AMD launches Zen+ 12nm Ryzen and X470 motherboards

1. CanardPC has no credibility so what Canard claims is irrelevant.

2. AsRock said right on its website that the AsRock A320M HDV (which is the motherboard that CanardPC used) only support up to 65W processors.

1. They are a respectable magazine. Moreover, Canard gets that 2700X is 3.4% faster than 1800X @1440p and AMD gets that 2700X is 5% faster than 1800X #@1080p. So the performance measured by Canard agrees with internal measurements made by AMD.

2. In their review the 1800X was a bit faster than the 8700k and the 2700X was 77% faster than the 1500X. Not bad for using a mobo that doesn't support the 1800x and the 2700X according to you and others. :D Of course, they didn't use a mobo limited to 65W chips. It is obvious.
 
1. They are a respectable magazine. Moreover, Canard gets that 2700X is 3.4% faster than 1800X @1440p and AMD gets that 2700X is 5% faster than 1800X #@1080p. So the performance measured by Canard agrees with internal measurements made by AMD.

CanardPC is "respectable" in the same way that the National Inquirer is respectable.

You should open the dictionary and look up what the word "canard" means.

2. In their review the 1800X was a bit faster than the 8700k and the 2700X was 77% faster than the 1500X. Not bad for using a mobo that doesn't support the 1800x and the 2700X according to you and others. :D Of course, they didn't use a mobo limited to 65W chips. It is obvious.

The motherboard (AsRock A320M-HDV) is on the front cover of the magazine.
 
Last edited:
Do you mean that A320 mobo that I mentioned above and that gives only 1.6%, 1.7%, 0.2%,... lower performance for the 1800X when compared to the fastest X370 mobo? Not bad for a board that only supports 65W chips according to you and others here. ROFL

CPC measured performance of 2000-series chips on different mobos and claims the difference between using A320 and X370/470 is only 1--2% on stock settings. Overclocking in France claims the same. And the above review of 24 mobos proves about the same for the 1800X.

So Joe Blow can pair the 2700X with a X370/470 motherboard, but the stock performance will be only 1--2% higher than the performance reported in the CPC magazine. If Joe plans to overclock, then he will see 5--8% higher performance when using one of the top mobos.

Your excuse about performance has been debunked with data from three different sources, and your claim that A320 mobos only support 65W CPUs is untrue.

I guess it would have been easier to just include the results from the x370 and the x470 and the b350 and the b450 motherboards they totally tested when the published the article. Since they had the testing totally done, they could have used REAL numbers to show the differences and the ever present automatic omnipresent cheats that those evil boards produce instead of a small quip in a sentence about 'there is a 1% or 2% difference' (I'm assuming they should put "we guess" in there as well.) Maybe they could just amend their article to actually show these differences but I guess they would need actual numbers to do that with. Well pretty soon, we shall have those numbers from other websites and maybe CanardPC could just steal those numbers when they become available.
 
auvnhwvii4q01.jpg


87101.png
 
So 4.0 stable oc, looks like on stock cooler? That's pretty good, 3.75 to 3.8 was max on stock cooling on gen 1 zen.
 
Not bad not bad. So what is the final boost clock for 2700k 4.25ghz? Looks like it can hit 4.3ghz. Curious if it can hit 4.4ghz.
 
CanardPC is "respectable" in the same way that the National Inquirer is respectable.

You should open the dictionary and look up what the word "canard" means.



The motherboard (AsRock A320M-HDV) is on the front cover of the magazine.

So you ignored all performance arguments that I gave you, and also the content of the review, including the power and clock measurements.

I guess it would have been easier to just include the results from the x370 and the x470 and the b350 and the b450 motherboards they totally tested when the published the article. Since they had the testing totally done, they could have used REAL numbers to show the differences and the ever present automatic omnipresent cheats that those evil boards produce instead of a small quip in a sentence about 'there is a 1% or 2% difference' (I'm assuming they should put "we guess" in there as well.) Maybe they could just amend their article to actually show these differences but I guess they would need actual numbers to do that with. Well pretty soon, we shall have those numbers from other websites and maybe CanardPC could just steal those numbers when they become available.

It is a print magazine. There are space limitations.


