Support NeoGAF

[Coulomb_Barrier]

Steady. Fake is a bit much given the things that have been right but the leak as a whole is otherwise still unconfirmed. The OP made it seem like this was a separate confirmation which is too far in the other direction, hence my post to try to keep things real.

Alongside the other info I've posted, worth considering the Zen 1 Engineering Samples.
The 8 core 95W TDP chip (among others) surfaced July 2016. Base clock 2.8GHz, boost to 3.2GHz.
Final 1800X chips (8 cores 95W) had 3.6GHz base, 4GHz boost. Flat 800MHz improvement there.
Works out as 28% uplift on base, 25% on boost from ES to retail.
Applied to that tasty 16 core would result in 4.2GHz base, 5.2 GHz boost which is amusingly close to a purported 3850X

Hold your objections - I know this isn't the same exact situation. To my knowledge, frequencies and the like don't scale linearly. There is also a longer period between that Zen 1 ES leak and retail release (July -> February) compared to TUM_APISAK's leak and the expected release. Different process. Different engineering challenges: brand new arch vs evolution of existing arch with new packaging strategy.
Entirely possible that 16 cores may be held back from initial release in a couple of months - there have been rumblings to that effect thus far.
Not PROOF™ but I personally still find it notable and interesting, even with the caveats.
A mere 400MHz improvement upon that sample just seems way too much of a low ball.

I see your logic but it's not a case of what's possible on 7nm it's a matter of what's realistic. Running not 8 not 12 but 16 cores with a 4.3Ghz base clock will result in a gigantic TDP so why would AMD do it when said CPU will run just as fast running at 3.3Ghz base so long as the boost algorithm is the same?

Anyway I've had enough on speculating these figures.

 

[Xdrive05]

Ha! I just built a 2600 rig on an MSI Tomahawk. Someday I might pick up a 3600 and drop in this board as a slight upgrade. Two more cores for almost the same price is wild!

 

[PhoenixTank]

I see your logic but it's not a case of what's possible on 7nm it's a matter of what's realistic. Running not 8 not 12 but 16 cores with a 4.3Ghz base clock will result in a gigantic TDP so why would AMD do it when said CPU will run just as fast running at 3.3Ghz base so long as the boost algorithm is the same?

Anyway I've had enough on speculating these figures.

I'm not looking to make anyone a believer. Believe if/when we see it. I know there were posters prematurely celebrating on the first page and one called you out, but I still maintain fake was too strong, hence my post.

I know it seems unrealistic, but only really against the background of Intel sandbagging, the failed run at the mobile market and the foundry side of the business stumbling hard against Moore's Law. That is business as usual for the past half decade or so. Intel persevering and continuing to struggle forward down the same road undaunted is admiral in some respects... but we'll probably never know how quickly or slowly Intel were planning to move away from monolithic chips before AMD caught them with their pants down.

Zen has some very nice clock/power gating but I've never really looked into it until now. If I'm to believe a post from Reddit and some less specific sleuthing elsewhere... under no load the frequency of a Zen core is basically irrelevant to power consumption. I'd imagine the idle clock will still be lower than base but it doesn't make it so impossible for high base clocks. For actual load, the power/performance curve may not be as unfavourable as you expect - you don't want more figures so I won't give you any. The purported 3850X will have to be cherry picked, but they just need any two great (and comparatively tiny) 8 core chiplets to make that happen.

 

[Ascend]

I see your logic but it's not a case of what's possible on 7nm it's a matter of what's realistic. Running not 8 not 12 but 16 cores with a 4.3Ghz base clock will result in a gigantic TDP so why would AMD do it when said CPU will run just as fast running at 3.3Ghz base so long as the boost algorithm is the same?

Anyway I've had enough on speculating these figures.

Not if you bin the best chiplets. And take into account that with the node shrink to 7nm, the optimal point on the performance curve between frequency and power consumption shifted to a higher point in terms of clocks. The question is how much.

