>Seems like Gen 3 Ryzen (Zen 2) is coming sooner than we think. wccftech has reported that some motherboards are currently getting BIOS updates to support "upcoming CPUs" > Currently, ASUS, MSI, and Biostar have released the new AGESA 0070 and AGESA 0072 updates for their respective motherboards. Not all motherboards are currently receiving the BIOS updates but would do so shortly. Both ASUS and MSI clearly list that the BIOS supports “New Upcoming AMD CPU” and because we know that there are no more Zen+ processors on the horizon for the X470 or X370 platform, the only processors that make sense to add support for is the AMD Ryzen 3000 series. wccftech.com/amd-ryzen-3000-valhalla-cpus-x370-x470-motherboard-bios-support/
It's happening, guys. Is it safe to expect a June release?
>resetting the AGESA version number once again AMD is retarded
Ian Reyes
>because we know that there are no more Zen+ processors on the horizon for the X470 or X370 platform, the only processors that make sense to add support for is the AMD Ryzen 3000 series
Wonder if it's worth sticking with my R5 1600 until AM5 generation CPUs or getting the final AM4 processors (assuming it works with my B350) and wait until 2nd generation AM5 processors that iron the bugs out?
Mason Fisher
The Bios updates are def for Zen 2. Updating Desktop Boards with Mobilechip AGESA would be beyond retarded. And the new Mobilechips where the only new thing in a while the anounced for release from the CPU side.
Justin Clark
and the Athlons are already suported in older Bios
Ethan Evans
I'm getting rid of both my r5 1600 and b350 pc mate. I bought a cheap motherboard specifically to upgrade to X570 (or X590) when PCIe 4.0 comes out. It was obvious 1st gen mobos will be messed up. Don't think there will be anything wrong with 3rd gen mobos, if anything, AMD may fuck up the I/O die but upgrading a mobo won't solve that anyway. I think next generation will be a refresh with no real improvements so x570 is the new z77.
Thomas Richardson
My b450 already got the new beta bios the asrock b350 itx i still got flying around probably to
Lincoln Ortiz
>June wow, it's so soon. Only 3 months, JUST WAIT!
Parker Baker
>if it isnt tomorrow its too far away
Repent Zoomer.
Elijah Lopez
To be honest i was going to stick my current mobo into a HTPC case and swap out the CPU for an APU
Brody Hernandez
I just hope there are X590 motherboards with 8 SATA ports right at launch, I need as much, and in the current offering they're quite rare already. I wouldn't mind going X470 to grab such a card, but I don't want to deal with getting one that wouldn't have its bios already updated. I'll want to use my new rig right away.
>C6H getting updates Oh baby. I can't believe ASUS didn't lie about that.
Christopher Edwards
When are Intel's 10nm high performance CPUs being released?
Joseph Harris
I'm gonna laugh when older AM4 MOBOs won't be able to handle the higher core count zen2 chips.
Matthew Allen
At rough the same time as AMD's probably. But history tells us Intel will pull marginally ahead. If AMD just released the CPUs in January they could've at least enjoyed a few months of being market leader.
Ethan Brown
I said 10nm, not 14nm+++++++
Ayden Allen
I'm kind of hoping they are released in late 2020 so it might have DDR5.
first servers, then phones, then laptops, then desktop
Kevin Collins
They just releasing all of them at the same time.
Austin Rodriguez
Why are you gonna laugh? Anyone with an old board would still be able to jump to 7nm 8c/16t, pretty good stuff. Unlike Intel, that you need to buy a new mobo for every fucking new launch, even if it doesn't have a new + in front of the 14nm.
Juan Taylor
Dunno about laptop, plenty seem to still use LPDDR3
William Lewis
mama su please release close to my birthday (6/17)
Austin Baker
any recent laptop uses LPDDR4 for newer CPUs, even cheap stuff like the latest HP Stream
Jace Baker
The real question is what mobos will still support windows 7 with zen2.
John Brown
Just get an extension card or something. Spamming mobos with legacy SATA ports is dumb and should definitely stop.
James James
There's no x86 CPU with lpddr4 support though.
Xavier Williams
it's not 2009 anymore user
Ethan Flores
Unfortunately, Zen 2 ain't going to live up to the hype. >Last month Mubadala Investment Co (the owner of global foundries) sold ALL of its AMD shares. They're probably aware of some critical flaw in the 7nm manufacturing node. >So far this year Lisa Su (AMD's CEO) has been gradually dumping her shares in million dollar batches despite the fact that she was gradually accumulating all through almost all of 2018. >The single threaded performance of the chip AMD demoed at CES was not very impressive. AMD's top executives are all panicking. See pic related.
