AMD launches Zen+ 12nm Ryzen and X470 motherboards

juanrga

2[H]4U
Joined
Feb 22, 2017
Messages
2,804
Wouldn't it depend on the comparison?

Since their baseline is the 7600k @ 100%, Zen is 150% and Zen+ is 170%, then one can say, in this test, Zen+ is 20% faster than Zen. (Since we are comparing against the 100% baseline) which is why I don't like those comparisons at all :D

If one is 150% and the other is 170% then the second chip is 13% faster, not 20%.
 

Digital Viper-X-

[H]F Junkie
Joined
Dec 9, 2000
Messages
14,966
If one is 150% and the other is 170% then the second chip is 13% faster, not 20%.

Nope. IT depends on the comparison. When you use a baseline, you reference the performance vs that baseline. When you word it as "In this test, Ryzen 2 gives you another 20% over the 7600k" , technically you are correct, however it's more word play, than it is an honest representation of what they were trying to show.

The number will again change, once you change your reference point.

EG.

Ryzen 1 = 150 fpfs
Ryzen 2 = 170 fpfs

170/150 = 13.3% , so when comparing the two, from the reference/baseline of the Ryzen 1 chip, its 13.3% faster than Ryzen 1.

When you chance your reference point to the Ryzen 2 cpu , 150/170 = 88.23%, or 11.77% slower. From the reference point of Ryzen 2, Ryzen 1 is 11.77% slower.

Though I could be wrong, I'm pretty sure this is used regularly to skew numbers in favor of one thing or another.
 

OrangeKhrush

[H]ard|Gawd
Joined
Dec 15, 2016
Messages
1,673
I don't have an issue with a cheap A320 board being used, I don't agree that it doesn't matter much, at first they say 1-2% then it goes to 5-8% and if we take the 8700K's "cheat mode activated" it scored up to 33% higher in CB which is significant. I assume the Asus X470 boards will have highest end quality caps, chokes and VRM's which will push the sustained boost modes higher. Canard PC is face value only, there is no reported clocks or methodology used, they could be throwing the worst case scenario which putting a 2000X chip on a A320 is, you will get choked memory performance, you will get turbo limits and no OCing, basically testing poorest quality equipment where they probably put Intel on the best and may even be running MCE.

At this point I will wait for more reputable reviewers to handle the matter, not Frenchies with a tendency to break NDA at will.
 

OrangeKhrush

[H]ard|Gawd
Joined
Dec 15, 2016
Messages
1,673
Objections:

1) Canard PC are excluded from AMD's testing program due to repeated NDA violations, this i confirmed with someone internally.
2) The chips are likely black market ES chips with old micro code to ensure they work on basic platforms
3) no methodology provided.
 

Dermac

Limp Gawd
Joined
May 6, 2005
Messages
152
Objections:

1) Canard PC are excluded from AMD's testing program due to repeated NDA violations, this i confirmed with someone internally.
2) The chips are likely black market ES chips with old micro code to ensure they work on basic platforms
3) no methodology provided.

So they will break NDA, use ES chips to signify retail chips, and have no discernible methodology to their testing. Sounds like they are just a click bait site, willing to circumvent the norms of the industry to put sensationalist information out on the net for little more than profit. This kinda ranks them with the news sites that have "10 Famous People pictures you won't believe" articles, total trash.
 

Pieter3dnow

Supreme [H]ardness
Joined
Jul 29, 2009
Messages
6,784
Objections:

1) Canard PC are excluded from AMD's testing program due to repeated NDA violations, this i confirmed with someone internally.
2) The chips are likely black market ES chips with old micro code to ensure they work on basic platforms
3) no methodology provided.

But at this point most of us already know to wait for websites to post their results on the new 2xxx range ;).
Never worry about some test results on early samples and a motherboard from a series next to no one uses.
 

juanrga

2[H]4U
Joined
Feb 22, 2017
Messages
2,804
Nope. IT depends on the comparison. When you use a baseline, you reference the performance vs that baseline. When you word it as "In this test, Ryzen 2 gives you another 20% over the 7600k" , technically you are correct, however it's more word play, than it is an honest representation of what they were trying to show.