The current best 8 core chips of Zen+ run 3.7 GHz base with 4.3 GHz boost with a TDP of 105W (2700X). The leak suggests that the 8 core CPUs can run at 4 GHz base and 4.8 GHz boost with a TDP of 95W (3600X). Does that seem unrealistic, taking into account the shift from monolithic dies to chiplets and the node shrink? To me, it doesn't. We can also expect AMD to bin the best chips for the 16 core CPU to save power, which means it should perform better than this 8 core...
The interesting part becomes the 3800X, a 16 core CPU with a core clock of 3.9 GHz and a boost clock of 4.7 GHz, and a TDP of 125W. It's clocked slightly lower than the 8 core 3600X, but uses 30W more power due to the additional cores. We can try and extrapolate a few things... But maybe I'll do that at a later time. Just remember that the I/O die is still 14nm, and it too requires power, although it's currently unknown how much it will use. You can't assume that double the cores should be double of 95W at the same clock speeds. Binning and the I/O die skew that in favor of the 16 core.

Most likely the power consumption doesn't increase much either, until reaching boost speeds, similar to the first gen Zen processors. See the chart below as a reference...;

Taken from;


Assuming a similar curve for Zen 2 (the graph is Zen 1), we can sort of extrapolate the difference in power that the 16 core and the 8 core Zen 2 CPUs have.

 

[JohnnyFootball]

MSI X570 Gaming Pro Carbon and X570 Gaming Plus Motherboards For AMD Ryzen 3000 CPUs Leak Out Too

The latest round of AMD X570 chipset powered motherboards has leaked out & this time we are looking at the MSI lineup for Ryzen 3000 CPUs.

wccftech.com

The MOBO leaks keep on coming!
MSI's X570.

 

[JohnnyFootball]

MORE LEAKS OF MOBOS!!!!!!! MSI, ASUS, and ASROCK

X570 Motherboards from MSI, Asus and Colorful Revealed (Update)

Computex is almost here, bringing along leaks and announcements of AMD X570 motherboards along with it.

www.tomshardware.com

ASROCK:

 

[JohnnyFootball]

I will probably go for an MSI board this time around. Kinda done with Gigabyte.

 

[Leonidas]

I wonder what's up with the chipset cooling fan on all these X570 boards.
Think I'll stick with my 400 series board...

 

[Ascend]

I wonder what's up with the chipset cooling fan on all these X570 boards.
Think I'll stick with my 400 series board...

Agreed... If PCI-E 4.0 is the only thing, which apparently is even getting support on some X470 and B450 motherboards, there really is no reason to go for an X570 with a noisy fan on it.

 

[The Skull]

I will probably go for an MSI board this time around. Kinda done with Gigabyte.

I had an x370 gaming pro carbon. It was horrendous. Swapped it for an Asrock Taichi. Fantastic board.

 

[JohnnyFootball]

I had an x370 gaming pro carbon. It was horrendous. Swapped it for an Asrock Taichi. Fantastic board.

I had an ASrock B350 and wasn't impressed. Asus is always a good choice, but supposedly their Z390 boards (Intel I know) were terrible. MSI is where I am looking. But I just want a well built board with high quality parts.

 

[PhoenixTank]

I wonder what's up with the chipset cooling fan on all these X570 boards.
Think I'll stick with my 400 series board...

Specifics aren't public knowledge but under some circumstances and workloads the chip will get too hot for passive cooling. Not expected to be all the time, but is a concern.

Agreed... If PCI-E 4.0 is the only thing, which apparently is even getting support on some X470 and B450 motherboards, there really is no reason to go for an X570 with a noisy fan on it.

Limited support on those older boards - less gen 4 lanes than X570. Tiny fans are usually annoying, though, so I hear you.

New video from AdoredTV. Still watching, but figure this is the best place to put it.


EDIT: Above video says m.2 RAID is the reason for the fan - see 12:37.

 

[kraspkibble]

Really hope that 3850X is real...

Last rumor I saw said that the 3700X will have higher clocks than the 16 core CPUs. I want the 3850X for 16 cores but if if the clocks are gonna be lower then I don't know. I guess 12 cores would be enough...

We'll see.

Also, not sure what board to get. Seems like nobody can settle on a good board. I was leaning towards Asus or Gigabyte.

 

[JohnnyFootball]

Really hope that 3850X is real...

Last rumor I saw said that the 3700X will have higher clocks than the 16 core CPUs. I want the 3850X for 16 cores but if if the clocks are gonna be lower then I don't know. I guess 12 cores would be enough...

We'll see.

Also, not sure what board to get. Seems like nobody can settle on a good board. I was leaning towards Asus or Gigabyte.

I want Asus, but supposedly they did a pretty bad job with their Z390 boards and it's giving me pause.

