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

[LePhuronn]

Go 7nm Vega Instinct officially gets moved to this year, which still leaves at least 10 months until Navi is supposed to land. I wonder what this Vega 12 is then...

Still, good to see 7nm is working nicely enough to get the CPUs rolling too. Can't wait to see the destruction that 7nm EPYC is going to do.

 

[Panos]

Go 7nm Vega Instinct officially gets moved to this year, which still leaves at least 10 months until Navi is supposed to land. I wonder what this Vega 12 is then...

Still, good to see 7nm is working nicely enough to get the CPUs rolling too. Can't wait to see the destruction that 7nm EPYC is going to do.

Who knows if we going to see 7nm Vega 20 on PC But imho not in hurry.
Vega 12 could be the IGP Vega, yet the AMD software director said that there is going something big coming at the end of the year but left it open.
So who knows what we going to see. Personally I believe the is the damn promise from January for HDMI 2.1 (incl VRR) support via driver update, on all RX cards.

 

[4K8KW10]

Who knows if we going to see 7nm Vega 20 on PC But imho not in hurry.
Vega 12 could be the IGP Vega, yet the AMD software director said that there is going something big coming at the end of the year but left it open.
So who knows what we going to see. Personally I believe the is the damn promise from January for HDMI 2.1 (incl VRR) support via driver update, on all RX cards.

Vega 12 could very well be some type of chip in the sweet spot regarding die sizes, guess around 250 sq. mm and for the discrete graphics card market.

 

[Panos]

Vega 12 could very well be some type of chip in the sweet spot regarding die sizes, guess around 250 sq. mm and for the discrete graphics card market.

If I remember correctly Vega 12 was the rumoured Polaris replacement.

 

[LePhuronn]

If I remember correctly Vega 12 was the rumoured Polaris replacement.

OK, so what was Polaris 30 that was also rumoured alongside Vega 12? I guess we'll have to wait another 5 months and not derail a CPU thread any further

 

[4K8KW10]

Apparently AMD Rome scores and picture of the chip leaked:

https://wccftech.com/amd-epyc-rome-7nm-64-core-cpu-performance-benchmark-leak/

Astonishingly good
Over 12.5 thousand SB R15 points is impressive.

 

[humbug]

Astonishingly good
Over 12.5 thousand SB R15 points is impressive.

25% faster than 4x Xeon Platinum 8160

https://hwbot.org/benchmark/cinebench_-_r15/halloffame

Let me repeat that, this CPU is 25% faster than four Xeon Platinum 8160's :O

 

[AMDPower]

25% faster than 4x Xeon Platinum 8160

https://hwbot.org/benchmark/cinebench_-_r15/halloffame

Let me repeat that, this CPU is 25% faster than four Xeon Platinum 8160's :O

That is just insane amount of cpu power

any idea of price? This surely must be upward of £3000

 

[NinjaCool]

AMD is taking things to a whole new level if that score is true!

 

[kitfit1]

25% faster than 4x Xeon Platinum 8160

https://hwbot.org/benchmark/cinebench_-_r15/halloffame

Let me repeat that, this CPU is 25% faster than four Xeon Platinum 8160's :O

Don't forget, each 8160 costs about £4250 (that bench cost the guy £17000)

 

[NinjaCool]

Maybe AMD was right after all!

 

[humbug]

Maybe AMD was right after all!

I like this too....

(Middle panel)

 

[4K8KW10]

25% faster than 4x Xeon Platinum 8160

https://hwbot.org/benchmark/cinebench_-_r15/halloffame

Let me repeat that, this CPU is 25% faster than four Xeon Platinum 8160's :O

64 cores vs 4x24 cores which are 50% more, but running at low 2.1 GHz.
Must be a good sign that the 7nm TSMC process delivers high frequencies.

This Engineering sample run at 3.5 GHz?

 

[humbug]

64 cores vs 4x24 cores which are 50% more, but running at low 2.1 GHz.
Must be a good sign that the 7nm TSMC process delivers high frequencies.

This Engineering sample run at 3.5 GHz?

Could also be IPC, these are Zen 2.

 

[Beren]

It is a cool leak, but does anyone know more than the basic chiphell screenshot? For all we know this could be a two socket variant.

It does suggest good IPC gains though, if this is real it is pretty big news.

 

[LePhuronn]

Holy cow that poster is brutal. AMD better be right this time, although I think they can be more confident and brash with CPUs given the tangible successes so far...

 

[Boot_Loop]

Very exciting to see these same 7nm chips eventually filter down to Threadripper and Ryzen 3000-series.

For the Ryzen 3700X I'm hoping for a boost clock of 4.8Ghz. That, coupled with increased IPC as was rumoured, would see it absolutely wipe the floor with the upcoming Intel 9900K, which will likely look like a hot, power hungry and inefficient relic compared to it, being as that is on the old 14nm+++++ arch.

 

[NinjaCool]

AMD have also said they will be an early adapter of PCIe4, so that makes it even more compelling to wait for Zen 2 if you are doing a new PC build.

 

[LePhuronn]

AMD have also said they will be an early adapter of PCIe4, so that makes it even more compelling to wait for Zen 2 if you are doing a new PC build.

That would probably depend on Navi and Nvidia's 7nm cards using it too. If AMD go all-in and release Ryzen 3000, Navi and 500-series motherboards about the same time then they can push a unified PCI-E 4 adoption and all their kit will work harmoniously as one, otherwise holding out for more bandwidth than you'd realistically use might be holding out for too long.

Of course, if Ryzen 3000 does use PCI-E 4 I don't see any reason why it couldn't work like Sandy Bridge and Ivy Bridge did with the move to PCI-E 3.

 

[4K8KW10]

AMD have also said they will be an early adapter of PCIe4, so that makes it even more compelling to wait for Zen 2 if you are doing a new PC build.

Adopter.
But one can simply sell the X470/370 motherboard and buy a new X570, using it with either generation of processor. Remember that the PCI-E lanes are also in the chipset.