The number will again change, once you change your reference point.

EG.

Ryzen 1 = 150 fpfs
Ryzen 2 = 170 fpfs

170/150 = 13.3% , so when comparing the two, from the reference/baseline of the Ryzen 1 chip, its 13.3% faster than Ryzen 1.

When you chance your reference point to the Ryzen 2 cpu , 150/170 = 88.23%, or 11.77% slower. From the reference point of Ryzen 2, Ryzen 1 is 11.77% slower.

Though I could be wrong, I'm pretty sure this is used regularly to skew numbers in favor of one thing or another.

No.

If the R2 is 13.3% faster than R1, then the R1 is 11.8% slower than R2. Independently of the baseline used.

CPCHardware used the i5-7600k as baseline for its performance graphs. They could change the baseline to i7-7700k, R7-1800X, i7- 8700k, or any other chip and the R2 will continue being 13.3% faster than R1 and the R1 will continue being 11.8% slower than R2.

"Faster than" and "slower than" are relational quantities, so they don't vary when changing the reference.

In the CPCHardware review the 2700X was 8.7% faster than the 1800X, NOT 14% as some sites are claiming. Add computerbase.de to the former list of sites that don't know basic math, because computerbase claims that 2700X was 15% faster than 1800X in applications, which is obviously false.
 
Last edited:

juanrga

2[H]4U
Joined
Feb 22, 2017
Messages
2,804
I don't have an issue with a cheap A320 board being used, I don't agree that it doesn't matter much, at first they say 1-2% then it goes to 5-8% and if we take the 8700K's "cheat mode activated" it scored up to 33% higher in CB which is significant. I assume the Asus X470 boards will have highest end quality caps, chokes and VRM's which will push the sustained boost modes higher. Canard PC is face value only, there is no reported clocks or methodology used, they could be throwing the worst case scenario which putting a 2000X chip on a A320 is, you will get choked memory performance, you will get turbo limits and no OCing, basically testing poorest quality equipment where they probably put Intel on the best and may even be running MCE.

At this point I will wait for more reputable reviewers to handle the matter, not Frenchies with a tendency to break NDA at will.

Objections:

1) Canard PC are excluded from AMD's testing program due to repeated NDA violations, this i confirmed with someone internally.
2) The chips are likely black market ES chips with old micro code to ensure they work on basic platforms
3) no methodology provided.

So they will break NDA, use ES chips to signify retail chips, and have no discernible methodology to their testing. Sounds like they are just a click bait site, willing to circumvent the norms of the industry to put sensationalist information out on the net for little more than profit. This kinda ranks them with the news sites that have "10 Famous People pictures you won't believe" articles, total trash.

Canard PC has as much credibility as the National Enquirer

(i) 1--2% is the performance improvement when replacing stock A320 mobo by a X470 mobo both on stock settings. 5--8% is the improvement when comparing stock vs auto-overclock.

(ii) Clocks are stock, so it is not needed to mention them on the graphs. Clocks are reported in the all the non-stock tests, like the test about latencies or the test about boost policies. Rest of metodologies such as applications tested or resolution and gaming settings is also given in the magazine.

(iii) Once again, the lack of overclocking is not a problem because they are testing stock chips. There is no MCE activated on Intel chips. All chips are working on stock settings. That is the reason why the 8700k was 160% whereas the 1800x was 161%. If MCE was enabled then the 8700k would score about 175%.

DZZSurJXcAEIqnQ.jpg


(iv) Obviously Canard didn't sign any NDA. So they didn't violate any NDA, and that is why they can ship the print magazine without violating any law.

(v) There is no ES for Pinnacle Ridge, so it is very difficult that Canard can test what doesn't exist. Also Canard wouldn't be confounded with sites as Guru3D. Guru3D use ES to signify retail chips, CanardPC doesn't. CanardPC mentions the use of ES when they test ES, as anyone minimally familiar with them know

771c63c2-75ad-4cd7-b2e1-3db9a7bf1fcd.png


Not only they mentioned explicitly the use of an engineering sample of Zen used in their magazine of 2016, but they gave the full identifier for the Zen sample tested then: 2D3151A2M88E4.

