AMD preps Ryzen 9 3900XT, Ryzen 7 3800XT, Ryzen 5 3600XT for June 16

[Void Master]

3 minutes ago, Jumballi said:

if these are ryzen 3000, then they should work without any bios updates.

MAYBE they want u to think that they can't...... 

Hmmm

someone should understand where am i going with this... 

[Jumballi]

Just now, Void Master said:

MAYBE they want u to think that they can't...... 

Hmmm

someone should understand where am i going with this... 

No we get that you're implying companies bad, the problem is that the enthusiasts are vocal and informed.

[Void Master]

Just now, Jumballi said:

No we get that you're implying companies bad, the problem is that the enthusiasts are vocal and informed.

IDK , i am just thinking of the worst case scenario .......

Most probably this rumor is fake...

[Jumballi]

10 minutes ago, straight_stewie said:

At how many years of refreshing do you draw the line? When exactly does it stop being an acceptable thing to do?

As long as it doesn't affect the health of the company, If intel releases the i7-3770 as the new pentium, I'd be all for that. Customers are not happy when the same CPU is rereleased with marginal improvements, but it still has all the flaws and vulnerabilities of the previous generation, that hurts trust(and the customers).

[Bombastinator]

39 minutes ago, straight_stewie said:

At how many years of refreshing do you draw the line? When exactly does it stop being an acceptable thing to do?

 

 

Oh I don't doubt that at all.

My point is that, as enthusiasts and machine builders, we shouldn't care at all about how many times a company has improved an existing design vs. making a new one. Nor should we care why they did it. Those things are of no importance what-so-ever to someone who does not have a financial stake in the companies under question.

We should only care about performance or performance/cost. I'm not even saying that Intel is better at either of those, just that we should stop making arguments about architecture refreshments, because they have absolutely no bearing on anything what-so-ever to those who do not have a financial stake in the company under question.

Mmm.  Maybe.  “Financial” is too specific I think.  Also should or shouldn’t is just as irrelevant.  If this was true it would apply to every consumer good.

[SpaceGhostC2C]

On 5/23/2020 at 7:24 AM, porina said:

I find this curious. Is there time for a refresh if we're to get Zen 3 later this year also? Feels a bit tight. So if this refresh is true, could Zen 3 be further out than we expect?

 

Alternatively, where is this extra performance coming from? I think it safe to assume they haven't changed the architecture at all. Has there been any changes in process? Like manufacturing tweaks and similar at TSMC.

My impression, if true: Epyc Zen 3 will not mean consumer Zen 3 for a while, and in the meantime they want

- get some more press

- make you pay for the improvements in manufacturing yields and binning: Zen 2 dies no longer needed for Epyc once Zen 3 is ready and refinements in TMSC's process

- maybe some of them can get the top spot of some chart off Intel at those clocks (see #1).

 

Cutting the prices on current Zen 2 CPUs to attack Intel on the value for money front is fine, but profit isn't made in low prices. These "refreshed" CPUs would just be a way to actually keep prices a bit higher.

I mean, why bother making CPUs' architecture match their names, like, I don't know, some Zen 2 APUs maybe? Don't the 3200G and 3400G need refreshing, perhaps? How about some spare Zen 2 dies to accompany those Vega graphics, a nice 3400G XT? They don't even need to keep the Vega side for that matter, they got Zen 2+RDNA2 ready for consoles, right?

No, no, just a 4790K-cough, sorry, I mean, some better binned clones that they would make anyway to meet demand, because it's not like they can get bad silicon on purpose, and because who needs a simple lineup when you can mess it with overlapping names that don't mean overlapping architectures and multiple SKUs at the same core count and 99% the same features.

Did anyone do that in the past? No? Maybe not, it just sounds vaguely familiar...

 

Meanwhile; Intel just announced a simplified lineup, no HT segmentation, no nonsense...

 

Can't we just have a third CPU manufacturer to be the permanent douche of the industry? I like both Intel and AMD better when they are trailing.

 

 

Of course, that's if these rumors hold some truth...