MSI is a company who I have never bought one product from, but everyone swears by them, so I'd say they're in the driver seat.

 

[JohnnyFootball]

AMD Zen 2 5GHz 12 Core, 4.3GHz 16 Core CPUs Leaked & Benchmarked

We're mere 5 days away from the announcement of AMD's highly anticipated next generation Zen 2 based Ryzen 3000 series CPUs which are set to make their debut next Monday on the 27th of May. In the meantime however, we've got a whopper of a leak for you to keep you entertained. Two of AMD's...

wccftech.com

If this is remotely close to accurate and the prices are true, then Intel is pretty much helpless until they can release their 10nm processors.

Let’s start with the 16 core part. Based on what’s being purported, this chip actually has a base clock speed of 3.2 GHz and only boosts to 4.3 GHz. This however, did not stop someone somewhere from overclocking it to 4.2 GHz on all cores and running a Cinebench R15 benchmark on it with quite an astonishing result.

If the leak is to be believed, this 16 core chip running at 4.2 GHz scored 4278 points in Cinebench R15. To put that into perspective an 8 core Ryzen 7 2700X scores 1828 points, a 16 core Ryzen 7 1950X Threadripper scores 3055 points and Intel’s 16 core i9 7960X scores 3163 points. In fact, it takes a core i9 7960X overclocked to nearly 4.8 GHz to match the leaked 16 core Ryzen 3000 chip at 4.2 GHz.

These results indicate that Zen 2 brings above 10% improvement in IPC in addition to higher clock speeds thanks to the 7nm process. Combine that with a doubling of cores, AMD’s Ryzen 3000 series could easily deliver double the multi-core performance of their predecessors up and down the stack.

The 3700X looks like it could be the next 2600K and possibly even surpass it.

Now, I am still skeptical of 12-cores having a base clock of 4.3 Ghz and a boost clock of 5.0 Ghz, but if true....................WOW.

 

[PhoenixTank]

If this is remotely close to accurate and the prices are true, then Intel is pretty much helpless until they can release their 10nm processors.

If true, even 10nm won't help them for a while. I can't remember if the leaked Intel roadmap was confirmed accurate or not, but they max out at 4 cores even in 2021. Probably not the full picture but everything points to 10nm yields being horrendous.

 

[SonGoku]

one called you out,

I was just having fun lol, and didn't realize at the time that the chart was regurgitated from AdoredTV.

 

[Ellery]

If it sounds too good to be true it probably is.

(I hope Zen2 is going to be a disruptive force raging through the CPU market, because I am going to buy one, but I am not sure yet if I believe the IPC gains and clockspeeds that are rumored)

 

[SonGoku]

If true, even 10nm won't help them for a while. I can't remember if the leaked Intel roadmap was confirmed accurate or not, but they max out at 4 cores even in 2021. Probably not the full picture but everything points to 10nm yields being horrendous.

How or why did Intel lose its fab prowess? How come TSMC surpassed them

 

[JohnnyFootball]

How or why did Intel lose its fab prowess? How come TSMC surpassed them

Probably due to Intel’s laziness due to lack of competition from AMD for so long.

 

[SonGoku]

Probably due to Intel’s laziness due to lack of competition from AMD for so long.

I get that for CPU designs... but they always prided themselves in being one or two steps ahead the competition in fab process
Intel is in love with mobile, makes no sense to neglect its fab

 

[PSlayer]

How or why did Intel lose its fab prowess? How come TSMC surpassed them

Their 10nm node is a nightmare from what insiders are saying.

They bet in the wrong technologies and when it didn't pay out they had to re-release 14nm again with slightly better clocks,then in the next year they added more cores to that at the expense of temperature(and still pretending the TDP was the same).Now they are in a situation where 10nm is still not ready,amd is about to reach competitive clockspeeds with similar IPC and equal or more cores and intel has no other choice but release 14nm again.

At this point,the only reasonable thing intel can do is drop the price of their products.They are not doing this right now probably because they don't want to change the public perception about the value of their product.