In this Zen+ review CanardPC has used retail chips of R5-2600, R5-2600X, and R7-2700X, which have been in the hands of some reviewers since past month. Below a pic of some Zen+ retail chips from computerbase.

Bg_ZziTa?format=jpg&name=600x314.jpg


(vi) CanardPC is a reputed site, with a tradition on providing information before anyone else. Those are the guys that leaked details about EPYC or Raven Ridge before anyone else. When they said in the 2016 that Raven Rdige would have 11CUs, many guys doubted that odd configuration, but it was correct.
 

Dermac

Limp Gawd
Joined
May 6, 2005
Messages
152
(v) There is no ES for Pinnacle Ridge, so it is very difficult that Canard can test what doesn't exist. Also Canard wouldn't be confounded with sites as Guru3D. Guru3D use ES to signify retail chips, CanardPC doesn't. CanardPC mentions the use of ES when they test ES, as anyone minimally familiar with them know

So a major line of chips from a manufacturer was not tested? Isn't an engineering sample used to test various stages of CPU development? So, you are saying that AMD hasn't tested the chips for an upcoming line of CPU's. Wait, what am I saying, of course they did because Juan says so.
 

Dermac

Limp Gawd
Joined
May 6, 2005
Messages
152
So where is the proof that AMD didn't test the Piunnacle Ridge CPU's, or is it just more made up 'facts'?
 

Pillars

Gawd
Joined
Jun 11, 2016
Messages
537
Hes referring to that one crap review like it's the Ryzen 2 Bible. I know to skip the 2 series and use my heavily discounted 1700 chip until 7nm comes. Don't need a single review to confirm this either :) Is my 8700K faster? Sure.. do I still love my Ryzen? Definitely.
 

OrangeKhrush

[H]ard|Gawd
Joined
Dec 15, 2016
Messages
1,673
Just because you don't sign a NDA, doesn't mean you are protected from the infringement policies, if it was that simple why would anyone sign a NDA? the answer is to get samples you need to sign and if you acquire through black market channels you still acquire the same title as the original signee.

As for retail, there are no retails as they are not in circulation, they are late samples with qualification bios's but they are still test samples even if 99% like the final product.
 

juanrga

2[H]4U
Joined
Feb 22, 2017
Messages
2,804
So a major line of chips from a manufacturer was not tested? Isn't an engineering sample used to test various stages of CPU development? So, you are saying that AMD hasn't tested the chips for an upcoming line of CPU's. Wait, what am I saying, of course they did because Juan says so.

You couldn't be more wrong. I am not saying what you pretend. Evidently, AMD and partners have tested Pinnacle Ridge. I even wrote about the QS of the 2600X in another thread. You would know if you actually read my posts and you were familiar with the terminology...

http://digiworthy.com/2017/12/19/aida64-ryzen-gen-2-cpus/

It appears that this website has seen evidence of the non-existent Pinnacle Ridge CPUs in December. What else we gonna make up Juan?

What evidence? A site mentioning another site where "AMD K18 ES" is mentioned in a comment in code? Where is stated that PR is K18? Where are those PR ES? No one has seen one. No OEM knows their existence. No one has leaked a single bit about any PR ES...

Just because you don't sign a NDA, doesn't mean you are protected from the infringement policies, if it was that simple why would anyone sign a NDA? the answer is to get samples you need to sign and if you acquire through black market channels you still acquire the same title as the original signee.

As for retail, there are no retails as they are not in circulation, they are late samples with qualification bios's but they are still test samples even if 99% like the final product.

If you don't sign any NDA then you don't violate any NDA. Not only CPCHardware didn't break any NDA, but they have given me the name of AMD representative in France for I can ask him directly about what you are saying about them.

PR has been in production for a while. The retail chips are shipping to stores. A handful of the retail chips are in the hands of reviewers (as those in the above image). They aren't test samples.

CPCHardware has confirmed that the BIOS used in the review "is a retail, final BIOS".
 