So stock 2700x is about 9.9% faster than 1800X, which agrees with the 8.7% found by Canard when we subtract the 1--2% extra by using X370/470 mobos.
 
just because the base design allows a higher speed, more capacity, reduced voltages and like does not mean it automatically will.

there are thousands possibly millions of changes that can be made to the design in question, some small, some very large.

base example GDDR memory from Hynix or from Samsung, just because one pushes the spec for 1600Mhz at 3v does not mean the other cannot do at 1900Mhz for 2.6v because they tweaked the PCB some.

a white paper and real world on the finalized product are vastly different things, all depends on how it was used, the voltages fed etc etc etc....these are not exactly lego block "put the blue block here, yellow block there and you will have a house" these are very complex things, every transistor, every mask added, every etch etc etc.....it is very complex, could very well be that TSMC, GF, IBM, Samsung etc base their "spec" on SRAM for the transistors so is a very "loose" approximation.

HBM/HBM2/GDDR5 among others have shown EXACTLY this difference, not counting if AMD want to "buy" the low voltage parts via a 9T library 12nm and choose to have the high performance on the 12T library instead....way too many possibilities to sum them up (currently)

TSMC/GF/IBM/Samsung/Intel obviously follow their own "recipes" so transistor count, cell density while the same "theory" are very much their own specific "baked apple pie" for lack of a better term.,

hell AMD was supposed to be 14nm to 7nm and all of a sudden TSMC/GF etc decided to "tweak" their respective 16nm/14nm and call it + version instead maybe in preparation for 10nm, 7nm or whatever to work out potential kinks for wafer yield, production amounts they want/need from the machinery etc.

east it seemed this way to me, 14nm to "12nm" seemed like a blink of the eye compared to how long it took to go from 45nm to 32nm to 28nm to 20nm or whatever....very odd indeed ^.^
 
So you ignored all performance arguments that I gave you, and also the content of the review, including the power and clock measurements.



It is a print magazine. There are space limitations.



So stock 2700x is about 9.9% faster than 1800X, which agrees with the 8.7% found by Canard when we subtract the 1--2% extra by using X370/470 mobos.

It's like somebody claiming that she would have finished her lap 5% faster, but she was wearing the wrong type of shoes.

The only way to know for sure is for her to change her shoes and run another lap.

Right now, all I have is her claim and her claim doesn't count.
 
hell AMD was supposed to be 14nm to 7nm and all of a sudden TSMC/GF etc decided to "tweak" their respective 16nm/14nm and call it + version instead maybe in preparation for 10nm, 7nm or whatever to work out potential kinks for wafer yield, production amounts they want/need from the machinery etc.

Glofo roadmap was 14LPP --> 10LP --> 7LP before 10LP was canceled and 12LP announced.
 
It's like somebody claiming that she would have finished her lap 5% faster, but she was wearing the wrong type of shoes.

The only way to know for sure is for her to change her shoes and run another lap.

Right now, all I have is her claim and her claim doesn't count.

We have now CPC review, the claims by overclocking made in France, the AMD internal measurements and the CB leak that you posted above. All the sources give the same performance for the 2700X...
 
We have now CPC review

Calling it a "review" is a joke.

No reputable source would test a 105W processor on a motherboard that only support processors up to 65W and call it a "review".

You can analyze the "results" all you want, but the methodology is wrong so the "results" are automatically invalided.

the claims by overclocking made in France, the AMD internal measurements

sources?
 
Calling it a "review" is a joke.

No reputable source would test a 105W processor on a motherboard that only support processors up to 65W and call it a "review".

You can analyze the "results" all you want, but the methodology is wrong so the "results" are automatically invalided.



sources?


Yep this is the equivalent of testing a b-360 cheapo with an 8700 before using xtu software:

https://www.techspot.com/review/1603-intel-b360-chipset/

The average CB run was just 1160 points! Compare that to the 1400 point average on other 8700 when not applying mcu.

Sorry, but I will wait for the real reviews with respectable motherboards in regards to Ryzen+ overclocking.
 
Yep this is the equivalent of testing a b-360 cheapo with an 8700 before using xtu software:

https://www.techspot.com/review/1603-intel-b360-chipset/

The average CB run was just 1160 points! Compare that to the 1400 point average on other 8700 when not applying mcu.

Sorry, but I will wait for the real reviews with respectable motherboards in regards to Ryzen+ overclocking.

You would think that there wouldn't be a PC hardware magazine that is stupid enough to try this when ASRock literally said (quote): Supports 65W Ryzen CPUs Only!
 
Last edited:
You would think that there wouldn't be a PC hardware magazine that is stupid enough to try this when ASRock literally said (quote): Supports 65W Ryzen CPUs Only!

