RTG already received first Zen 2 Sample! (Rumour)

[cj09beira]

52 minutes ago, GoldenLag said:

Zen 2 is a bigger difference than what Zen+ was. It is the first actual iteration of Zen. There is probably quite some improvements, but its far from a ground up design. It still builds very much on Zen

 

Well, this is an engineering sample most likely with no XFR. Key word engeneering sample. Though it might not look all that impressive now it is still 3-5 months untill Rome launch and 6 months untill Ryzen 3000. Max XFR potential is probably not avavible on that chip. Im expecting 20% up from 2700x (4,3ghz) and that will be about 5,1ghz. Offcourse plus minus widening of the architecture and IPC.

sorry but the probability of this new cpus reaching 5Ghz is very low, i dont expect clocks higher than 4.7ghz

[AlTech]

1 minute ago, Mihle said:

I think I ment 12nm and not 14. Remember that is after Epyc is on 7nm.

 

Didn't even notice the lack of 4 core 12nm Ryzen, those might be in gen 3?

They exist.  R3 2300X and R5 2500X

[Mihle]

14 minutes ago, PianoPlayer88Key said:

 

Some things I'd like to see for Zen 2 / Ryzen 3000:

• Ryzen 7: 16c/32t, CPU

 

Just won't happen, the best you will possibly get is 12c/24t

[GoldenLag]

1 minute ago, cj09beira said:

sorry but the probability of this new cpus reaching 5Ghz is very low, i dont expect clocks higher than 4.7ghz

You are probably right, but im putting 20% on the table. If AMD cant meet that in singlethreaded then im going to be unimpressed. 

 

Im putting a high bar, but its what AMD should strive for.

 

3 minutes ago, AluminiumTech said:

They exist.  R3 2300X and R5 2500X

True, but i really havent seen them for sale anywhere at all. Not all tgat surprising as what do these chips offer compared to cheaper older offerings like the 1200 and 1600

[Mihle]

3 minutes ago, AluminiumTech said:

They exist.  R3 2300X and R5 2500X

I haven't really followed the lowest end stuff so I don't know, you and the other person say different stuff so let me Google.

[GoldenLag]

3 minutes ago, Mihle said:

Just won't happen, the best you will possibly get is 12c/24t

I was genuinly hoping for 6core ccx all year, but as the time comes a smaller 4 core ccx seems way more likely as AMD wants as small chiplets as possible for 7nm

[cj09beira]

3 minutes ago, Mihle said:

Just won't happen, the best you will possibly get is 12c/24t

its not impossible they could release a 2 die cpu though the probability is very low 

 

1 minute ago, GoldenLag said:

You are probably right, but im putting 20% on the table. If AMD cant meet that in singlethreaded then im going to be unimpressed. 

 

Im putting a high bar, but its what AMD should strive for.

 

True, but i really havent seen them for sale anywhere at all. Not all tgat surprising as what do these chips offer compared to cheaper older offerings like the 1200 and 1600

problem is that they will most likely be limited to the node capabilities, and i don't have much faith in tsmc's node, hoping to be wrong though

 

Just now, GoldenLag said:

I was genuinly hoping for 6core ccx all year, but as the time comes a smaller 4 core ccx seems way more likely as AMD wants as small chiplets as possible for 7nm

ya it seems they have gone with more dies to increase the advantage of dividing the die even more (for the epyc cpus), this for us consumers isn't as good as a 12 core die though

[dizmo]

2 hours ago, sazrocks said:

I’m a bit confused here... I thought you can get 3000MHz working with pretty much any memory manufacturer (assuming the stick can actually do it of course), especially with ryzen 2? Much above 3200 is pretty useless too (except for very specific tasks)

 

On topic:

8c/16t? I thought we were going to see at least 12c/24t, especially since we’ve been told that AMD’s 7nm is about half the size. Hopefully this is just a cut down model (3600?).

@Dan Castellaneta basically said exactly what I was going to in the post below your reply. So I'll leave that at that.

 

I think 12c/24t in Ryzen is a little over the top. It's not needed. They should keep those SKUs in the TR platform for at least a couple generations.

 

1 hour ago, AluminiumTech said:

No?

 

Vega 56 usually destroys or matches the 1070. 1070Ti easily beats Vega 56 but that's not the 1070.

Really? Hardware Canucks shows otherwise.

The gap between the 1070Ti would only be much larger.

 

[Dabombinable]

Anyone else getting Pentium III Coppermine 1100/1133 flash backs?

[Brooksie359]

1 hour ago, GoldenLag said:

You are probably right, but im putting 20% on the table. If AMD cant meet that in singlethreaded then im going to be unimpressed. 

 

Im putting a high bar, but its what AMD should strive for.

