Official AMD Ryzen Benchmarks, Reviews, Prices, and Discussion

[raghu78]

Ok, thanks for that.

But you don't know as a fact that AMD went from HD (78) to HP (84) with PR?

No. Its not a fact but my deduction. We will only know when PR launches.

 

[rainy]

Isn't that for Raven only?

Yes, it is for Raven Ridge, Pinnacle Ridge would be in April.

 

[formulav8]

Isn't that for Raven only?

My bad, Sorry. Got RR and + mixed up. R+ is April.

 

[french toast]

We can confirm that Zeppelin used 14LPP High density libraries (CPP=78nm, 9T) from the ISSCC Zen core specification (CPP=78nm, MMP=64nm)

https://www.pcper.com/news/Processors/AMD-Details-Zen-ISSCC

14LPP High performance (9T) and Ultra high performance (10.5T) both use relaxed CPP of 84nm.

https://pc.watch.impress.co.jp/img/pcw/docs/733/713/html/9.jpg.html

btw GF confirmed support of the performance option (which i guess is CPP=84nm) with all of their libraries at 12LP. My guess to the 12LP libraries that PR is built using (MMP=56nm CPP=84nm, 10.5T). I believe 12FDX and 12LP are not arbitrarily named 12nm. They share the same MMP. We already know 12FDX specs

https://www.eetimes.com/document.asp?doc_id=1332328
https://m.eet.com/content/images/eetimes/1 7 12 14 copared x 800_1505972923.jpg

GF has stated 12LP is an optical shrink to Daniel Nenni of semiwiki. All these pieces of information allow me to guess with quite a bit of confidence.

I swear you and fragman are one and the same person lol.
I hope you are right, but I doubt it.

 

[amd6502]

Hey guys, does anybody have a B350 board and ryzen? If so, is CPB (core boost freq) overclocking possible. The majority of the OCing seems to be multiplier overclocking. I'm not interested in multiplier overclocking, but would like higher boost frequencies. It says Ryzen 1600 only has XFR of 3700. Has anyone gotten it going to 3900 or higher, for CPB overclocking---not multiplier overclocking?

I prefer the 1600's lower 65W under full load over the 1600x's 95W profile and also like its cheaper price.

 

[Asterox]

Athlon ride again, 200GE Desktop 2/4 APU+Vega GPU or "by price=iGPU performance Pentium 2/4 killer".

https://overclock3d.net/news/cpu_ma...ega_graphics_appears_on_sisoftware_database/1

 

[LTC8K6]

Athlon ride again, 200GE Desktop 2/4 APU+Vega GPU or "by price=iGPU performance Pentium 2/4 killer".

https://overclock3d.net/news/cpu_ma...ega_graphics_appears_on_sisoftware_database/1

Not sure how much it matters, but the 200GE seems to be the Banded Kestrel embedded chip from what I can gather on the interwebs.

 

[Asterox]

Not sure how much it matters, but the 200GE seems to be the Banded Kestrel embedded chip from what I can gather on the interwebs.

Well, embedded or not embedded it is Desktop APU.If will be only embedded version no problem, or like this Mini-iTX motherboards "with first AMD APU E-350".

 

[Shivansps]

Not sure how much it matters, but the 200GE seems to be the Banded Kestrel embedded chip from what I can gather on the interwebs.

About that, the results seems too high to be a small core like BK with 2/4... That thing even lives? I would have guess AMD cancelled it after the failure of AM1 in general.

 

[dark zero]

And soon.. Expecting those

Sempron 2C/2T 3.4 Ghz and Vega Graphics Dual Core. Perfect office desktop.

And...

Duron 1C/2T at 3.8 Ghz and Vega Single Core. Ultimate HTPC chip.

 

[DisEnchantment]

Ryzen 3800+ XP, 16C/32T 4.6 Base/5.4 XFR
Probably 16/32 would be BW starved with only two channels.

Athlon Thunderbird... My first CPU. I think the first mainstream x86 CPU at 1GHz+

 

[Roger Wilco]

Ryzen 3800+ XP, 16C/32T 4.6 Base/5.4 XFR
Probably 16/32 would be BW starved with only two channels.

Athlon Thunderbird... My first CPU. I think the first mainstream x86 CPU at 1GHz+

Wait, wut?

 

[Drazick]

Anyone tried ThreadRipper with MATLAB and compared to Intel?

MATLAB relies heavily on Intel MKL and Intel IPP and I wonder if those packages discriminate AMD Zen based CPU's.

I really wonder how ThreadRipper fits the Scientific Programming world.

Remark:
This also holds for any Python distribution utilizing Intel MKL (For instance, Anaconda) or R (I think Microsoft distribution uses Intel MKL).

@The Stilt , Any knowledge on that?

 

[VirtualLarry]

Ryzen 3800+ XP, 16C/32T 4.6 Base/5.4 XFR
Probably 16/32 would be BW starved with only two channels.

Pretty sure that's just random speculation, at this point.

Looking forward to Ryzen+ 2000-series CPUs, AND APUs, soon.

 

[Abwx]

Well, embedded or not embedded it is Desktop APU.If will be only embedded version no problem, or like this Mini-iTX motherboards "with first AMD APU E-350".

https://www.computerbase.de/2018-02/amd-ryzen-embedded-v1000/

 

[DisEnchantment]

Pretty sure that's just random speculation, at this point.

Looking forward to Ryzen+ 2000-series CPUs, AND APUs, soon.