ASRock probably idiot-proof the motherboard so that it would forced the processor to run at 65W.
Don't get so worked up over this the only reason those people keep posting their message is they thrive on posting things that are far from the truth and get a lot of people worked up over it.

There is no A320 testing scene nor has anyone ever came up with why there should be one (for valid reasons).
 
Calling it a "review" is a joke.

No reputable source would test a 105W processor on a motherboard that only support processors up to 65W and call it a "review".

You can analyze the "results" all you want, but the methodology is wrong so the "results" are automatically invalided.



sources?

The analysis of the results prove that the review is accurate. Clocks, power, and performance reported are correct. This is the same kind of critical analysis that proves that the TPU review that you mention often is a piece of nonsense.

CPC got that the 2700X is 8.7% faster than 1800X. Adding the 1--2% correction for the mobo change A320-->X370/470 implies the 2700X is about 9.8--10.8% faster when everything is stock and a top mobo is used.

The CB leak that you posted in #286 shows that 2700X (stock) is 10% faster than 1800X

1786/1625 = 1.099

So the CB leak that you posted agrees with the performance measured by CPC. You can admit this is the performance of Pinnacle Ridge or you can continue looking for lame excuses to reject the CPC review.

What will be the next lame excuse? That the user that purchased the 2700X and run CB on it and OC up to 4.3GHz is using a motherboard that only supports 65W? And please don't ask for sources have been given two or three times before.
 
Review: AMD Ryzen 7 2700X (Chipset X370)
BY IVÁN MARTÍNEZ
04/06/2018
CPUS & APUS
AMD-Ryzen-Logo-740x339.jpg


Fast forward to the launch date on April 19, we present in a world premiere the review of the new AMD Ryzen 7 2700X , the flagship of the second generation of AMD Ryzen processors. Like its predecessor, the Ryzen 7 1700X , it has 8 cores and 16 threads , which increase its frequencies up to 3.70 GHz , and can reach 4.35 GHz in Boost mode. with an AMD X470 motherboard.

Let's see first its complete technical specifications :

AMD Ryzen 7 2700X
Socket AM4
Fabrication process 12 nm
Cores / Threads 8 cores / 16 threads
Base Frequency 3.70 GHz
Turbo Frequency 4.35 GHz
Cache L3 2 × 8 MB
Compatible Memory DDR4-2933 Dual-Channel
PCIe lines 16 lines
TDP
Before starting the review, we would like to thank AMD for not sending us their products and therefore not forcing us to sign their NDA , so we can offer this review before anyone else. Next week, the AMD Ryzen 5 2600 .

Index of contents [Ocultar]

Packaging and Accessories
AMD-Ryzen-7-2700X-01-694x600.jpg

AMD-Ryzen-7-2700X-01-1-1-694x600.jpg

AMD-Ryzen-7-2700X-02-694x600.jpg

AMD-Ryzen-7-2700X-02-1-694x600.jpg

The AMD Ryzen 7 2700X arrives in a box similar to the previous generation, showing the processor from the side and protecting it perfectly inside, as well as the heatsink.

AMD-Ryzen-7-2700X-03-740x555.jpg


AMD-Ryzen-7-2700X-04-656x600.jpg


In addition to the warranty card and a sticker of Ryzen 7, this model comes with the AMD Wraith Prism , a model that acquires configurable RGB LED lighting from a USB header or from plates compatible with RGB strips , so we can customize your lighting to taste.

AMD-Ryzen-7-2700X-05-659x600.jpg


AMD-Ryzen-7-2700X-06-740x530.jpg


AMD-Ryzen-7-2700X-07-659x600.jpg


AMD-Ryzen-7-2700X-08-740x532.jpg


AMD Ryzen 7 2700X
AMD-Ryzen-7-2700X-09-740x573.jpg


As we anticipated, the AMD Ryzen 7 2700X offers 8 cores , which in turn offer 16 threads that start from a base frequency of 3.7 GHz to reach 4.35 GHz in Turbo mode thanks to the XFR 2 automatic overclocking technologies and Precision Boost Overdrive . It offers a total of 2 × 8 MB of L3 cache , as well as 8 x 512 KB of L2 cache and 8 x 32 KB + 8 x 64 KB of L1 cache, thus offering a cache distribution identical to that of the Ryzen 7 1700X .