 

True, but i really havent seen them for sale anywhere at all. Not all tgat surprising as what do these chips offer compared to cheaper older offerings like the 1200 and 1600

You are that parent that is upset when their kids get a 3.9 instead of 4.0. 5ghz is not that easy to hit and I wouldn't be expecting it as you will only be disappointed. If they do it 5ghz I would be overjoyed but I also don't expect them to do such a big leap in one generation. 

[Stefan Payne]

3 hours ago, cj09beira said:

sorry but the probability of this new cpus reaching 5Ghz is very low, i dont expect clocks higher than 4.7ghz

I don't know.

We are at 4,3GHz right now with an improved 14nm (12nm), 5GHz would allow around 15% higher clockrates.

From 4.3GHz to 5 is only 16%, so its not that much...

 

4.7GHz would be only 9% more. Ähm, no...

That's highly improbable.

 

3 hours ago, GoldenLag said:

You are probably right, but im putting 20% on the table. If AMD cant meet that in singlethreaded then im going to be unimpressed. 

No, 20% would be 860MHz more -> 5.2GHz

5GHz are roughly 15% more....

 

You are that parent that is upset when their kids get a 3.9 instead of 4.0. 5ghz is not that easy to hit and I wouldn't be expecting it as you will only be disappointed. If they do it 5ghz I would be overjoyed but I also don't expect them to do such a big leap in one generation. 

Well, we are talking about around +15% higher clockrates with the switch from 12/14nm to 7nm...

That doesn't seem that high, relatively speaking...

 

A long long time ago, you could overclock a Celeron by 50%...

But today, we are happy with 10%...

[D13H4RD]

Definitely sounds like an engineering sample.

 

Though because it’s WCCF...

[MMKing]

4 minutes ago, Stefan Payne said:

I don't know.

We are at 4,3GHz right now with an improved 14nm (12nm), 5GHz would allow around 15% higher clockrates.

From 4.3GHz to 5 is only 16%, so its not that much...

 

4.7GHz would be only 9% more. Ähm, no...

That's highly improbable.

The Ryzen 2700x will NOT boost all cores to 4.3ghz on it's own. I am unfamiliar with the specifics, but without manual overclocking, all 8 cores will only boost to around 4 to 4.1GHZ. The 8700K will boost all of it's cores to 4.3GHZ (according to Intel).

 

Going from 4.0 - 4.1 to 5GHZ would entail an increase of 22-25%. Going forward on rumors alone, it's difficult to guess with any real amount of certainty as to how high the Zen 2 architecture will clock. Which is made more difficult due to the fact that Zen 2 will be on TSMC 7nm, not on Globalfoundries 12nm. New process, and a new foundry. In any case, a 15% IPC gain (rumored) compared to the 3% IPC gain going from Zen to Zen+ is very impressive.

 

The current sample clocked at 4GHZ all cores base is already promising, and if we see 4.5GHZ base all cores i would be very impressed.

[GoldenLag]

4 hours ago, MMKing said:

The Ryzen 2700x will NOT boost all cores to 4.3ghz on it's own. I am unfamiliar with the specifics, but without manual overclocking, all 8 cores will only boost to around 4 to 4.1GHZ. The 8700K will boost all of it's cores to 4.3GHZ (according to Intel).

 

Going from 4.0 - 4.1 to 5GHZ would entail an increase of 22-25%

Actually the 2700x can do 4,2 ghz all core doing a relativly easy OC. 

 

Also is 20% all that much to expect from 1. Node change to something significantly denser. 2. 2+ years of architectural improvement over the base Zen design?

[asus killer]

8 hours ago, cj09beira said:

sorry but the probability of this new cpus reaching 5Ghz is very low, i dont expect clocks higher than 4.7ghz

why? isn't this 7Nm, a huge jump from 14 or 12Nm of previous generations. I really don't trust either way this rumors. Wait and see.

[jagdtigger]

9 hours ago, RejZoR said:

I don't get it why people continue to parrot the same IPC bullshit about Zen being the same as Haswell. Intel hasn't really improved IPC at all for ages, the only real gains they make between generations is from higher clocks, not from IPC itself.

Yep, based on cinebench there is only 300 MHz difference between the 4th and the 8th gen if i want to match the 8th gens per core performance(not surprising, after all its just a 7th gen design but with more cores).

 

/ON

Lets hope its true, but even then the DRAM prices gonna crash the party .

Edited October 1, 2018 by jagdtigger

[RejZoR]

18 minutes ago, asus killer said:

why? isn't this 7Nm, a huge jump from 14 or 12Nm of previous generations. I really don't trust either way this rumors. Wait and see.