Last edited:

juanrga

2[H]4U
Joined
Feb 22, 2017
Messages
2,804
"They claimed that X470 is practically identical to X370 except for a few cosmetic changes." There is a rumor that X470 is simply a rebranded X370 with an improved BIOS. I will ask about it
 

juanrga

2[H]4U
Joined
Feb 22, 2017
Messages
2,804
and the fact that it can properly run the new chips to their full potential rather than crippling them...

According to CPCHardware and Overclocking made in france, replacing the A320 mobo by X470 only brings 1% or 2% moar performance. That is no "crippling" by any bit of imagination.
 

Rvenger

2[H]4U
Joined
Sep 12, 2012
Messages
2,992
According to CPCHardware and Overclocking made in france, replacing the A320 mobo by X470 only brings 1% or 2% moar performance. That is no "crippling" by any bit of imagination.


There's no way the VRM is overheating causing throttling on that 105w CPU. No way man, A320 is robust. /s



Here let me give you this quote.

"The rumored processor was supposedly tested on an entry-level A320 motherboard that doesn’t support XFR2. With that in mind, this chip could perform even better – or at least, achieve even higher core frequencies – with a X370 or X470 motherboard." https://www.techradar.com/news/the-first-amd-ryzen-7-2700x-review-reveals-brute-force-performance


A320 boards have 4 phases. 4+2 which means 4 for the CPU and 2 for the memory. NO way in hell that is not throttling. Sorry bud.
 
Last edited:

sirmonkey1985

[H]ard|DCer of the Month - July 2010
Joined
Sep 13, 2008
Messages
22,287
There's no way the VRM is overheating causing throttling on that 105w CPU. No way man, A320 is robust. /s



Here let me give you this quote.

"The rumored processor was supposedly tested on an entry-level A320 motherboard that doesn’t support XFR2. With that in mind, this chip could perform even better – or at least, achieve even higher core frequencies – with a X370 or X470 motherboard." https://www.techradar.com/news/the-first-amd-ryzen-7-2700x-review-reveals-brute-force-performance


A320 boards have 4 phases. 4+2 which means 4 for the CPU and 2 for the memory. NO way in hell that is not throttling. Sorry bud.

even some b350 boards had vrm issues running the 1700x with even a slight overclock.
 

__hollywood|meow

[H]ard|Gawd
Joined
Feb 20, 2006
Messages
1,500
so juan, serious question for ya: after a years time & over seventeen hundred flailing posts, have you begun to realize that nobody here is buying all the substanceless circumlocution? [H] is a forum of hardcore hardware enthusiasts, plenty of whom are intimately familiar with technical discussion. people can follow your argumentation. when folks rightly point out facile reasoning, that is not the time to double down.

back to lurking for me! dont bother responding. yall have a nice chat. ill wait until [H], guru3d, TPU, anand, et al get their hands on zen+ & we have more concrete information regarding the topic
 

juanrga

2[H]4U
Joined
Feb 22, 2017
Messages
2,804
There's no way the VRM is overheating causing throttling on that 105w CPU. No way man, A320 is robust. /s



Here let me give you this quote.

"The rumored processor was supposedly tested on an entry-level A320 motherboard that doesn’t support XFR2. With that in mind, this chip could perform even better – or at least, achieve even higher core frequencies – with a X370 or X470 motherboard." https://www.techradar.com/news/the-first-amd-ryzen-7-2700x-review-reveals-brute-force-performance


A320 boards have 4 phases. 4+2 which means 4 for the CPU and 2 for the memory. NO way in hell that is not throttling. Sorry bud.

The same board was used for all the 1000 and 2000 series chips tested. The 2600 was 7.6% faster than 1600 in applications. The 2600 is rated at 65W. Which is here the excuse?

According to AMD the A320 mobos support XFR2. Are you suggesting that AMD, CPC, and Overclocking made in france are wrong? And isn't this the same techradar site I mentioned in #192 as part of the collection of sites are giving incorrect information about the performance of Pinnacle?

