It just remains to be seen if we get those more cores on AM4 or they save them for threadripper, hoping its am4 myself and when quickish
-
Competitor rules
Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.
AMD Zen 2 (Ryzen 3000) - *** NO COMPETITOR HINTING ***
- Thread starter Jamin280672
- Start date
It just remains to be seen if we get those more cores on AM4 or they save them for threadripper, hoping its am4 myself and when quickish
Ah so you're still at 1080p playing CS:GO, got it.
So AMD need to copy Intel's incremental tick tock screw up that put them in this current position of being beaten by AMD at it's own game. Yeah good idea mate...
A 16c 32t flagship mainstream CPU at the same price as the 9900K or cheaper and having a higher profit margin and outperforming said 9900K (if the IPC gains are anything to go by, then it will) will make them sell like hot cakes. Even Dell, HP and Lenovo et al will want to get in on that pie. You don't win by playing the other players game. You win by making the rules. If the rule for the flagship is now 16c then Intel hasn't got a response. They can't market a 10c CPU vs a 16c and expect to do well. Hell the 28c ice cooled monster is just about to release to AMD's 16c mainstream CPU and later this year a likely 64c ThreadRipper.
Pull your head out of your arse mate.
One cinebench score which has the ability to be analyzed and information extrapolated from it. There's a lot of information there and also a lot of questions. However the one fact is that the 8c single CCX die CPU just equalled the 9900K as an Engineering Sample. That's a fact. I say equalled as it's within error margin. There was a result prior to the main show which the 9900K beat the 3xxx by about 12 points. So yeah run to run variance. If there's an uplift of between 10% and 15% in IPC over previous Ryzen arch and there's a potential for clock speed increase we can work out pretty much where the frequency of the ES was. Purely on IPC increase it would put the ES at 4.08GHz - 4.27GHz. Seeing as most of the other leaks have shown 4GHz for the ES we could assume it was 4.1GHz for this test. Though I don't particularly like assumptions it's an approximation.
If that's the 3600 SKU (X or no X is mostly irrelevant at this stage) then it's a 5GHz potential SKU going off the leaked information, with a full functioning 8 core die. Assuming it can reach 5GHz then that would equate to ~2500 CB score vs the ~2180 for a 5GHz 9900K and as posted on Guru3D ~2300 for a 5.3GHz 9900K which validates the scaling of the 9900K on frequency. Kinda highlights that even at 5GHz the Ryzen 3xxx in a CB workload will beat the 9900K at 5.3GHz.
I've heard 16c 32t can help with that
Caporegime
- Joined
- 9 Nov 2009
- Posts
- 25,457
- Location
- Planet Earth
GN says the X570 is the biggest issue ATM currently and it looks like a June release. PCI-E 4.0 won't be on the chipset and it appears PCI-SIG will be using AM4 as their reference platform for PCI-E 4.0!
Caporegime
- Joined
- 18 Oct 2002
- Posts
- 33,188
NO one markets their next chip that isn't available for half a year as so good you absolutely shouldn't buy any of our stuff till this launches, because it's shooting yourself in the foot. You do the super hype full on marketing bash at launch, with product available when the old stuff has had time to sell off most of the inventory and price cuts are about to be implemented to clear the rest and the new stuff is ready to sell.
If you hype up the chip and give details/prices today, and not have them available for months you tank 10s of millions of dollars in inventory value as you destroy demand. You kill your Q1/Q2 revenue, take a massive stock hit as you have to write down earnings by 30-40mil at least for each quarter.
CPUs have a long lead time, that means a very small amount of info several times leading up to the launch.
Associate
- Joined
- 27 Apr 2007
- Posts
- 967
Well said drunken one.
In one way it was a beautifully understated presentation that only hinted at the possibility of what's to come.
16C on AM4 at a friendly TDP will be a killer at the sub £500 price point.
The real meat and potato is about pricing and performance below £350 which could well include 12C and a class leading or matching 8C.
Although they could price 16C at £350 - £400 and go for the jugular
In one way it was a beautifully understated presentation that only hinted at the possibility of what's to come.
16C on AM4 at a friendly TDP will be a killer at the sub £500 price point.
The real meat and potato is about pricing and performance below £350 which could well include 12C and a class leading or matching 8C.