The critical flaw is probably *doesn't hit 5ghz* which was a hold over of rumours from ibm/GF 7nm that has been shitcanned. I don't think TSMC ever published their target fmax but the claimed 15% performance is probably 4ghz (2700x) increasing to ~4.6ghz which is pretty good but the internet will blow it totally out of proportion to be a failure on the scale of itanium. Regardless 5ghz is hard even on power and heat even if the process is capable of it.
Mubadala seems like they want out of the semi conductor business altogether. Halting gloflo, pushing hard for profit only, my bet is they'll sell GF the first moment they can.
Which paper? Is this the one that's talking about each ccd having a memory channel each in certain modes?
Parker Edwards
The sandnigger wants to get rid of GloFo anyway. Who's going to buy it? Samsung? If it doesn't reach 5GHz while still beating Intel I'll be ok with it
Ian Adams
It was at an ISSCC session, showed the VRF curve, inflection point at 4.6ghz IIRC, that doesn't exactly change with design or temperature
When will they release the price list or any other information about Zen 2?
Cooper Morales
>beat a same core count intel chip at lower clock speed SMT undoubtedly helped AMD. Also, the zen2 chip won by a hair. The 9900k's single threaded performance is almost certainly higher than the single threaded performance of the chip they demoed.
The Zen 2 CES demo vaguely reminds me of what AMD did leading up to the Polaris launch. They did a demo with an underclocked card in a cherry-picked benchmark, bragged about the performance per watt, and people assumed that the final card would smash anything nvidia had to offer. In reality, Polaris turned out to be meh/10.
It would be funny if AMD intentionally sent AdoredTV fake info just to pump the share price while the top executives sell. Their stock is currently up 39% YTD due to Zen 2 hype despite all the insider selling that is going on.
>>Last month Mubadala Investment Co (the owner of global foundries) sold ALL of its AMD shares. They're probably aware of some critical flaw in the 7nm manufacturing node. Probably huge latency from having the IO die seperate from the Core die. Which fucks latency over hard. If you though it was bad for gaming before, then boy this will be surprise for amd fucks
Isaiah Butler
Zen2 ticks all the right boxes and the only way AMD can fail it is if Zen2 were deliberately sabotaged.
There's legit no way to lose (from a design standpoint). Better SMT design, wider pipes, a doubled FPU, they built working product (albeit a poorly binned housefire GPU) on the node already. Zen2 with real IPC gains and even a half decent clock speed increase is a market killer. Intel has nothing, nothing at all, unless they've been sandbagging 10nm so hard that nobody knows it actually works flawlessly and is ready for mass market this year.
All this talk...hmmm....seems like SA or MF shilling to short investors before another stock spike. AMD is now poised to double their server marketshare by 2020's end
If this is true then they're fucked, but with all the improvements they made in the IF it can't possibly be worse than what they have now.
Jose Thomas
>Ryzen 5 3600G >overclocks to 4.7 Ghz >with RX 560ish like GPU >for $200
Grayson Richardson
Should be coming sooner, otherwise motherboards manufacturers would have waited longer to release bios updates.
Charles Young
I hope Zen2 can be overclocked 50% on all cores with stock cooler
Robert Turner
It's AMD retards they will be good CPUs but not godlike like everyone always hypes them up to be.
Thomas Morris
That's my board, what did they not lie about?
Josiah Murphy
>the owner of global foundries >global foundries brainlet. they don't even have 7nm.
Logan Hall
But this time it's different, they're going from 12nm to 7nm, unless they somehow butcher the 7nm process I have no idea how pic related wouldn't be real. It's the first time I'm actually wondering if Intel will be able to beat AMD, there's no way any of Intel's 14++ nm chips are better than AMD's zen2 ones.
I want them to do good, but let's get real it's always the same shit. Overhyped with ok performance. Just like at the Radeon VII
Liam Lewis
why they gonna double the consumer core count and burn all TR owners. 16 core fake af, 12 core probably real tho.
Joseph Collins
who the hell hyped Radeon VII?
Jayden Russell
We all knew Radeon VII was gonna be shit though, Zen2 leaks, if real, indicate that's it's actually gonna be good. Anyway, I can't wait to see how it's going to turn out and I can't wait for actual news from AMD or even Intel.
Lincoln Price
>burn all TR owners Core count alone doesn't make HEDT, it's the PCIe lanes
Liam Young
So, are they going to actually release new APUs? I'm getting conflicting answers.
Juan Baker
But things not optimized for amd still won't be optimized for amd right? So for stability and consistency wouldn't Intel still be the go to?