According to AMD, the 2700X is about 5% faster than the 1800X at 1080p. According to CPC the 2700X is about 3.4% faster than the 1800X at 1400p. Both results are compatible. So where is the problem? Where is the "crippling"?

Isn't the Gigabyte 4+3?
 
Last edited:

ManofGod

[H]F Junkie
Joined
Oct 4, 2007
Messages
12,719
The same board was used for all the 1000 and 2000 series chips tested. The 2600 was 7.6% faster than 1600 in applications. The 2600 is rated at 65W. Which is here the excuse?

According to AMD the A320 mobos support XFR2. Are you suggesting that AMD, CPC, and Overclocking made in france are wrong? And isn't this the same techradar site I mentioned in #192 as part of the collection of sites are giving incorrect information about the performance of Pinnacle?

According to AMD, the 2700X is about 5% faster than the 1800X at 1080p. According to CPC the 2700X is about 3.4% faster than the 1800X at 1400p. Both results are compatible. So where is the problem? Where is the "crippling"?

Isn't the Gigabyte 4+3?

I see you are still attempting to bring AMD down. I would have thought after the smackdown you received from Linus, you would be a bit more humble but, I guess not. AMD is bringing it and Intel is responding, got to love competition. :)
 

ManofGod

[H]F Junkie
Joined
Oct 4, 2007
Messages
12,719
so juan, serious question for ya: after a years time & over seventeen hundred flailing posts, have you begun to realize that nobody here is buying all the substanceless circumlocution? [H] is a forum of hardcore hardware enthusiasts, plenty of whom are intimately familiar with technical discussion. people can follow your argumentation. when folks rightly point out facile reasoning, that is not the time to double down.

back to lurking for me! dont bother responding. yall have a nice chat. ill wait until [H], guru3d, TPU, anand, et al get their hands on zen+ & we have more concrete information regarding the topic

I will respond for one reason, are you going to be upgrading that machine in your sig to Zen? Admittedly, that is still a good, fast build but, if you are not upgrading soon, you have more self control than I do. :D
 

N4CR

Supreme [H]ardness
Joined
Oct 17, 2011
Messages
4,948
The more I repeat things the more truth there is, it's the thread.
The only reason someone is still allowed to 'post' here is because if they didn't, the amd section would have half the post counts it does.
 

OrangeKhrush

[H]ard|Gawd
Joined
Dec 15, 2016
Messages
1,673
I just laugh how he thinks AMD graciously give out samples to people without signing any NDA, I don't need no fake french contact with AMD to know it is more CPC BS. They acquired a sample off the black market which is why they are vague in their reviews, no screen shot validations in certain suites that will tell you what the CPU is, it isn't the first time they have been cutting corners. They can't even get power readings right. The only thing the frenchies are good for is surrendering.

No difference between a A320 and X470, okay awesome so I can save a ton of money and buy the A320 to get the same features and performance, sounds about right Juan. I mean you are a genius.
 

OrangeKhrush

[H]ard|Gawd
Joined
Dec 15, 2016
Messages
1,673
One of the major features of the B350 and X370 is the VRM's, the X370 have 12+ and the B350 6-8 phase, there are none on the A320 meaning a 1800X would likely throttle hard under load, doubt it will even hit a turbo frequency without the board heating up and throttling.
 

Rvenger

2[H]4U
Joined
Sep 12, 2012
Messages
2,992
The same board was used for all the 1000 and 2000 series chips tested. The 2600 was 7.6% faster than 1600 in applications. The 2600 is rated at 65W. Which is here the excuse?

According to AMD the A320 mobos support XFR2. Are you suggesting that AMD, CPC, and Overclocking made in france are wrong? And isn't this the same techradar site I mentioned in #192 as part of the collection of sites are giving incorrect information about the performance of Pinnacle?

According to AMD, the 2700X is about 5% faster than the 1800X at 1080p. According to CPC the 2700X is about 3.4% faster than the 1800X at 1400p. Both results are compatible. So where is the problem? Where is the "crippling"?

Isn't the Gigabyte 4+3?