Although they could price 16C at £350 - £400 and go for the jugular
was there any rumours about 16c and 12c ryzen 3000 working with higher end x470 boards? Some are suggesting that X470 boards will only support 8 core cpu, which would be a real shame for those of us with higher end x470 boards that have VRM suited to run higher core chips, like a 16core cpu.
I'm actually really angry. I came on here for chat about ryzen 3000 before work and it's just post after post by Easyryder and Dg trolling the ***p out of this forum. How are they not banned? On any other forum they would be.
Caporegime
- Joined
- 18 Oct 2002
- Posts
- 33,188
My take is this, most x470 boards will run a 16 core just fine. A ridiculously cheap, and I mean really really cheap low end board might not but any x470 board is basically designed for overclocking and can easily push a 2700x up to using well over 150W with no issue.
Lately you have a bunch of guys, particularly youtubers making all these claims about bad VRMs, for the most part when they say VRMs are bad they are really saying, but not specifically so being stupid, that the one high end board with way WAY overspec VRMs has enough juice to push say 300W comfortably and the 'crap' overclocking board with the 'crap' VRM is only capable of pushing 270W... so it can't match the board that can push 300W via VRMs. In reality, at stock or a light overclock there isn't a high end overclocking board.
There are two issues here, do the manufacturer in AMD or Intel, allow the mobo makers to update bioses to support and validate the new chips or do they want to sell more chipsets and get a cut of motherboards, and then if they allow it, will lets say Asus decide to use that permission to update their boards which may lead to them losing a sale as someone flashes their x470 instead of buying a new x570.
Intel have made a huge deal about intentionally blocking older chipsets using newer chips which stops most motherboard makers enabling support because they risk the wrath of Intel for going against them. AMD usually go the other way and would be more likely to let their partners update their boards.... but the partners really don't want to. Really it becomes if one big manufacturer wants to win long term customers by doing the right thing and being seen as the reliable company, then the others will probably fall in and support them also.
Realistically it would also be VERY easy to simply say a 16 core chip will use 95W and lets say for arguments sake, 4.2Ghz all core boost on x470 but will enable 125W mode on x570 enabling 4.7Ghz all core boost. Allow that to be the 'stock' setting that can be easily validated on all x470 boards, but in reality that is only a stock mode to meet a specification, you can overclock the 16 core just like any other chip and either use the normal higher clock speeds or overclock even further.
Basically AMD have talked a lot about supporting older motherboards so I can't see them getting in the way.
Then on the 8/12/16 core. I think Intel aimed 10nm at a good couple of years ago and Zen 2 was supposed to come out quite a long while after Icelake. AMD were likely banking on 14nm Intel staying 4 core and Icelake a year or two ago maybe moving to 6 core, but being on a process that it would be easy to scale to 12 core or so without much trouble. AMD planned Zen 2 to go against Icelake, they've said so. It's why it's designed to be 16 core on desktop. The thing is Intel couldn't have missed 10nm/Icelake releases worse than they have. Even now every sign is desktop Icelake is going to be a fair way into 2020. This means AMD doesn't have to go 16 core asap to beat Intel, however with 8 core 9900k.... 8 core Zen 2 is a waste imo. 8 Core puts them on par with Intel, maybe slightly ahead and for maybe 6-12 months. If they release at min a 12 core, if not 16 core at launch then they absolutely kill Intel mainstream immediately and utterly dominate for a year at which point Intel will likely only catch up, or worst case won't even catch up.
So they could go the route of I think at worst, 12 core at launch, hold back 16 core to smack Intel with at a later date and milk the 12 core for a while as with the 1800x, then bring in the 16 core at that price and drop 12/8 core pricing at a later date. Personally I think they are going for the straight killer blow, straight 8-16 core product range available from launch and nothing Intel produce in desktop is remotely competitive till into 2020.
Lately you have a bunch of guys, particularly youtubers making all these claims about bad VRMs, for the most part when they say VRMs are bad they are really saying, but not specifically so being stupid, that the one high end board with way WAY overspec VRMs has enough juice to push say 300W comfortably and the 'crap' overclocking board with the 'crap' VRM is only capable of pushing 270W... so it can't match the board that can push 300W via VRMs. In reality, at stock or a light overclock there isn't a high end overclocking board.