 

[xool]

I assumed the 7nm node is "pure" for both parties, tbf Intel is jumping straight to 7nm EUV

How or why did Intel lose its fab prowess? How come TSMC surpassed them

Table time :

(the International Technology Roadmap for Semiconductors (ITRS) publishes standards for node sizes )

ITRS​	Intel equivalent​	Samsung / TSMC /GloFo equivalent​	Notes​
14nm	14 nm	~10nm	Non-intel 10nm is slightly better than ITRS rules for 14nm
10nm	10 nm	7 nm	Everyone but intel is "exaggerating"
7nm	7 nm	(5nm probably)	Everyone using EUV

I think the deviation started around 14/16nm node - TSMC etc made a node ("10nm") that was slightly better than their previous 14/16nm designs - so they called it "10nm" when it was closer to (but a bit better than) the standards for 14nm.

Then because they already had "10nm" when they made a new smaller node they called it "7nm" despite it being equivalent to the 10nm [edit corrected] standard.

Intel stuck close to the standard. (the rest is marketing lies)

In reality, Intel and the others are actually pretty close -- TSMC/Samsung are just starting to mass produce 7nm and Intel is just starting to mass produce their 10nm - but the two products are actually comparable .. it's neck and neck ..

[edit - if you want to compare the numbers - I recommend https://en.wikichip.org/wiki/technology_node -and see the individual articles for https://en.wikichip.org/wiki/10_nm_lithography_process etc - you'll always see Intel's node being better (smaller) for a specific "node name" at after around 20nm ]

 

[SonGoku]

Table time :

(the International Technology Roadmap for Semiconductors (ITRS) publishes standards for node sizes )

ITRS​	Intel equivalent​	Samsung / TSMC /GloFo equivalent​	Notes​
14nm	14 nm	~10nm	Non-intel 10nm is slightly better than ITRS rules for 14nm
10nm	10 nm	7 nm	Everyone but intel is "exaggerating"
7nm	7 nm	(5nm probably)	Everyone using EUV

I think the deviation started around 14/16nm node - TSMC etc made a node ("10nm") that was slightly better than their previous 14/16nm designs - so they called it "10nm" when it was closer to (but a bit better than) the standards for 14nm.

Then because they already had "10nm" when they made a new smaller node they called it "7nm" despite it being equivalent to the 10nm [edit corrected] standard.

Intel stuck close to the standard. (the rest is marketing lies)

In reality, Intel and the others are actually pretty close -- TSMC/Samsung are just starting to mass produce 7nm and Intel is just starting to mass produce their 10nm - but the two products are actually comparable .. it's neck and neck ..

Ok so they are roughly on par with tscm having a slight edge (no horror stories)
Still begs the question how did the rest of fab business caught up with intel?

 

[PhoenixTank]

How or why did Intel lose its fab prowess? How come TSMC surpassed them

Unfortunately not a singular simple answer. A few years old but still relevant to the whole picture:

How Intel Lost $10 Billion and the Mobile Market - ExtremeTech

How did Intel -- the most advanced foundry in the world -- manage to spend so many billions of dollars on mobile technology with so little to show for it?

www.extremetech.com

The second part: https://www.extremetech.com/computi...le-market-part-2-the-rise-and-neglect-of-atom
Mobile focused but relevance is that, while Intel had no real competition, their focus and money were elsewhere and ultimately largely a waste. They didn't change when and where they needed to.
Physics plays a big part here too.

In reality, Intel and the others are actually pretty close -- TSMC/Samsung are just starting to mass produce 7nm and Intel is just starting to mass produce their 10nm - but the two products are actually comparable .. it's neck and neck ..

Yes, widely believed that Intel's 10nm and TSMC's 7nm should have similar performance. I'd say TSMC are still ahead, though. TSMC 7nm is in full production while Intel's 10nm is at best nearing the top of the hill of problems. By the time Intel are hitting mass production there TSMC 7nm+ should be here.

 

[xool]

Ok so they are roughly on par with tscm having a slight edge (no horror stories)
Still begs the question how did the rest of fab business caught up with intel?

Historically eg 65nm (PS3) IBM/Global Foundries and others were equal to Intel - then everyone but Intel started having problems at 45, 28, and 16/14 nm, dropping to 1 and then 2+ years behind - these are the same time as AMD's dark days when they just couldn't compete - a big factor was GlobalFoundries was years behind Intel on chips..

That's been the story for the last 10 years, but it wasn't always like that.

idk the full story behind the others falling behind, or why Intel lost it's lead recently.

 

[JohnnyFootball]

Their 10nm node is a nightmare from what insiders are saying.