No excuse, these A320s were not designed for any 6 - 8 core configurations. Let me be clear even though the TDP of the CPU is 65w, XFR can increase the TDP above that 65w level. My Ryzen 1700 at stock did not use 65w. If I turned off core boost, then it would run around 65w.

Premature bios with a board not designed for these chips in the first place is probably common sense at this point.

Wow, big difference. 3 phases for RAM which is a waste. Still only 4 for CPU which you would need to run a fan over those VRMs to keep them cool. X370 board that can't handle 8 cores without active cooling -


I am going to wait and see instead of making premature predictions that are off base. According to your findings these are the bulldozer of CPUs... and I will never forget the Torvolds moment. ;)
 

Dermac

Limp Gawd
Joined
May 6, 2005
Messages
152
So, CanardPC theorizes that they are justified in using the crappiest motherboard for testing because of "automatic omnipresent cheats" that boards designed for high core count chips may provide. They theorize that using a proper board will only give about a 1%-2% boost to the scores (i.e. without automatic omnipresent cheats enabled) and about 4% or more with the automatic omnipresent cheats enabled. Common sense would tell me that if there is any difference, then there is an outside effect negatively impacting the results (i.e. CPU throttling could be a cause.) If there is throttling happening, that means they aren't testing in a proper test set-up. If they aren't testing in a proper test set-up, why should anybody care what they say/print?

In response to the NDA stuff, it is true CanardPC didn't violate an NDA they were never given an option to sign. But NDA violations will transcend time, meaning they could have committed a violation in the past that would make it so that AMD would never give them an NDA option again, meaning that AMD won't give CanardPC the tools and the NDA in order for them to produce their content for their site, within non-combative industry guidelines addressed in the NDA. I would assume that AMD would want a major French tech publication to test their products because French speaking people buy CPU's too, why is it that CanardPC didn't get the option to sign? While you are contacting your 'sources' Juan, might as well tweet CanardPC on why AMD doesn't think they are worthy to review their products with their support. Although I don't know for sure, I'm theorizing they violated an NDA in the past and got them blacklisted by AMD. Considering their non-existent methodology of testing, I don't blame AMD for having common sense.
 

juanrga

2[H]4U
Joined
Feb 22, 2017
Messages
2,804
I see you are still attempting to bring AMD down. I would have thought after the smackdown you received from Linus, you would be a bit more humble but, I guess not.

So mentioning that AMD refutes what techradar says about XFR2, mentioning that CPC review agrees with internal tests made by AMD, and asking if the GB mobo is 4+3 is... "attempting to bring AMD down"?

And don't forget that all what I wrote to Linus was correct. :D

I just laugh how he thinks AMD graciously give out samples to people without signing any NDA, I don't need no fake french contact with AMD to know it is more CPC BS. They acquired a sample off the black market which is why they are vague in their reviews, no screen shot validations in certain suites that will tell you what the CPU is, it isn't the first time they have been cutting corners. They can't even get power readings right. The only thing the frenchies are good for is surrendering.

No difference between a A320 and X470, okay awesome so I can save a ton of money and buy the A320 to get the same features and performance, sounds about right Juan. I mean you are a genius.

I don't think what you imagine. It is obvious to me that they didn't got the hardware from AMD.

No one said that there is "no difference between a A320 and X470". You are just making up things again, instead reading what was written in this thread. I even posted the measured performance variation between A320 and X470. Go figure!

No excuse, these A320s were not designed for any 6 - 8 core configurations. Let me be clear even though the TDP of the CPU is 65w, XFR can increase the TDP above that 65w level. My Ryzen 1700 at stock did not use 65w. If I turned off core boost, then it would run around 65w.

XFR is an automatic overclock feature that pushes clocks above the stock TDP, when there are extra cooling installed. Also your 1700 at stock isn't a 65W chip. The real TDP is 90W.


So, CanardPC theorizes that they are justified in using the crappiest motherboard for testing because of "automatic omnipresent cheats" that boards designed for high core count chips may provide. They theorize that using a proper board will only give about a 1%-2% boost to the scores (i.e. without automatic omnipresent cheats enabled) and about 4% or more with the automatic omnipresent cheats enabled. Common sense would tell me that if there is any difference, then there is an outside effect negatively impacting the results (i.e. CPU throttling could be a cause.) If there is throttling happening, that means they aren't testing in a proper test set-up. If they aren't testing in a proper test set-up, why should anybody care what they say/print?