 

 

[GDRRiley]

2 minutes ago, SpaceGhostC2C said:

I mean, why bother making CPUs' architecture match their names, like, I don't know, some Zen 2 APUs maybe? Don't the 3200G and 3400G need refreshing, perhaps?

they will but they go with 4000 products. it will be their laptop chips like 4800h stuck in desktop.

They have always based APUs on last gen silicon because they are repurposed laptop chips. my 3500U is a 3200g with SMT or a 3400g with some GPU removed.

their laptops and APUs are 1 complete die not chiplets.

 

This mostly is a result of higher yields and a maturing node. notice how intel can get a few extra mhz every year out of 14mn that is because it is becoming more mature.

Unofficially AMD is now at 90%+ full working dies up from around 70% at launch.

[SpaceGhostC2C]

3 minutes ago, GDRRiley said:

they will but they go with 4000 products. 

 

I know. 

And that's terrible.

 

[GDRRiley]

4 minutes ago, SpaceGhostC2C said:

I know. 

And that's terrible.

 

it isn't terrible a 4800H in a desktop package would be faster than a 3800x. That is why they don't do it

[SpaceGhostC2C]

Just now, GDRRiley said:

it isn't terrible a 4800H in a desktop package would be faster than a 3800x. That is why they don't do it

No, it is terrible: the only reason to make Zen APUs part of the "2000 series", Zen+ APUs part of the "3000 series", etc is to mislead.

What they are doing: releasing APUs way later than everything else. What they make it look like: APUs are just current gen CPUs with graphics. See the issue?

 

And no, Zen 2 with less cache is not faster than Zen 2 with more cash. Needless to say, neither the 3200G nor the 3400G are 8-core parts, so I don't even know what we're talking about here...

[descendency]

15 minutes ago, GDRRiley said:

they will but they go with 4000 products. it will be their laptop chips like 4800h stuck in desktop.

The 4xxx laptop chips are Zen 2 parts, not Zen 3. I really hate AMD's naming scheme. 

[GDRRiley]

1 minute ago, SpaceGhostC2C said:

No, it is terrible: the only reason to make Zen APUs part of the "2000 series", Zen+ APUs part of the "3000 series", etc is to mislead.

What they are doing: releasing APUs way later than everything else. What they make it look like: APUs are just current gen CPUs with graphics. See the issue?

 

And no, Zen 2 with less cache is not faster than Zen 2 with more cash. Needless to say, neither the 3200G nor the 3400G are 8-core parts, so I don't even know what we're talking about here...

it isn't to misslead. they are mobile parts suck in a desktop socket.

 

they don't have any CCX latency. look at single core scores and look at clock for clock benchmarks.

4000APU are based on 4000 mobile parts.

 

Just now, descendency said:

The 4xxx laptop chips are Zen 2 parts, not Zen 3. I really hate AMD's naming scheme. 

they've done it sense launch

[Bombastinator]

10 minutes ago, GDRRiley said:

it isn't terrible a 4800H in a desktop package would be faster than a 3800x. That is why they don't do it

Not easy to test but testable.  Have to convert from bga which would mean desoldering it off the laptop and putting on pins.  Expensive and time consuming. Not impossible though.

[GDRRiley]

6 minutes ago, Bombastinator said:

Not easy to test but testable.  Have to convert from bga which would mean desoldering it off the laptop and putting on pins.  Expensive and time consuming. Not impossible though.

you can do it by matching clockspeeds.

AMD Ryzen 7 3800X 508 5054 vs 4900hs 492 4104 at a much lower clock

 

your missing the microcode to do it as well.

[Bombastinator]

1 minute ago, GDRRiley said:

you can do it by matching clockspeeds.

AMD Ryzen 7 3800X 508 5054 vs 4900hs 492 4104 at a much lower clock

Estimate it anyway.  “ it could happen” and “it happened” are different things.

[GDRRiley]

4 minutes ago, Bombastinator said:

Estimate it anyway.  “ it could happen” and “it happened” are different things.

amd has always taken laptops chips and moved them over to desktops.

look at codenames

[SpaceGhostC2C]

2 minutes ago, GDRRiley said:

it isn't to misslead. they are mobile parts suck in a desktop socket.

Those are two unconnected statements. Putting an Excavator APU on an AM4 socket didn't make it a Zen CPU; naming it "Ryzen 1100G" would of course have been misleading.

