AMD launches Zen+ 12nm Ryzen and X470 motherboards

[Mockingbird]

A return to heatpipes and clips? Huh? I wonder if that slide is true or not, seems potentially fake.

How can it be fake when the slide came from AMD?

 

[IRobot23]

I was quite surprise by stock cooling on my R7 1700, definitely AMDs solder and stock cooler are great for everyone.

 

[Atari2600]

How can it be fake when the slide came from AMD?

#FakeNews

#Faux&Friends

 

[Magic Hate Ball]

A return to heatpipes and clips? Huh? I wonder if that slide is true or not, seems potentially fake.

It's pretty much identical in design to the Wraith Max I have on my 2400G currently. Just with even more RGB bling!

 

[eek2121]

Pinnacle Ridge AGESA (which is rolling out to all AM4 motherboards these days in preparation for Ryzen 2000 series launch) has finally fixed pstates, you can set them to whatever you want in BIOS without breaking any functionality. Then you can enable cstates (should be enabled by default, but it doesn't hurt to force them on) to let the hardware properly idle, and use an appropriate power plan in Windows.

Here's the result:

• P0 is set to 3.8GHz and 1.375v (no LLC, droops down to ~1.31v under P95 AVX small FFTs load)
• vSOC is set to 1.1v to support 3466MHz memory speed, vDIMM 1.425v
  
• Ryzen Balanced power plan with CPU idle speed set to 0% instead of 90% because I'd like to let Windows use the entire clock speed range

See, proper idle (hwinfo managed to "detect" P3 at 0.4v) with overclocked speeds. You could more or less get the same result since launch day on the C6H with voltage offsets and other workarounds, but thankfully now it's much simpler and can be done on every board as far as I'm aware.



Anyway, having cstates enabled lets Ryzen take control of power management all by itself and that's what you want to do in this platform. Entering OC mode doesn't disable power savings.

https://www.reddit.com/r/Amd/comments/6w793f/how_to_reduce_idle_clock_speed_on_a_manually/

I'm old fashioned and that's why I take that 90% idle down to 0%, but it's not needed.


Obviously, power management this advanced is miles ahead of any faildozer part ever made, and Pinnacle Ridge is probably going to make even more improvements in this area.

Saw this post and tried it on my 1950x. Sure enough, 0% took it down to 2 GHz when idle. NICE!

 

[cellarnoise]

Love my cheap 1700x, my msi b350 gaming pro carbon voltage droops bad when trying to overclocking the cpu. currently running 3200 cl14 ram at 3333 cl14 stiltz for many months. Afraid to upgrade bios after latest-est new processors. Read not good.

Very stable at stock CPU and O.C. ram. Does all our audio and video editing in enjoyable times. And pushing intel to better and better after many years of blah.......

 

[B-Riz]

Saw on Reddit that CarnardPC is supposed to release a 2700X review today, but on a 320 series board, lol.

 

[Gideon]

Some info about the CanardPC article from this thread.

Please buy CPC if you are French (online if possible because of their distributor problems... ) but:

• Improved latency, slowly closing the gap with Intel
  
  
• Much smoother turbo
  
  
• Fantastic "brute force" performance, still worse than a 8400 in games.
  
  
• Power consumption goes wild.

According the CPC, the X470/B450 and X370/B350 chipsets are ENTIRELY IDENTICAL, they are literally renames. XFR2 Enhanced and Precision Boost only depend on the board but because not all old gen boards were up to spec AMD wanted to avoid a situation were some boards had XFR2/PrecBoost and some board didn't and decided to rebrand the chipset and impose more stringent standards.

Latencies should be really good, especially if these chips support 3466+ Mhz more widely. Precision Boost 2.0 also seems very nice (especially when considering that the benches were run on a A320 chipset:

 

[Atari2600]

That's a pretty nice curve for AMD.

Prob not great for overclockers, but good for AMD.

 

[Bouowmx]

Are 'Precision Boost 2' freqency targets pre-programmed and static?

 

[moinmoin]

Are 'Precision Boost 2' freqency targets pre-programmed and static?

They are supposedly dynamic based on the headroom available. Though still restricted seeing how the leaked slides points to "XFR 2 Enhanced" and "Precision Boost Overdrive" unlocking even more headroom...

Going by that slide CanardPC's numbers should be using XFR 2 Enhanced.