There are two issues here, do the manufacturer in AMD or Intel, allow the mobo makers to update bioses to support and validate the new chips or do they want to sell more chipsets and get a cut of motherboards, and then if they allow it, will lets say Asus decide to use that permission to update their boards which may lead to them losing a sale as someone flashes their x470 instead of buying a new x570.
Intel have made a huge deal about intentionally blocking older chipsets using newer chips which stops most motherboard makers enabling support because they risk the wrath of Intel for going against them. AMD usually go the other way and would be more likely to let their partners update their boards.... but the partners really don't want to. Really it becomes if one big manufacturer wants to win long term customers by doing the right thing and being seen as the reliable company, then the others will probably fall in and support them also.
Realistically it would also be VERY easy to simply say a 16 core chip will use 95W and lets say for arguments sake, 4.2Ghz all core boost on x470 but will enable 125W mode on x570 enabling 4.7Ghz all core boost. Allow that to be the 'stock' setting that can be easily validated on all x470 boards, but in reality that is only a stock mode to meet a specification, you can overclock the 16 core just like any other chip and either use the normal higher clock speeds or overclock even further.
Basically AMD have talked a lot about supporting older motherboards so I can't see them getting in the way.
Then on the 8/12/16 core. I think Intel aimed 10nm at a good couple of years ago and Zen 2 was supposed to come out quite a long while after Icelake. AMD were likely banking on 14nm Intel staying 4 core and Icelake a year or two ago maybe moving to 6 core, but being on a process that it would be easy to scale to 12 core or so without much trouble. AMD planned Zen 2 to go against Icelake, they've said so. It's why it's designed to be 16 core on desktop. The thing is Intel couldn't have missed 10nm/Icelake releases worse than they have. Even now every sign is desktop Icelake is going to be a fair way into 2020. This means AMD doesn't have to go 16 core asap to beat Intel, however with 8 core 9900k.... 8 core Zen 2 is a waste imo. 8 Core puts them on par with Intel, maybe slightly ahead and for maybe 6-12 months. If they release at min a 12 core, if not 16 core at launch then they absolutely kill Intel mainstream immediately and utterly dominate for a year at which point Intel will likely only catch up, or worst case won't even catch up.
So they could go the route of I think at worst, 12 core at launch, hold back 16 core to smack Intel with at a later date and milk the 12 core for a while as with the 1800x, then bring in the 16 core at that price and drop 12/8 core pricing at a later date. Personally I think they are going for the straight killer blow, straight 8-16 core product range available from launch and nothing Intel produce in desktop is remotely competitive till into 2020.
My advice is block posters you think aren't participating in a reasonable manner. Makes the world of difference
The thing is, by this post it seems you would like to ban anyone who doesn't share your'e blinkered view on a Company or hardware release. The people who should be banned are the people posting fake information and data about an unreleased part.
The discussion with regard to the 9900k is valid due to it being the direct competitor currently with Zen 2.
Its not complicated. Am I allowed to post if I agree with you instead? Because that is not a discussion forum..Thats a ideological repressive state apparatus.
Agree with Us or be banned.
So by the same token. I find your post offensive to my free will and own thought processes in a discussion forum.
What gives you the right to police what I say or choose who posts in certain Hardware threads? Is the caveat now " you can only post if you own the hardware and speak postively about it"
Well that is fundamentally wrong to the Constitution of free speech. Something that are forefathers lost their lives for...So think about that on your way to work this morning.
Caporegime
- Joined
- 8 Jan 2004
- Posts
- 32,614
- Location
- Rutland
You are absolutely dominating a thread on Ryzen with only one thing you want to talk about, your 9900K.
It's ok to let other people have a different opinion just like you say. It's ok to let people get excited, it's not hurting you.
It's ok to let other people have a different opinion just like you say. It's ok to let people get excited, it's not hurting you.
Its not the only thing I wanted to talk about. The 9900k focus was comparing and discussing the Cinebench performance that was shown in the CES stream.
My point was...any discourse that doesn't whole heartedly worship Zen 2 is deemed as trolling which is clearly wrong. The Radeon 7 is lacklustre 2 years after 1080ti but offers the same performance....will I be vilified if I'm negative about that?
The Forum needs balance this should be encouraged.