The thing is, you can't just slap a new smaller node process and just clock the thing to infinity. The chip needs to be designed almost from ground up to clock high. If it's not, it won't clock even if you make it using 1nm process. You do automatically gain some headroom because you have lower heat and lower leakage, but you have to design chip's pipeline in such a way to allow chip to clock high. And that usually means using narrower and longer pipeline. And even that doesn't mean a guaranteed success. Pentium 4 had very high clocks using very long pipeline and they weren't any better than AthlonXP CPU's clocked way lower.

[MMKing]

1 hour ago, GoldenLag said:

Actually the 2700x can do 4,2 ghz all core doing a relativly easy OC. 

 

Also is 20% all that much to expect from 1. Node change to something significantly denser. 2. 2+ years of architectural improvement over the base Zen design?

You know you're not countering my point by stating that the 2700x can reach 4.2 with a relatively* easy OC. When my original point was that exceeding 4.1ghz all cores required manual overclocking.

 

 

I'm not particularly interested in how well Intel or AMD CPUs can overclock with expensive motherboards and cooling solutions. What matters is what guarantee AMD or Intel is willing to put on their product, with cooling NOT exceeding the rated TDP (Thermal Design Power). I'm very much against the press overclocking the product, given the premium components and cooling solutions they usually utilize as well as the issue of Intel and/or AMD cherry picking CPUs for the media.

 

As it stands. AMD guarantee that the 2700x will hit 3.7GHZ on all cores, with a rated TDP of 105W. It is expected that no matter how cheap the motherboard, as long as it is supported. And as long as the cooling is able to dissipate 105W, the CPU is guaranteed to hit 3.7GHZ. If AMD is willing to guarantee 5GHZ all cores on their Zen 2 CPUs, that is fantastic news and extremely impressive.

 

https://community.amd.com/community/gaming/blog/2017/11/27/asdasd

''Precision Boost allows an AMD Ryzen processor to dramatically increase clockspeeds when the CPU has electrical, thermal, and/or utilization headroom to spare.''

 

https://community.amd.com/community/gaming/blog/2018/08/13/understanding-precision-boost-overdrive-in-three-easy-steps

''Precision Boost Overdrive requires a 2nd Gen AMD Ryzen™ Threadripper processor with AMD X399 chipset motherboard. Because Precision Boost Overdrive enables operation of the processor outside of specifications and in excess of factory settings, use of the feature invalidates the AMD product warranty and may also void warranties offered by the system manufacturer or retailer.''

 

https://www.intel.com/content/www/us/en/products/docs/processors/core/8th-gen-core-family-datasheet-vol-1.html

''Intel Turbo Boost Technology 2.0 will increase the ratio of application power towards TDP and also allows to increase power above TDP as PL2 for short periods of time. Thus, thermal solutions and platform cooling that are designed to less than thermal design guidance might experience thermal and performance issues since more applications will tend to run at the maximum power limit for significant periods of time.''

 

 

I'm not swallowing marketing hogwash that try to pass off a 28 core running 5GHZ as normal product behavior, while they accidentally forget to mention certain features the end user may not have. The boost technologies are just differences in degree. AMD and Intel are not willing to guarantee that their products will perform at ''boost'' speeds, with ANY supported motherboards while being cooled using any cooler matching or exceeding the manufacturers rated TDP. The boost speeds represent the absolute maximum clock speed the CPU will clock itself to, provided the end user provides enough cooling, power and possibly an unreasonably expensive motherboard.

 

In the case of the 28 core farse. Intel used a 600USD cooler and a motherboard with 4 * 8 pin power connectors and what looks to be a 29 phase power design. Though the power phase design is more difficult to judge without a close inspection of the motherboard and the parts.

 

The 2700x is a 3.7GHZ 8 core 16 thread CPU, with a boost technology that MAY allow 4.3GHZ. Zen 2 8 core 16 thread CPU hitting 5GHZ would be a 35% improvement over the current flagship 8 core 16 thread. A 20% increase would be 4.44GHZ, and the rumored engineering sample of 4GHZ is a 10.8% improvement.

 

Hardware unboxed released a video a couple weeks ago, testing the 2990WX on 4 different motherboards.

 

Even using a closed loop water cooler. The Asrock Fatality X399 is NOT able to maintain a 4GHZ clock at 1.25v, bellow the AMD stated boost clock of 4.2GHZ. 

[GoldenLag]

13 minutes ago, MMKing said:

-snip-

Yes, intel or AMD only guarantee certain spec. Wich is why boost tech is involved at all. 

 

Lets take a look at Zen. Its running on a 14nm node that has a target of 3ghz. Zen + is a copy-paste Zen at 12nm with slight cache change that improved IPC. 

 

Practically there has been no development between Zen and Zen+. 

 

Zen 2 on the other hand is a architectural iteration on Zen. There are actual changes to speak of. 