AMD-Ryzen-7-2700X-10-740x369.jpg


As we can see in CPU-Z , the set of instructions supported has MMX (+), SSE, SSE2, SSE3, SSSE3, SSE4.1, SSE4.2, SSE4A, x86-64, AMD-V, AES, AVX, AVX2, FMA3 and SHA, also the same as in the first generation of Ryzen .

AMD-Ryzen-7-2700X-11-740x573.jpg


Obviously it is a processor for the AM4 socket and therefore makes use of the " Pin Grip Array " connection system, where the pins are in the processor and not in the motherboard, highlighting the use of a total of 1,331 contacts.

Test Equipment
AMD-Ryzen-7-2700X-12-638x600.jpg


To analyze the performance of the AMD Ryzen 7 2700X processor we have assembled the following test equipment:

AMD-Ryzen-7-2700X-13-740x493.jpg


We have done the tests with the AMD Wraith Prism that accompanies the processor, which offers a sensational design once turned on and equipped with its RGB LED lighting .

Tests
Next, we show you the results of the performance tests performed on this and other processors in calculation benchmarks (wPrime), rendering (Cinebench), coding (x264), memories (Aida64) or focused on games (3D Mark).

AMD-Ryzen-7-2700X-Tests-01.jpg


AMD-Ryzen-7-2700X-Tests-02.jpg


AMD-Ryzen-7-2700X-Tests-03.jpg


AMD-Ryzen-7-2700X-Tests-05.jpg


AMD-Ryzen-7-2700X-Tests-04.jpg


AMD-Ryzen-7-2700X-Tests-06.jpg


AMD-Ryzen-7-2700X-Tests-07.jpg


Note: We have not been able to test games because the platform became unstable when opened , with crashes shortly after starting or even blue screenshots. Not having specific drivers for the new generation, it's something that could happen, but the funny thing is that it only happens in games , which would foreseeably be his weak point against Intel, so we'll have to wait a little longer to know if this will continue to be the case

Temperature, Overclocking and Consumption
Under an ambient temperature of 20ºC, the AMD Ryzen 7 2700X offers temperatures of about 31ºC at rest and about 56ºC in its stress test, a really good value that slightly improves that of the previous generation. The heatsink complies correctly in terms of loudness, around 40 dB at full performance, although it could rise to 45 dB if we take it to the limit.

On the consumption , to say that it has improved notably remaining around 200W for the complete equipment in the stress test of Aida64 (without discs or GPU). It is similar to that offered by Intel i7 processors, improving the Ryzen 7 1700X by 50W .

AMD-Ryzen-7-2700X-14.jpg


Overclocking is still not the strong point of these processors , because although they have increased their frequencies compared to the previous generation, we have only managed to bring the 8x cores to 4.20 GHz with 1.4V . The Ryzen Master OC application has not been able to use it for not having it updated to give compatibility to these processors, but without problems from the BIOS.

conclusion
AMD-Ryzen-7-2700X-99-740x584.jpg


The AMD Ryzen 7 2700X is a good review of the Ryzen 7 1700X with which AMD surprised us last year, improving it in all aspects.

On the one hand, we have higher frequencies , which even increase significantly with the new XFR 2 mode. This supposes an improvement around 10% in the different benchmarks that we have passed in front of its predecessor, something similar to what Intel offers in each new generation (up to the 8th). However, the memories seem not to behave correctly, although it may be due to not having specific drivers for this generation yet. We will check it soon.

In games we could not analyze if they have improved their poor performance against Intel, because we have had problems as we have discussed with all of them. Again, we will reanalyze it when we have X470 motherboards (or access to a new BIOS for the X370 chipset).

If the prices that we saw in Amazon Germany yesterday are true , it will go on sale around 335 euros , competing one on one with the i7-8700K , so we are in a situation similar to that of the previous generation.

Main advantages
+ Good performance / price ratio for professional environments
+ Excellent multi-core performance
+ Very good temperatures and consumption
+ 8 cores / 16 threads
+ Compatible with previous generation base plates
Negative aspects
- The memories do not finish going fine, at least for the moment
- Little overclocking capacity
Other Aspects
* We do not know his performance in games
Therefore, from El Chapuzas Informático we awarded the Gold Award to this processor for its good improvement over the previous generation, although it still has much to prove, especially in games.

Galardon-Oro.png




Read more https://elchapuzasinformatico.com/2018/04/amd-ryzen-7-2700x-review/
 
Last edited:
So according to El chapuzas, the 2700X is 8% faster than 1700X on Wprime MT, 12.8% faster in CB15, and 11.6% faster in x264. The average is about 11%.

