive me batty as I have to sit there and watch it and can't walk away like I normally do during transfers. I don't have the confidence to use external media actively either at the moment for fear of the port just "dying" and the transfers just stop. No errors, warnings or nothing. It simply stops working.

 

Yeah, it's too bad that ASUS mobos are so unreliable and their warranty service is so horrible. They have some great features that would be really nice to have in a different brand.

I am looking forward to seeing your results.

 

Welp, I learned something new today. Thanks for this nugget of information. I've been wondering why in the world my external hard drive enclosure likes to drop off when connected to my desktop (which also has an ASUS motherboard), but always works fine with my Ranger. I thought something was wrong with the cable.

You have received cookies from me as a result.

Also another interesting tidbit of information: my external drive never drops when connected to my desktop through a USB 2.0 cable. You may want to try that and see if that stops the port from dying.

 

USB disconnects are not an Asus problem but an AMD one. Their latest AGESA was supposed to have cured it, but if the issue is still present, it might be a good idea to hop onto the AMD subreddit to see if there are still active discussions as company reps seem to be present there.

 

An interesting piece on Rocket Lake. Maybe Intel realized that the 11900K would be a dog. But, that sometimes you need to take baby steps to get where you want to be. Intel 11th gen - Intel is willing to bite the bullet on the title (real or hyped) of being the best CPU. But, could Intel leverage lessons learned to make a comeback with 12th gen?

 

Depends on how yoo look at it. Could still be a problem due Asus and not so much with the other brands http://forum.notebookreview.com/thr...-frequent-random-freezes.773174/#post-9961012

Here is the problems Red camp struggle with. From my post here

AMD confirms USB problems with 500 series motherboards computerbase.de
The increasing number of users who complain of problems with USB ports on motherboards with B550 and X570 chipsets calls AMD onto the floor.

 

If there was ever a video that needed a time index....this is it.

$530 Motherboard....too rich for my blood.

I like his summation:

100-200mhz less than 10th gen OC overall
Railing on YT reviews that double down on sticking to PLs
Pulling 300w+ and hitting highs of 97c on custom water...(once the confidence level and technique increases, incoming delids all around)
287w non-AVX, 325w AVX
Custom tweaking > ABT

 

I actually ended up rerouting a long cable from the rear USB 2.0 ports for a sense of stability, but those transfer speeds are atrocious. The front Type-C port also eventually just stalls too.

Hopefully MSI issues a new bios update sooner than later.

Still there, but the latest BIOS for the MSI x570 Tomahawk (1.6.1 March 2021) is still 1.2.0.1 based.

 

Great video. The best part starts at 25:35 and he is totally speaking my language.

Exactly. It's really stupid if you're buying something with a focus on performance and overclocking. Stupid if you want better performance. Stupid if you're going to overclock because running at default Intel specs is probably never going to happen. You don't want it to and never will.

 

I know, right? And, once again... close, but no cigar. Typical Gigabyte design bipolar disorder, unfortunately.

• Nice overclocker-centric features... crazy good array of buttons/switches (I love this part)
• Not enough power phases for an extreme board (Z490 Dark has 18)
• VRM cooling design may be OK, but doesn't seem particularly robust (I'd be worried)
• Ugly design due to lack of IO cover (inexcusable at this price point)
• Ugly and inconvenient placement of dual 8-pin power ports
• Puzzling inclusion of too many gamer-boy features that detract from an extreme overclocking focus
  • WiFi, BT, lots of RGB headers, TB add-in card support, etc
  • Wasted the extra 4 CPU lanes on gamer crap (i.e. extra M.2) instead of an extra PCIEX4 slot
• Feature attributes are inconsistent and seem to clash.
  • If you're doing the gamer-boy thing, at least go to the effort of making it look nice.
  • If you're catering to extreme overclockers, then actually do that 100% on everything.
• Gigabyte firmware UI is the same clunky and somewhat chaotic design as before
• Lack of driver support for legacy OS is not an extreme overclocker attraction

 

Credit where it's due - the Core i5 11400 is the new value CPU champion, at least when it comes to gaming. If nothing else, it'll function as a stop-gap if you can't wait for Alder Lake but are short on cash.

 

Agreed, even the 11600k has value @ $269.99 vs the 5600x @ $299.99 if you're going to overclock. For locked frequencies, the 11400 really is the new budget buddy.

 

https://hothardware.com/news/amd-wa...el-security-vulnerability-in-zen-3-processors

Interesting -
"The attack is similar in scope to Spectre and involves a feature introduced with Zen 3 called Predictive Store Forwarding (PSF). PSF, in effect, guesses the result of a load and uses speculative execution with subsequent commands. "In typical code, PSF provides a performance benefit by speculating on the load result and allowing later instructions to begin execution sooner than they otherwise would be able to," AMD explains.

"Most of the time, the PSF prediction is accurate. However, there are cases where the prediction may not be accurate and cause incorrect CPU speculation."

AMD outlines two cases where an incorrect PSF prediction can occur. The first is when either a store or load address changes while a program is executing. The second is when "a store/load pair which does have a dependency may alias in the predictor with another store/load pair which does not."


While AMD was able to escape most of the ill-effects of previous Spectre and Meltdown attacks, it explains that there is a possibility that malicious actors could use the PSF to carry out side-channel attacks on a Zen 3 system:

Because PSF speculation is limited to the current program context, the impact of bad PSF speculation is similar to that of speculative store bypass (e.g., Spectre v4). In both cases, a security concern arises if code exists that implements some kind of security control which can be bypassed when the CPU speculates incorrectly. This may occur if a program (such as a web browser) hosts pieces of untrusted code and the untrusted code is able to influence how the CPU speculates in other regions in a way that results in data leakage. This is similar to the security risk with other Spectre-type attacks."

 

CPU Vulnerabilities, Meltdown and Spectre, Kernel Page Table Isolation Patches, and more

It was only a matter of time. Predictable and not surprising, but it was kind of sad that the AMD fanboys actually believed they were "more secure" than they were with Intel. We will never live in a world without security issues and vulnerabilities. Anyone that thinks so has lost their grip on reality. Nothing made by Intel, AMD, NVIDIA, Micro$lop, Apple, or any other technology provider will ever be secure as long as bad people have a heartbeat and a pulse. As AMD-powered platforms become more popular they will naturally become more vulnerable.