AMD launches Zen+ 12nm Ryzen and X470 motherboards

[The Stilt]

Ah, so all-core turbo and memory overclocking are considered cheats?

If the default configuration is altered when the system is supposed to be at stock, then yes.
The type of scenario we saw with 8700K launch. A competent reviewer will check that power limits or the frequencies are not altered, before publishing any results run as "stock".
This was not the case in several, if not in the majority of 8700K reviews.

 

[IRobot23]

If the default configuration is altered when the system is supposed to be at stock, then yes.
The type of scenario we saw with 8700K launch. A competent reviewer will check that power limits or the frequencies are not altered, before publishing any results run as "stock".
This was not the case in several, if not in the majority of 8700K reviews.

i7 8700K was/is extreme case. Problems with power, stability and cooling. If my R7 1700 goes to 3,2GHz on stock bit over TDP is normal. Well I don''t know if it really went above TDP limits, but faster ram = more soc power.. above tdp.

 

[raghu78]

If the default configuration is altered when the system is supposed to be at stock, then yes.
The type of scenario we saw with 8700K launch. A competent reviewer will check that power limits or the frequencies are not altered, before publishing any results run as "stock".
This was not the case in several, if not in the majority of 8700K reviews.

I disagree. If Intel or AMD send their CPU with a high end Intel Z370 or AMD X470 chipset board with auto overclocking turned on by default in BIOS to reviewers its ok as long as the power consumption results too are measured with the auto OC on. If the user does not like the settings he can turn auto OC off and manually tweak clock and voltage. btw the whole purpose of high end motherboards is to push the CPU to its highest perf, within safety limits though. The option of auto OC vs manual OC is still with the user. Ideally the reviewer should post stock and auto OC perf and power results to provide the user with the information needed to go with one of them.

 

[Atari2600]

I disagree. If Intel or AMD send their CPU with a high end Intel Z370 or AMD X470 chipset board with auto overclocking turned on by default in BIOS to reviewers its ok as long as the power consumption results too are measured with the auto OC on.

Disagree.

No launch reviews should ever look at overclocking conditions, whether triggering one switch in bios or tuning manually for weeks.

First task for any reviewer should always be baselining the product against its ancestors and its competition. This means no overclocking for anything.


If, say, turbo modes are a standard feature (or XFR), then run with. But anything that requires tinkering with bios (even setting one switch) and anything that would be sensitive to cherrypicked* CPUs should never be entertained in a first/launch review.

*power consumption will be sensitive to it, but there is only so much you can do.

 

[raghu78]

Disagree.

No launch reviews should ever look at overclocking conditions, whether triggering one switch in bios or tuning manually for weeks.

First task for any reviewer should always be baselining the product against its ancestors and its competition. This means no overclocking for anything.


If, say, turbo modes are a standard feature (or XFR), then run with. But anything that requires tinkering with bios (even setting one switch) and anything that would be sensitive to cherrypicked* CPUs should never be entertained in a first/launch review.

*power consumption will be sensitive to it, but there is only so much you can do.

What if the default setting in BIOS is with auto OC turned on. Why are you going to tweak the BIOS if out of the box performance and power is what you are testing ?

 

[The Stilt]

What if the default setting in BIOS is with auto OC turned on.

NPRP bioses are required to have such features disabled.

 

[IRobot23]

Hopefully these chips will OC easy to 4,3-4,4GHz @ safe voltage for 24/7. Even R5 1600 (r5 2600 should be even better) at 4,3GHz should be decent and with decent low latency DRAM like 3200CL14/3466CL15/3600CL16 should be great gaming chip. Yet, people might have problems... My experience with C6H is great, but timings on XMP profile are way off.

 

[The Stilt]

Hopefully these chips will OC easy to 4,3-4,4GHz @ safe voltage for 24/7. Even R5 1600 (r5 2600 should be even better) at 4,3GHz should be decent and with decent low latency DRAM like 3200CL14/3466CL15/3600CL16 should be great gaming chip. Yet, people might have problems... My experience with C6H is great, but timings on XMP profile are way off.

XMP is pretty useless for Ryzen as it doesn't support storing the most critical parameters.
AMD should definitely bring back updated AMP.

 

[raghu78]

NPRP bioses are required to have such features disabled.

What does NPRP stand for ? Thanks for the explanation.

 

[The Stilt]

What does NPRP stand for ? Thanks for the explanation.

Media / partner review program.

NPRP bioses are internally audited and then provided to the press.

 

[CatMerc]

XMP is pretty useless for Ryzen as it doesn't support storing the most critical parameters.
AMD should definitely bring back updated AMP.

Not really familiar with XMP/AMP. What's XMP missing for Ryzen?

 

[IRobot23]

AMP would be cool.

Not really familiar with XMP/AMP. What's XMP missing for Ryzen?

short version : xmp on ryzen = some timing are wrong which makes instability.

xmp/amp = sets od timings for dram that runs at higher speed that jedec standard.

In my case XMP only sets CAS 16 18 18 18 correctly other timing are same as 2133MT/s jedec standard or maybe some timings are random numbers. Once it booted to tRC 75 next time 81, 84, 91and 62... all other timings were same.

I was like.. okay no more auto xmp.

 

[The Stilt]

Not really familiar with XMP/AMP. What's XMP missing for Ryzen?