Just click where it says DG or Easyrider and then when the box pops up click Ignore.. problem over.
Im a bit disappointed by the Vega 7 thing, its clearly a pro card rebadged as enthusiast, but the Zen2 stuff looks incredibly promising. 8c midrange ES sample matching Intels top end i9 at lower TDP... looks like Intel may have some people getting fired soon lol, they'll be doing all they can right now to get their 10nm stuff competitive and out the door ASAP is my guess.
Will be interesting to see if the higher core count ryzens actually have the higher clocks as the heat generation may be an issue on that small packet, thats a lot of cores to cool at high clocks. Might be that the 8c 16t single chiplet chips end up clocking the highest, which would also avoid any CCX crossing issues / latency etc. Obviously we have those rumored Zen2 leaks, but like any leak / rumor they might not be true.
Interesting times ahead for sure, but im a bit bummed that we have to wait til mid point of the year, i was hoping April / May at the latest for my new build
Im a bit disappointed by the Vega 7 thing, its clearly a pro card rebadged as enthusiast, but the Zen2 stuff looks incredibly promising. 8c midrange ES sample matching Intels top end i9 at lower TDP... looks like Intel may have some people getting fired soon lol, they'll be doing all they can right now to get their 10nm stuff competitive and out the door ASAP is my guess.
Will be interesting to see if the higher core count ryzens actually have the higher clocks as the heat generation may be an issue on that small packet, thats a lot of cores to cool at high clocks. Might be that the 8c 16t single chiplet chips end up clocking the highest, which would also avoid any CCX crossing issues / latency etc. Obviously we have those rumored Zen2 leaks, but like any leak / rumor they might not be true.
Interesting times ahead for sure, but im a bit bummed that we have to wait til mid point of the year, i was hoping April / May at the latest for my new build
One question I have is...If that was a mid range part and its beating a 9900k why didn't they show a 16/32 chip destroying the 9900k
It maybe that the 8c/16 thread chips are cheaper say £250 offering 9900k performance for less. While the R7 and R9 versions are stacked with cores and offer brilliant Productivity , come in cheaper than threadripper, but are clocked lower and dont perform so well in games?
There are plenty of variables.
A 5ghz £250 8/16 thread 3600x gaming CPU, beating a 9900k would be some feat..Exciting if it pans out that way
So you reckon AMD have a near 20% IPC advantage over Intel now, then? Interesting.
Assuming Zen 2 isn't a regression from Zen+, we know the chip can't have been running at over 4.7 GHz. It was likely running somewhere between 4 and 4.5 GHz.
Last edited:
There is a very simple answer to that. Who would give a damn if AMD showed a 12c or 16c chip beating the 9900k in cinebench? Absolutely no one. It means nothing.
The fact it is news worthy is the very fact that it was 8c vs 8c. Something AMD have not been able to do till now. It shows they have both caught up and overtaken (slightly) Intel at IPC/Clocks. (Ergo.... GAMING).
Last line is exactly why, sales of current models for half a year would be a major reason as @drunkenmaster mentions above also. Or just not stable yet who knows, all signs point to it being inevitable... also the recent Intel stupidity in comparison bench marks, no way you would want to compare non like for like.
This is a major problem but turning it on it’s head if you were Asus do you really want to keep on having to R&D new boards, packaging etc every year? Would you not make a far greater profit if your board was listed as compatible with more CPU skews? Obviously none of us have the numbers to work that out but I would think it’s possible it’s worse and particularly for smaller board vendors. If you look at the price of some of the mid/top tier boards it’s getting pretty punchy, partly due to continued increasing costs perhaps or simply the need to differentiate from the last product still in market?
That said the board vendors still make more variants of AM4 but without going and counting them all probably (hopefully) ultimately less in its life by some way, complete agree on the the lack of need and pretty inaccurate stigma of ‘bad VRM’ which doesn’t matter for many consumers. I would personally much rather see fewer skews and much better differentiation between them, less rinse and repeat will also no doubt bring down cost for all which is never a bad thing and more focus on areas like a better BIOS which is a relatively lacking area for most (looking at the work Elmor put in at Asus formerly shoes what that can achieve).
If you really wanted to be a bit of an eco warrior it’s also incredibly wasteful of finite resources to encourage such an unnecessary update. Not my bag though