According to the CPC review the 2700X was 13.6% faster than the 1700X in MT.

13.6% > 11%

And el chapuzas used a top X370 mobo. :D
 
So according to El chapuzas, the 2700X is 8% faster than 1700X on Wprime MT, 12.8% faster in CB15, and 11.6% faster in x264. The average is about 11%.

According to the CPC review the 2700X was 13.6% faster than the 1700X in MT.

13.6% > 11%

And el chapuzas used a top X370 mobo. :D

See, and there isn't all the questions about this or that, they tested on a motherboard that doesn't put a 10% negative impact on the CPU. If CPC just did this they wouldn't have all the questions, about their testing standards. Also isn't that higher than the jump from 6700 to 7700 though?
 
See, and there isn't all the questions about this or that, they tested on a motherboard that doesn't put a 10% negative impact on the CPU. If CPC just did this they wouldn't have all the questions, about their testing standards. Also isn't that higher than the jump from 6700 to 7700 though?

So the "10% negative impact on the CPU" was only in the imagination of some people, because the 2700X/1700X performance gap in the CPC review was higher than in El chapuzas review: 13.6% vs only 11%.

The CPC review was basically correct. Contrary to BS and silly accusations CPC didn't cripple the performance of the 2700X. At contrary. CPC favored the 2700X, the 2600X and the 2600 in their review, because CPC tested all on stock settings. So the 1000-series chips used 2666MHz memory whereas the 2000-series chips used 2993MHz, which increased the performance gap a bit. El chapuzas tested all the chips with 3200MHz RAM, which reduced the advantage. That is why CPC got 13.6% performance gap but El chapuzas only got 11%.

The 2700X is a replacement for the 1800X as well. The performance gaps are

1600-->2600: 7.6% faster

1600X-->2600X: 8.7% faster

1800X-->2700X: 8.7% faster

The average is 8.3% faster.
 
Last edited:
So Ryzen 7 2700X ~ Core i7-7820X

AMD basically went from Broadwell => Skylake with a minor update.

That’s not bad at all.
 
Definitely looking forward to the 2600X, since it also comes with the Wraith Spire cooler.

I came ever so close to getting things together for a 1600X system, but held off, since this is supposed to be around the corner, and would cost about the same, if not less. My FX-4100 platform, while still capable of playing all of my favorite games at 1080, is getting rather long in the tooth. I'm probably going to make my FX-4100 (running at 4.4 GHz) my new work PC, and make my main work PC (an A3850-based PC) the secondary.
 
I wonder if the X based models are going to be more heavily binned than last time around. Only basing this off of the fact that a few non-X buyers are reporting 4ghz max overclocks while the 2700X owners seem to be able to to hit 4.3 with enough (probably unsafe) voltage.
 
Definitely a possibility early on, as we're not on a super-mature fabrication process and their ability to make perfect chips all-day-every-day is probably limited. That landscape may change in 6 months of production, where the vanilla vs "x" difference might only be their stock clocks, as seems to be the case with the 1X00(x) series Ryzens. To be determined.
 
I wonder if the X based models are going to be more heavily binned than last time around. Only basing this off of the fact that a few non-X buyers are reporting 4ghz max overclocks while the 2700X owners seem to be able to to hit 4.3 with enough (probably unsafe) voltage.
All depends on the yeilds since they still have to bin for threadripper.

Either way cpuz's voltage is slightly higher than the actual voltage.. it's probably closer to 1.45v which is the recommended voltage limit from amd though 1.5v should be just fine unless you plan to use the processor for the next 10 years.
 
My Ryzen 1700 is only good to 3.5GHz under water. At 3.7GHz weird shit starts happening with my desktop like I will launch a program and it instantly closes. This behavior leads to eventual lockups after the system has been live for a week or more. I'd buy a binned chip if I had to do it over again.
 
My Ryzen 1700 is only good to 3.5GHz under water. At 3.7GHz weird shit starts happening with my desktop like I will launch a program and it instantly closes. This behavior leads to eventual lockups after the system has been live for a week or more. I'd buy a binned chip if I had to do it over again.

sounds similar to what my r5 1600 was doing when i was having memory stability problems.. no matter what my processor did not like me setting the SoC voltage below the auto 1.21v
 
I came ever so close to getting things together for a 1600X system, but held off, since this is supposed to be around the corner, and would cost about the same, if not less.
Not to debate your decision to wait, but how is the new system going to cost the same or more importantly, LESS?
 
Back
Top