I was just going with the Athlon XP series , while we are at the Duron, Sempron, Opteron thingy, hence the reference to my first CPU, Athlon Thunderbird 1.4 GHz. It ran hot and fast and with such a tiny fan, and I recollect reapplying paste once a while directly on the die.
If you recollect there were debates about the AMD naming schemes to match equivalent intel offerings.
Athlon XP 3200+ etc etc

 

[Shivansps]

And soon.. Expecting those

Sempron 2C/2T 3.4 Ghz and Vega Graphics Dual Core. Perfect office desktop.

And...

Duron 1C/2T at 3.8 Ghz and Vega Single Core. Ultimate HTPC chip.

joking aside, they may rebrand the A6-9500 as a Sempron.

 

[Topweasel]

joking aside, they may rebrand the A6-9500 as a Sempron.

Doubtful. Just in the sense that AMD's roadmap has a Zen alternative for just about all BD based CPU solutions and eventually Zen based replacement for Cat core solutions as well. Chances are that they let the model sit in retail till they have a Zen based CPU that replaces it. I could see them bring back Sempron though for some solutions. But outside Athlon that still carries some cache I don't know if AMD is big on looking back like that. I would like to see a Rebirth of Duron personally. It died a beast, instead of being beaten down like Athlon and Sempron.

 

[dark zero]

joking aside, they may rebrand the A6-9500 as a Sempron.

Sadly, seems that BD chips are EOL at all... time to move forward... maybe those are rated 15 Watts?

 

[Shivansps]

Sadly, seems that BD chips are EOL at all... time to move forward... maybe those are rated 15 Watts?

EOL never stopped them before, i hope they dont, but considering i can still get A4-7300 and A4-4000 in industrial amounts i have a hard time beliving they will just let BD go.

 

[dark zero]

EOL never stopped them before, i hope they dont, but considering i can still get A4-7300 and A4-4000 in industrial amounts i have a hard time beliving they will just let BD go.

Actually 16 and 14 nm are starting to get cheaper and eventually it might be cheaper than 28 nm... So BD might end soon... Unless they release 14 nm BD...

 

[Edgemeal]

AMD will host a live webinar on Mar.12, 2018 at 1:30 p.m. PDT to celebrate the One Year Anniversary of Ryzen.

AMD to Host Ryzen One Year Anniversary Webinar

 

[DAPUNISHER]

Howdy! I am posting from the future. The year is 2025.

Gaming performance is crap for the same reason the google octane score is crap. This is why I been ranting for months... "where is the google octane score?" Now we see that it is garbage. Gaming performance is going to be garbage. Nothing changes. i5-7600k vastly and dramatically destroys all of these ryzen SKUs. Intel has no reason to change anything.

zin quoted this post -

OK.

Let's just quote this one for posterity.

The 7600K had a MSRP of $242, inflation adjusted that's $322. For a 4/4 CPU only 8 years ago. Truly a gaming monster that everyone still praises...wait, what's that? No one talks about it anymore? 7th generation aged like warm milk, you say? And Intel did indeed have reason to change things? They in fact changed lickety-split and had a 6/12 i7 out the door before the year was out? LMAO at "Intel has no reason to change anything". The cope was strong.

The i7 8700K was $359, $478 in today's money. I'll write it again for all of the people that were whinging about the MSRP of the 9800X3D; Since at least the 2600K released way back in 2011, $450-$500 adjusted is what top tier gaming performance has cost.

This thread is also a prime example of how insufficient many bigger bar better reviews had already become. There were games like Battlefield 5, Assassin's Creed, and Witcher 3 with frame pacing issues with 2/4 i3 and 4/4 i5 of the era. Those charts showing how 7th gen was better for gaming were not even fully accurate for when they were published.

 

[Hans Gruber]

Howdy! I am posting from the future. The year is 2025.


zin quoted this post -


The 7600K had a MSRP of $242, inflation adjusted that's $322. For a 4/4 CPU only 8 years ago. Truly a gaming monster that everyone still praises...wait, what's that? No one talks about it anymore? 7th generation aged like warm milk, you say? And Intel did indeed have reason to change things? They in fact changed lickety-split and had a 6/12 i7 out the door before the year was out? LMAO at "Intel has no reason to change anything". The cope was strong.

The i7 8700K was $359, $478 in today's money. I'll write it again for all of the people that were whinging about the MSRP of the 9800X3D; Since at least the 2600K released way back in 2011, $450-$500 adjusted is what top tier gaming performance has cost.

This thread is also a prime example of how insufficient many bigger bar better reviews had already become. There were games like Battlefield 5, Assassin's Creed, and Witcher 3 with frame pacing issues with 2/4 i3 and 4/4 i5 of the era. Those charts showing how 7th gen was better for gaming were not even fully accurate for when they were published.

Battlefield 5 was the first game that said no to 4-core 4-thread CPU's. That is unless you didn't mind the stuttering in game. With regards to the 9800x3D pricing. If AMD followed their tried and true strategy in the past. A reasonable MSRP that declines quickly throughout the product cycle. They would have many more PC builders upgrade every generation or every other generation. You sell it cheap and you sell more chips with each product cycle.

At this point it seems like Intel CPU's are toast (figuratively and literally.) They can't blame TSMC for the current generation Intel chips. Everybody knows TSMC makes quality silicon.

I think a lot of people are burned out with the out of control GPU pricing. The generational performance increase is not what it used to be.

 

[MadRat]

Literally GPU manufacturers are driving up your electrical costs with everything but gaming.