Gabriel Brooks
Well we won't know until people actually get the chips in their rigs
Jaxson Murphy
What I expect AMD to do is have a bit of overlap (not much) between Ryzen and Threadripper.
6-12 cores would be purely Ryzen. 16C would have a Ryzen and a Threadripper, and from there it would be threadripper. My question is what the Threadrippers are going to top out at. Are they going to cap them at 32 cores? 48? or are they going to go the whole way like they did with the 2990WX and put the same amount of cores Epyc tops out at, just with less I/O?
Robert Cox
>muh add in cards Flex the quad channel memory and the sexy mobos who needs that many lanes unless they really like their fast storage
David Howard
64 cores on quad-channel RAM is a bit too much, 48 cores might be doable though
Lincoln Butler
Thats what I've been thinking. Even if AMD did enormously improve IF bandwidth and IMC clock rate compatibility, its just not enough bandwidth to feed that many cores. The cache boost is expected to help some, but we don't know how much just yet.
Ethan Peterson
>What I expect AMD to do is have a bit of overlap (not much) between Ryzen and Threadripper. Yup, Zen 1 also had some overlap. The top ryzen chip and bottom tr both had 8 cores. The only thing differentiating them was memory and pcie lanes.
>My question is what the Threadrippers are going to top out at. They'll probably keep their top tr chip at 32 cores. The 2990WX already has problems with cache misses and shitty memory bandwidth. It would be dumb to add more cores without moving to a socket which supports 6+ memory lanes.
Asher Jenkins
>The 2990WX already has problems with cache misses and shitty memory bandwidth Cache is expected to be handled by the doubled L3, and bandwidth could be blamed on the fact 2 of the dies didn't have a memory controller to directly access, and had to fight for resources on the IF. If AMD has improved IF (rumor has it it's not hard-coupled to memory clocks anymore) then it'd enormously help that. I dunno, we'll see what AMD's got waiting for us.
Cameron Adams
>doubled L3 This won't help when the windows task scheduler goes full retard and moves a thread from one CCX to another. AMD would either need to pressure Microsoft into getting its act together or add a gargantuan L4 cache on the IO die.
>Just like at the Radeon VII lmao Anyone that knew shit of tech knew it was just ported MI** with how they presented it. Vega is great at some things, but it's very mediocre on gaymen. Stop falling for the drone delusions.
Josiah Turner
>64 cores on quad-channel RAM is a bit too much This ha been disproved already with the 2990WX, it's just Windows 10 being absolute dogshit.
Alexander Jackson
32 cores on quad-channel is doable with some caveats, and yes part of the issue is that window's scheduler is dogshit and will forever be dogshit. 64 cores though is literally double the bandwidth requirements. Unless AMD has significantly improved the IMCs to allow for higher memory clocks, or stuffed in an L4 cache on the I/O die to assist with feeding the cores, 64 probably will not happen on quad-channel. 48, maybe, but not 64.
Brayden Jones
People set themselves up for failure to justify Nvidia (in the context of the VII). The VII is exactly what AMD said it would be - a 2080 competitor. Nothing more, nothing less.
Anthony Jackson
Dual ranks 80GB ddr5 when.
Charles Mitchell
The new Gen 11 graphics for icelake, look at the last row and think about it. Next Desktop Socket from AMD might need 4 Channel. am4 till 2020 that would be next year. So am5 with 4 Memechannel ?
i just bought a 2600x yesterday and a msi x470 gaming plus, together with 32gb ddr4.
dont know if regret or not.
Jacob Sanchez
Plenty of people have overclocked 95w parts past 130w, so power draw isn't going to matter.
Lucas Davis
HOW UPDATE BIOS OP? HALP!! I want download better performance!
Eli Fisher
The difference is that the Polaris demo wasn't particularly impressive performance-wise. Even if you ignore the power figure, the Zen demo is still AMD beating Intel's absolute top offering.
Alexander Kelly
DDR5 has twice the bandwidth so is similar to quad channel DDR4.
Anthony Morgan
The only virtualization feature I know of that Zen can't do is Hyper-V nested virtualization, and that's something that can be fixed as soon as AMD/Microsoft start giving a shit.
I can understand why they don't, because that's pretty fucking niche.
Kevin Jackson
Can't be worse, probably like ~10% better
Nicholas King
I'm on X370 and I'm gonna upgrade to a 3600X. What's the problem with that?
Aiden Martinez
The bios won't have enough memory for the update
Nathan Long
they could, it wont be anything interesting though, as APUs are one generation behind. 3000 series APU = zen+.
Lucas Miller
My 1600 and x370 becomes my home server when 7nm comes out. As long as it is not a flop that is.