They bet in the wrong technologies and when it didn't pay out they had to re-release 14nm again with slightly better clocks,then in the next year they added more cores to that at the expense of temperature(and still pretending the TDP was the same).Now they are in a situation where 10nm is still not ready,amd is about to reach competitive clockspeeds with similar IPC and equal or more cores and intel has no other choice but release 14nm again.

At this point,the only reasonable thing intel can do is drop the price of their products.They are not doing this right now probably because they don't want to change the public perception about the value of their product.

If AMD can meet (the absolute bare minimum expectation) the 9900K in terms of performance then Intel will likely have little choice.

 

[xool]

Yes, widely believed that Intel's 10nm and TSMC's 7nm should have similar performance. I'd say TSMC are still ahead, though. TSMC 7nm is in full production while Intel's 10nm is at best nearing the top of the hill of problems. By the time Intel are hitting mass production there TSMC 7nm+ should be here.

I'll rest when the parts in the shops and in stock .. Remember the Core i3-8121U ? - supposedly Intel had a 10nm chip in production back in early 2018 .. which at the time remember being hailed as another example of Intel's lead.

 

[scydrex]

I was thinking of building a PC with a 2200g and a b450 but i will wait for the Zen 2 to come out and the b550. I just hope AMD release the G version this year too.

 

[PhoenixTank]

I'll rest when the parts in the shops and in stock .. Remember the Core i3-8121U ? - supposedly Intel had a 10nm chip in production back in early 2018 .. which at the time remember being hailed as another example of Intel's lead.

Oh of course, a lot can change and these TSMC 7nm products largely aren't available to buy. Think it is pretty solid info that AMD are/were trying to get as much capacity as they could before Apple swoop in for 7nm parts i.e. The whole thing is far along at this point.
I believe the 8121U was an obligation to stockholders or something along those lines: "See!? We have a 10nm part in a product now - we weren't technically misleading you / performing damage control as to when 10nm would be here."
Quote from Anandtech's article on the 8121:

Intel’s official line is that this processor was released with a specific target market in mind, and it fulfils the role required. What exactly this market is, and at what price point, is still a mystery, even in 2019. However some analysts believe the graphics was a dodo out of the door due to uneconomically viable yields, as well as this chip not making any sense commercially for the product segment it ended up in – it was put into the market just to fulfil a promise to investors.

I thoroughly look forward to Monday.

 

[xool]

I believe the 8121U was an obligation to stockholders or something along those lines: "See!? We have a 10nm part in a product now - we weren't technically misleading you / performing damage control as to when 10nm would be here."

I almost feel sorry for Intel - to think they could make 10 (7) nm parts back in early 2018, and have been struggling ever since, presumably to get yields up to viable numbers.

 

[SonGoku]

I almost feel sorry for Intel - to think they could make 10 (7) nm parts back in early 2018, and have been struggling ever since, presumably to get yields up to viable numbers.

Why do you think that happened?

 

[xool]

Why do you think that happened?

idk.

Intel's basically said they "bit off more than they could chew" :

Krzanich explained that the company "bit off a little too much on this thing" by increasing 10nm density 2.7X over the 14nm node. By comparison, Intel increased density by only 2.4X when it moved to 14nm. Although the difference may be small, Krzanich pointed out that the industry average for density improvements is only 1.5-2X per node transition.

 

[JohnnyFootball]

Ryzen 3000-Series CPUs Allegedly Support DDR4-3200 Out of The Box, Overclock to DDR4-4400

DDR4-3200 is reportedly official memory speed supported on the upcoming Ryzen 3000-series, codenamed Matisse, processors.

www.tomshardware.com

Ryzen 3000-Series CPUs Allegedly Support DDR4-3200 Out of The Box, Overclock to DDR4-4400


Things just keep getting better and better.

 

[JohnnyFootball]

AMD Zen 2 6 Core CPU Benchmark Leaked, Faster Than Ryzen 7 2700X in Geekbench 4

Another day another Zen 2 Ryzen 3000 series CPU leak and to call this one a shocker would be an understatement! So let's get straight into it! A peculiar CPU entry has been spotted in the wild, specifically in Geekbench's database. The entry carries the same string we've seen on all other...

wccftech.com

Are you excited yet? Computex needs to hurry the fuck up and get here. These babies need release dates!