 

[The Stilt]

Are 'Precision Boost 2' freqency targets pre-programmed and static?

Dynamic operation within specific limits (power, thermal and reliability "FIT").
Fixed utilization (or core) limit was already gone in Raven so I assume that applies on PR as well.

 

[ao_ika_red]

Some info about the CanardPC article from this thread.


Latencies should be really good, especially if these chips support 3466+ Mhz more widely. Precision Boost 2.0 also seems very nice (especially when considering that the benches were run on a A320 chipset:

That's quite a hard work for A320 board.

 

[DisEnchantment]

Some info about the CanardPC article from this thread.


Latencies should be really good, especially if these chips support 3466+ Mhz more widely. Precision Boost 2.0 also seems very nice (especially when considering that the benches were run on a A320 chipset:

So with Precision Boost Overdrive and XFR 2.0 Enhanced we could see 4.5 GHz SC boost, when not thermal constrained?
Looks like low 60s ns Latencies possible at > 3200 MHz , tight timings?
With a faster RAM kit and 400 series board the review might have looked a bit different.

When I got my 1700X I had 97ns latency for two or three months XD, until I got it to work at 3200@CL14.

 

[CatMerc]

So with Precision Boost Overdrive and XFR 2.0 Enhanced we could see 4.5 GHz SC boost, when not thermal constrained?
Looks like low 60s ns Latencies possible at > 3200 MHz , tight timings?
With a faster RAM kit and 400 series board the review might have looked a bit different.

No, all three modes have the same max clockspeed. XFR 2.O and Overdrive just increase the limits so that when you load more cores you don't lose as much frequency.

 

[exquisitechar]

Very nice multi core performance improvement, the improved turbo is great. Power consumption doesn't seem that bad to me, it's a nice improvement over gen 1 Ryzens, especially when you take into account how the turbo boost is much higher for multiple cores. Gaming performance is a let down, not even a 4% improvement over the 1800X, but obviously we'll need to see a lot more in-depth testing.

 

[sandorski]

A return to heatpipes and clips? Huh? I wonder if that slide is true or not, seems potentially fake.

Especially the Clips, ugh.

 

[NinjaCool]

There is some new info on videocardz.
https://videocardz.com/75504/cpc-hardware-amd-ryzen-7-2700x-review-goes-online

 

[Bouowmx]

Which limits are in effect for Precision Boost 2 at low core counts? They don't include power because 2 cores at 4175 MHz using 105 W (the TDP) doesn't sound right?

 

[Mockingbird]

There is some new info on videocardz.
https://videocardz.com/75504/cpc-hardware-amd-ryzen-7-2700x-review-goes-online

Now, you have to remember that CanardPC uses an A320 motherboard for testing.

You would think that a magazine about PC hardware would at least buy a proper X370 motherboard for testing.

 

[CatMerc]

Which limits are in effect for Precision Boost 2 at low core counts? They don't include power because 2 cores at 4175 MHz using 105 W (the TDP) doesn't sound right?

Temperature, Power, and "Reliability" whatever it means in practice.

Ryzen can currently push 40W through two cores, maybe 4.3 on two cores is like 60W-80W.

 

[Kenmitch]

Now, you have to remember that CanardPC uses an A320 motherboard for testing.

You would think that a magazine about PC hardware would at least buy a proper X370 motherboard for testing.

Just consider it as a sneak peak if nothing else. It will be interesting to view and compare results from different chipsets once the NDA lifts and proper reviews drop.

 

[Kenmitch]

Now, you have to remember that CanardPC uses an A320 motherboard for testing.

You would think that a magazine about PC hardware would at least buy a proper X370 motherboard for testing.

Just consider it as a sneak peak if nothing else. It will be interesting to view and compare results from different chipsets once the NDA lifts and proper reviews drop.

Why is SMT off in the chart with turbo clocks?

 

[Olikan]

Ddr4 is propably be using JEDEC standarts... cas 19

 

[DisEnchantment]

How are these guys even publishing stuffs like this? They got their sample from someone else so no NDA?
I understand the need to get the review out, but it such a weird thing, publishing data but withholding something because of not wanting to show everything but already disclosing anyhow. Then using a Low end board for a high end part? That too in print when you want to get first coverage ?
Also that choice of games...
I cannot fully understand this Canard PC.