 

2 minutes ago, GDRRiley said:

they don't have any CCX latency.

They also don't have a) the cache, b) 2-12 cores, as the 3000G parts are quad cores, and all I'm talking about is Zen 2 versions of those instead of pointless refreshes. Even if they had some advantage for single threaded lab benchmarks applications, they would still not be a replacement for something like a 3800X.

Oh, but wait, we also just got two quad-core CPUs in the 3000 lineup! ... with no graphics.

Yep, doc, I think we have a case of  cancerous market-segment multiplication here...

 

2 minutes ago, GDRRiley said:

they've done it sense launch

We're not questioning that it's consistent, we're saying that it's been consistently bad since the beginning.

[GDRRiley]

Just now, SpaceGhostC2C said:

Those are two unconnected statements.

 

They also don't have a) the cache, b) 2-12 cores, as the 3000G parts are quad cores, and all I'm talking about is Zen 2 versions of those instead of pointless refreshes. Even if they had some advantage for single threaded lab benchmarks applications, they would still not be a replacement for something like a 3800X.

Oh, but wait, we also just got two quad-core CPUs in the 3000 lineup! ... with no graphics.

Yep, doc, I think we have a case of  cancerous market-segment multiplication here...

 

We're not questioning that it's consistent, we're saying that it's been consistently bad since the beginning.

the point was AMD doesn't do them early because they first want them for laptops.

There are no zen2 versions of those chips. they rebuilt them a renior for 4000 laptops and APUs.

thats all those moible parts are. quad cores. you can have with or without hyper threading a varying level of graphics

the 3100x and 3300x are based on desktop silicon not mobile.

 

[nick name]

There is a guy over at overclock.net that has a 3600 that will run 4.6GHz at 1.325V on all cores.  If he didn't simply win the lottery then it might be the process has matured enough to produce much better results.  So if they're ending up with a lot of top-end silicon then I can see them just starting to use that and differentiating that they are taking advantage of better yields with new marketing materials.  

 

And I suspect that if they're going to be releasing a 3900XT and 3950XT then it might just be that they're putting two good CCDs in them now instead of the one good and one weaker CCD scheme they started with.    

[GDRRiley]

4 minutes ago, nick name said:

 So if they're ending up with a lot of top-end silicon then I can see them just starting to use that and differentiating that they are taking advantage of better yields with new marketing materials. 

that is all this is

[SpaceGhostC2C]

3 minutes ago, GDRRiley said:

the point was AMD doesn't do them early because they first want them for laptops.

Which, by the way, are also named in the wrong generation.

They just launched the "4000" mobile processors, and I can't find any in my market: exact release date aside, we had both "3000 APUs" and "3000 mobile" simultaneously for a while, and they are both Zen+.

It's not so much APU after mobile as it is APU+mobile being one product cycle late - just named as if they weren't.

 

[Bombastinator]

15 minutes ago, GDRRiley said:

amd has always taken laptops chips and moved them over to desktops.

look at codenames

As is “similar things have happened in the past” third thing anyway.

[GDRRiley]

3 minutes ago, SpaceGhostC2C said:

Which, by the way, are also named in the wrong generation.

They just launched the "4000" mobile processors, and I can't find any in my market: exact release date aside, we had both "3000 APUs" and "3000 mobile" simultaneously for a while, and they are both Zen+.

It's not so much APU after mobile as it is APU+mobile being one product cycle late - just named as if they weren't.

 

I can find lots of asus 4000 based laptops here in the US.

they can be simultaneously but it is about launch and the demand.

[GDRRiley]

2 minutes ago, huilun02 said:

So has anyone yet decided what the good news will be for my X470 mobo?

• Status quo and next gen is Zen 3
• Rumors true and it gets Zen 2+ as well as Zen 3

these aren't zen2+ these are just better binned zen2 which may mean no bios update needed

[ravenshrike]

Could just be that between process refinement at TSMC, stepping improvements, and microcode changes there was a large enough improvement overall that they could advertise better performance. At which point there's no reason not to do so and sucker punch intel's release, setting them up for even further humiliation when zen 3 launches.