 

[JohnnyFootball]

The live press briefing for this is actually in under 11 hours. 10 pm EST.

 

[The Skull]

Tempering expectations but man the hype for zen 2 is real.

 

[Leonidas]

I'm gaming at 4K so for me there won't be much of a difference between Zen 2 and my current 8-Core.

Will be interesting to see if AMD can match Intel in high refresh gaming though.

I can't see myself upgrading CPU till Zen3 (2020) or Ice Lake.

I'm more excited for Navi personally...

 

[Leonidas]

TLDW

-conference today (in 11 hours) = X570 unveil, limited Ryzen 3000 info
-Ryzen 3000 full specs revealed 6/10 (E3)
-16C power draw at 300w (alledgedly)
-7/7 launch review date

 

[Celcius]

E3 keeps giving me more and more to look forward to
The 16 core CPU sounds like it runs toasty.

 

[Slings and Arrows]

NM

 

[PhoenixTank]

E3 keeps giving me more and more to look forward to
The 16 core CPU sounds like it runs toasty.

[snip]

TLDW

-conference today (in 11 hours) = X570 unveil, limited Ryzen 3000 info
-Ryzen 3000 full specs revealed 6/10 (E3)
-16C power draw at 300w (alledgedly)
-7/7 launch review date

Regarding the bold parts of your posts: Very important to note the (now) pinned comment from GN:

There seem to be some overly ambitious viewers who are not listening to the video fully: We specifically and explicitly said the 16-core CPU can reach upwards of 300W with H2O overclocks. We absolutely did not say that's the stock CPU. Please stop spreading misinformation. The video title even says "OC Power," meaning "overclocked power." Not stock.

 

[Celcius]

My concern wasn’t based on the GN video but just from having 16 cores in general, especially if overclocking on air like I do. I’m definitely looking forward to the official clocks though.

 

[PhoenixTank]

My concern wasn’t based on the GN video but just from having 16 cores in general, especially if overclocking on air like I do. I’m definitely looking forward to the official clocks though.

Fair enough - you know what you're getting yourself into with 16 cores
I'm personally hoping XFR/PBO are going to make things nice and easy, so I can be a lazy shit about it.

 

[JohnnyFootball]

Im getting a big time hard on thinking about how awesome the 16-core CPU could be.

But....Ill probably get the 12-core CPU if it can reach significantly higher clocks.

 

[PhoenixTank]

Im getting a big time hard on thinking about how awesome the 16-core CPU could be.

But....Ill probably get the 12-core CPU if it can reach significantly higher clocks.

I hear you, but there are still plenty of things that could be subpar. Really have to hope that Infinity Fabric 2 and the IO die curtail the problems that Threadripper has in latency sensitive workloads.
Fingers crossed that we'll know enough in ~3 hours.
Apparently the official livestream will appear on this url: https://www.amd.com/en/events/computex

 

[Kenpachii]

Already said a few months ago i think that only one i am interested in is a 16/32 core solution specially if consoles use a 8/16 solution and i want a 5ghz clock and a 299 price tag on it with it.

People laughed that it was going to be totally unrealistic. Yet here we are. Those 6/12 cores are not going to age well.

 

[JohnnyFootball]

All AMD Ryzen 3000 CPUs Will Feature Soldered IHS, Expect Impressive Thermals & Great Overclocking Capability

AMD has confirmed that all of their upcoming 3rd Gen Ryzen 3000 CPUs featuring the Zen 2 core architecture will feature a soldered design.

wccftech.com

Ryzen 3000 CPUs will feature soldered heatspreaders.

AWESOME news for overclockers!

 

[Trogdor1123]

All AMD Ryzen 3000 CPUs Will Feature Soldered IHS, Expect Impressive Thermals & Great Overclocking Capability

AMD has confirmed that all of their upcoming 3rd Gen Ryzen 3000 CPUs featuring the Zen 2 core architecture will feature a soldered design.

wccftech.com

Ryzen 3000 CPUs will feature soldered heatspreaders.

AWESOME news for overclockers!

Why is that better?

 

[JohnnyFootball]

Why is that better?

Soldering a heat spreader to a CPU yields much much better thermal performance than thermal paste, which was causing heat issues on Intel CPUs.

Granted this isn't HUGE news since they have been using soldered since the first Ryzen CPU.

Nevertheless, it's good to have confirmation for the 3000 series.