AMD Zen 2 (Ryzen 3000) - *** NO COMPETITOR HINTING ***

[DragonQ]

4300 the value itself isn't even close to the absolute records achieved with 2990WX.
The more important metric is the 14% IPC increase. This will put Ryzen 3000 on top of the IPC charts with ~estimated 168 CB points @3.5GHz.

What's the all-core boost of the 2950X? 4278 is a lot higher than 3210, 33% in fact. I imagine 4.2 GHz is faster than the 2950X would be running at in Cinebench but not by much. EDIT: I think it's 4.1 GHz?

We need to consider though that we haven't seen anything about single threaded performance yet. It'll be helped by clock speeds regardless but everything we've seen could be just due to improvements in SMT or inter-core communication. Not saying that's the case but we haven't got enough evidence yet.

 

[Stephanie Peterson]

It's the difference between 5W and 15W on the chipset - it won't impact the CPU in any meaningful way.

Its unlikely to have any affect on the cpu frequency i would have thought as both are separated by several inches.

 

[Stephanie Peterson]

Source?

Adoreds latest video.
While many dont trust or like him i find his stuff to be close enough - its prob all nearly true.

 

[EsaT]

It's the difference between 5W and 15W on the chipset - it won't impact the CPU in any meaningful way.

Especially compared to all those non-reference cooler graphics cards dumping anything up to 300W of heat into case to warm it.

If its not 5Ghz with Intel beating IPC with 16 cores then I'll be disappointed given the absolute hype that adored is attempting lol.

If I had Intel Skylake Rebrand x+y+z I would be already disappointed to constantly growing list of vulnerabilities.
With Intel's resources and time they've had when AMD was out of competition those old ****** design choises still being there are inexcusable.

 

[MassiveJim]

Adoreds latest video.
While many dont trust or like him i find his stuff to be close enough - its prob all nearly true.

Just watched the bit about the cinebench score.

What is his track record like for accuracy in his Leaks?

 

[Jokester]

Its unlikely to have any affect on the cpu frequency i would have thought as both are separated by several inches.

You would be maxing out the bus into the CPU end as well, but the load at that end is negligible in terms of the CPUs thermal envelope. Either way it would only maybe have an impact if the CPU was also at or near full load simultaneously.

 

[Stephanie Peterson]

What is his track record like for accuracy in his Leaks?

Better than most Youtube commentators.

Bar a few minor things i think we are looking at the real Zen2 launch setup.
The one i have a big "?" over is the high closed 16 core part he has listed.
While i dont doubt on seeing a 16 core AM4 chip i dont know about the 5.1Ghz boost speed its claimed to do - ok the 12 core 5 or 5.1 boost on say 2 cores yeah... and 4.4 on all cores ... yeah..

But 5.1??? naa not out the box... well i think not, maybe 52% no / 48% yes...

 

[Jono8]

So what do we reckon? Is this finally going to be the release that completely changes the CPU market? A 16 core 32 thread 5ghz CPU for under $500 with better ipc than Intel? That will obliterate anything Intel have right now.

 

[Martini1991]

So what do we reckon? Is this finally going to be the release that completely changes the CPU market? A 16 core 32 thread 5ghz CPU for under $500 with better ipc than Intel? That will obliterate anything Intel have right now.

Sounds too good to be true.
But I'd buy it.

 

[beany_bot]

So what do we reckon? Is this finally going to be the release that completely changes the CPU market? A 16 core 32 thread 5ghz CPU for under $500 with better ipc than Intel? That will obliterate anything Intel have right now.

I don't think so. I think we will get something that is equal to Intel in gaming. (yes including 240p space invaders on a RTX Titan). With more cores (better in productivity) for less money. Certainly making it a great upgrade for me. But not utterly game changing like Athlon 64.

 

[Sargatanas2511]

Better than most Youtube commentators.

Bar a few minor things i think we are looking at the real Zen2 launch setup.
The one i have a big "?" over is the high closed 16 core part he has listed.
While i dont doubt on seeing a 16 core AM4 chip i dont know about the 5.1Ghz boost speed its claimed to do - ok the 12 core 5 or 5.1 boost on say 2 cores yeah... and 4.4 on all cores ... yeah..

But 5.1??? naa not out the box... well i think not, maybe 52% no / 48% yes...

It's very unlikely I agree. I am staking my money on 4.7 boost for a 16 core part, any more than that and it'll only make me happier.

I will be upgrading regardless, but clocks will decide between a 12 and 16 core cpu.

 

[Stephanie Peterson]

I do think so, we are on the cusp (potentially) of a radical change in CPU building not seen since the day of 64 bit (AMD) and multicore cpu's (AMD)

Building a CPU out of discrete multicore modules and packaging them on a carrier small enough to fit in a socket that is not oversized and cheep enough for consumers is huge imho.
Building that cpu with its memory controller and other potential interesting things like a graphics core as a chiplet or bigger cache on the IO die is even more interesting.
Building that whole setup with a class leading IPC and frequency along with a brand new 7nm node... well its a biggie.

Its the start of a new thing for computer processors, 10 years time i bet most if not all PCs CPUs will be built like this. Arm stuff might follow as well.

What will we have in 10 years with this kind of build? 128Core 512 thread 6 or 7 Ghz (lol!) cpu with 4 chiplets, an integrated graphics core with a couple of thousand processors as well as a few hundred Mbit cache on an io chiplet with vast memory bandwidth.

Meanwhile Intel readys Skylake mk23 on 14Nm +∞ which can run GS:go at 500fps/720P low settings.

 

[smilingcrow]

NVMe Raid, apparently PCIe 4 is causing the heat when fast NVMe drives are running in Raid 0.

Did he actually test this with PCIe 3 drives or is it just speculation?
Unless the chipset supports hardware RAID isn't most of the work done by the CPU anyway?
It sounds as he's speculating so dual PCIe 3 SSDs might not be an issue which is the most the majority would run and even then rarely both from the chipset if you hang one off the CPU and the other off the chipset.
As for PCIe 4 SSDs, sure that might be a bigger deal but pure speculation.

 

[4K8KW10]

From a business perspective. Why would you release a chip that utterly destroys your competition with massive leaps in performance when you can release one that just beats the competition and leaves you with plenty more on the back shelf to release later. I'm just not expecting anything "epic" I expect "enough".

Because you have very weak market penetration, maybe?!

Sounds too good to be true.
But I'd buy it.

5.0GHz isn't a target and shouldn't be one's target and desire.

But it's just what it is.

Prepare for ground breaking performance, as per the AMD's own slide: https://videocardz.com/80720/sapphire-preparing-two-radeon-navi-models-for-july-7th-launch

 

[humbug]

Because you have very weak market penetration, maybe?!



5.0GHz isn't a target and shouldn't be one's target and desire.

But it's just what it is.

Prepare for ground breaking performance, as per the AMD's own slide: https://videocardz.com/80720/sapphire-preparing-two-radeon-navi-models-for-july-7th-launch

So AMD are on thier own hype train, Good Grief i hope all this pans out or at least close to this hype or the crockery will start flying....

 

[Martini1991]

Q3 2019, ew. But that's what I did see at the Stone HQ.

Hopefully it means July.
I just want to buy it God damn it.

 

[humbug]

Q3 starts in 2 weeks, its probably July.

 

[Journey]

Q3 starts in 2 weeks, its probably July.

6 weeks, not two weeks.

 

[humbug]

6 weeks, not two weeks.

You're right, 1'st July.

 

[Martini1991]

Q3 starts in 2 weeks, its probably July.

Q3 starts with July.

AMD need to get this launch all right with the later than anticipated launch.