SoC voltage and ProcODT are an actual issue in terms of general functionality.
xSCL timings are an issue in terms of performance lost.

AMP was much more flexible than XMP ever has been.

 

[IRobot23]

SoC voltage and ProcODT are an actual issue in terms of general functionality.
xSCL timings are an issue in terms of performance lost.

AMP was much more flexible than XMP ever has been.

Then Asus did job on XMP here. It changes SoC V when I run XMP to 1-1.1V.

 

[CatMerc]

Set tRDRDSCL, tWRWRSCL to 2 CLKs and tRFC to 308 CLKs.
tWR you can set to 12.

Those will give you the biggest boost, which should be quite significant compared to those timings you posted.

These can all be found under the standard memory timings, each and every board has them.



Try disabling GearDownMode first while leaving Cmd2t to "Auto".
Once the system has posted re-enter the bios and set Cmd2t to 1T.

I suggest that you use >= 1.05V VDDCR_SoC as these settings will put more strain on the memory controller, than the standard configuration.

Tried your suggestion, it managed to boot with Cmd2t and GDM disabled at 1.1v SoC, but when I set it to Cmd1t it failed to boot.

 

[The Stilt]

Tried your suggestion, it managed to boot with Cmd2t and GDM disabled at 1.1v SoC, but when I set it to Cmd1t it failed to boot.

This is at 3200MHz CL14 right, on 2x8GB modules?
Are you running on PinnaclePI based bios or?

 

[CatMerc]

This is at 3200MHz CL14 right, on 2x8GB modules?
Are you running on PinnaclePI based bios or?

Yup.
PinnaclePi 1.0.0.0, BIOS version F20.

F22 refuses to boot at any core clock higher than 3.7GHz lol

 

[.vodka]

Yup.
PinnaclePi 1.0.0.0, BIOS version F20.

F22 refuses to boot at any core clock higher than 3.7GHz lol

Either your motherboard hates you, or your 1700 lost the silicon lottery in many ways..

This is from the K7's manual. Are you by chance not using slots 1 and 2?
What procODT are you using? Single rank B-die sticks are usually stable with 53.3 or 60 ohms.

Can you boot CL14 and 1T+Geardown? It could be thought of a "1.5T" mode and it only works with even CAS latency. Odd will round up.

 

[CatMerc]

Either your motherboard hates you, or your 1700 lost the silicon lottery in many ways..

This is from the K7's manual. Are you by chance not using slots 1 and 2?
What procODT are you using? Single rank B-die sticks are usually stable with 53.3 or 60 ohms.

Can you boot CL14 and 1T+Geardown? It could be thought of a "1.5T" mode and it only works with even CAS latency. Odd will round up.

I'm indeed not using them. RAM currently installed on DDR4_4 and DDR4_3. I'll try moving them.
ProcODT is 60, trying 53 doesn't really help.
And I can boot 1T+Geardown with tight timings just fine, in fact it's what I'm running right now and tested for 24 hours for any errors with RAMTest.

The clockspeed thing is a known issue, many people can't boot higher clocks with the latest BIOS for some reason. I can take my chip to 24/7 stable 3.9GHz at 1.375V and even do cinebench runs at 4.1GHz and 1.45V, but latest BIOS won't boot at 3.8 or higher. If I set my clockspeed in Ryzen master after I've booted it works just fine.

 

[CatMerc]

Whelp, moved my DRAM, and it won't boot even the same configuration as before lol

Won't even boot XMP.

Edit: Reseated the RAM and it's working, though no improvement in the GDM situation.

 

[DisEnchantment]

Somebody in Italy got their sample yesterday,
Before it is taken down, I took a snapshot.
Quick Summary
Ryzen 2700X stock, DDR4 3200 CL14, Gigabyte AX370 Gaming 5, GTX 1080Ti
CB MT 1780/ST 174
Memory Latency 66.4 ns
Max Clock 4.32 GHz

He did not perform OC, Now if those 5-8 % gains of 2700X over stock are real as said by CPC, this chip is a good improvement over Ryzen Gen 1.
There is a YouTube video one and half hour long of the test.

Looks like a repeat of RR release, some guy nicely stream the unbox and test of RR for 8 hours , couple of days before NDA when he delivered a chip by Amazon.

Sandra Link

http://ranker.sisoftware.net/show_r...d4ecdee9ddebdef88ab787a1c4a19cac8af9c4fc&l=en

Spoiler: YouTube Video

Spoiler: AIDA Cache & Memory

Spoiler: CB15 Single Core

Spoiler: CB15 Multi Core

Spoiler: SiSoft Results

 

[Justinbaileyman]

Why is the video in your link showing the 2700x is 105w in cpu-z??

 

[exquisitechar]

Hm, 174 points in CB seems low if it's boosting to 4.3GHz, from what I know about the first generation Ryzen's scores.

Why is the video in your link showing the 2700x is 105w in cpu-z??

Because its TDP is 105W... according to leaks, at least.

 

[jonijs]

Looking at the picture it says 3.7Ghz

 

[epsilon84]

Looking at the picture it says 3.7Ghz

Pretty sure thats the base clock, not the actual operating frequency. If not, thats an insanely high ST score for 3.7GHz, higher than Intel in fact.