 

We also have the Node change that is actually major. 14nm to 12nm resulted in no dieshrink and was to all intents and purposes a 14nm node with slightly better performance.

 

7nm is aimed at 5ghz. Its density is much better and powerdraw is down. 

 

Also, show a mobo with XFR2 and a CPU with its stock cooler that doesnt meet its boost. Im actually very interested in knowing what mobo that would be.

 

Also the 2700x is actually guaranteed to hit X clockspeed at Y voltage hence the implementation of XFR2. Otherwise it would be an 2700. 

 

Also im not talking 5ghz all core. Im just expecting them to hit somewhere near what the node is actually rated for. 

[yian88]

14 hours ago, Emberstone said:

Due to the constraints of the Infinity Fabric (e.g. latency), even at 8c/16t and 4.5 GHz it probably won't be overtaking the likes of the 8600K/8700K in gaming, but in productivity it'll kick ass. I'm excited to see what becomes of this.

 

That said, if there's a supposed "Ryzen 5 3600" that sells for $189 or similar with 6c/12t and 8600K/8700K clock speeds, that'll be absolutely tasty for new system builders considering Intel's prices for similar speeds and threads.

It wont overtake, but it will be neck and neck, even a tiny 5-10% IPC improvement + 4.5ghz is enough to make Ryzen equal to i7/i9 in games productivity.

If they make infinity fabric improvement to reduce latency then they will definately be equal in performance.

I hope they do all above and provide more cores for AM4, with 7nm if they cant bring 10-12 cores Ryzen thats just sad, they would have to seriously improve IPC like 15-20% along 4.5-4.8ghz OC potential, else intel is beating them with old 14nm for years.

[MMKing]

@GoldenLag

I'm just trying to put boost technologies and overclocking into a more grounded perspective.

 

Also what do you mean about the 2700 and the 2700x? The 2700 has a stock clock, 500mhz lower than the 2700x and an advertised boost clock, 200mhz lower while rated at 65W TDP compared to 105W of the 2700x. Both CPUs support SenseMI, which is AMDs overall brand for boost technologies.

 

The Ryzen 7 2700 product page even directly links to AMDs XFR 2 video

[Taf the Ghost]

12 hours ago, GoldenLag said:

I was genuinly hoping for 6core ccx all year, but as the time comes a smaller 4 core ccx seems way more likely as AMD wants as small chiplets as possible for 7nm

The "Everything is Zen" approach is actually something of a layering of 4-way connection parts. Each layer of interaction will stay in the 4-way alignments because it's the most efficient, mathematically, for direct connections. This is why even the controller die discussions are really just setting up 4-way, socket to socket connections.

 

In other words, we're not seeing a move away from the 4-core CCX for a while. Though once they're up to 4 CCX per die, things might get more interesting. Zen is best viewed as an advanced HyperTransport with a GenZ approach layered over a completely different implementation of Bulldozer's CMT design. We really should look at each CCX as a "core", from a design perspective. Similar to a SMT8 design from IBM, however it's built up from smaller, efficient traditional cores. This approach will play out better as we get up to the 5nm & 3nm generations. 

 

From AMD's research papers (that AdoredTV found), it looks like 4x4 CCX will be the final building block for server parts. That'll be the new CCX structure, and then they go full Chiplet with it.

[GoldenLag]

1 hour ago, MMKing said:

 Also what do you mean about the 2700 and the 2700x? The 2700 has a stock clock, 500mhz lower than the 2700x and an advertised boost clock, 200mhz lower while rated at 65W TDP compared to 105W of the 2700x. Both CPUs support SenseMI, which is AMDs overall brand for boost technologies.

Whenna CPU doesnt meet the requirement of a higher end product it gets downgraded to the tier that suits it. 

 

The 2700 might qualify for XFR2, but i doesnt hold up to the standard of the 2700x and gets a different clock and voltage at boost.

[ThePD]

Very interested to see what AMD does with Zen 2 Threadripper models. Will be a very welcomed performance improvement over my i7-5820k.

[Taf the Ghost]

3 minutes ago, ThePD said:

Very interested to see what AMD does with Zen 2 Threadripper models. Will be a very welcomed performance improvement over my i7-5820k.

We'll know a lot more when Rome launches and how AMD is putting those together. The 8+1 chiplet/controller approach has been too much in the rumor mill to not have weight to it, so it's really a question if they bring the controller-die version to TR3. They could just repeat the X399 approach without too much issue, as we expect Zen2 will have AVX2 units, which means there's upwards of ~25-30% IPC improvement in a bunch of "professional" type tasks on the table.

 

The real interesting thing is going to be the PCIe lanes. It's all but confirmed that Rome comes with PCIe 4.0.