In response to the NDA stuff, it is true CanardPC didn't violate an NDA they were never given an option to sign. But NDA violations will transcend time, meaning they could have committed a violation in the past that would make it so that AMD would never give them an NDA option again, meaning that AMD won't give CanardPC the tools and the NDA in order for them to produce their content for their site, within non-combative industry guidelines addressed in the NDA. I would assume that AMD would want a major French tech publication to test their products because French speaking people buy CPU's too, why is it that CanardPC didn't get the option to sign? While you are contacting your 'sources' Juan, might as well tweet CanardPC on why AMD doesn't think they are worthy to review their products with their support. Although I don't know for sure, I'm theorizing they violated an NDA in the past and got them blacklisted by AMD. Considering their non-existent methodology of testing, I don't blame AMD for having common sense.

They don't theorize. They benchmarked performance in different boards, including X470. It is all in page 5 of this thread.

About NDA violations:

CanardPC said:
Total bullshit. We never had any “agreement” with AMD and we never “breached” any NDA because we don’t sign them.

CanardPC said:
PS: AMD’s PR in France is @Julien_Jay, you can ask him directly when we signed an NDA with AMD for the last time
1f601.png
 
Last edited:

sirmonkey1985

[H]ard|DCer of the Month - July 2010
Joined
Sep 13, 2008
Messages
22,287
So much debate over early half assed benchmarks.. wait a few weeks and get the real thorough reviews and save yourselves the headaches of arguing over this stuff.

nah we're just bored and waiting for the NDA to expire.. :)
 

OrangeKhrush

[H]ard|Gawd
Joined
Dec 15, 2016
Messages
1,673
So mentioning that AMD refutes what techradar says about XFR2, mentioning that CPC review agrees with internal tests made by AMD, and asking if the GB mobo is 4+3 is... "attempting to bring AMD down"?

And don't forget that all what I wrote to Linus was correct. :D



I don't think what you imagine. It is obvious to me that they didn't got the hardware from AMD.

No one said that there is "no difference between a A320 and X470". You are just making up things again, instead reading what was written in this thread. I even posted the measured performance variation between A320 and X470. Go figure!



XFR is an automatic overclock feature that pushes clocks above the stock TDP, when there are extra cooling installed. Also your 1700 at stock isn't a 65W chip. The real TDP is 90W.




They don't theorize. They benchmarked performance in different boards, including X470. It is all in page 5 of this thread.

About NDA violations:

So basically Canard PC support Black Market products, they got their samples by buying them from a person that may have also been in the cycle of buying ES samples. One term of NDA is that you cannot sell ES chips. They have no integrity and their testing reasoning is full of contradictions. They claim no difference but then don't post X470 scores when there a re clear reviews out there that show the vast performance difference going A320 to X370 notably load line calibration which is not on A320 boards and chips throttling because the VRM's overheat. There is no justification to put AMD's flagship parts on a entry level board, the 8700K reviews were done on the best Z370 motherboards on the market, not a H110.
 

Derfnofred

Gawd
Joined
Dec 11, 2009
Messages
606
ES samples, depending on where they are in the development cycle, shouldn't be too far off end-performance. This has been shown on plenty of occasions and makes sense when you start breaking down the whole development cycle. At engineering sampling, the yield rate may be absolutely atrocious, but the chips that do get validated for use are generally okay.

*that said* taking any one review, especially one this with so many holes in its methodologies, as any sort of gospel gets you exactly what you deserve. Canard PC may be entirely correct that you will get really mediocre performance using a high-power chip in a budget board incapable of feeding it the power it needs! Doesn't really help anyone except to say, "don't do that". Or there may be other A320 boards with better ability to handle the heat (but still silly to pair with). The folks stating "let's wait for multiple reputable sites to give us their reviews" are 100% correct.
 

DuronBurgerMan

[H]ard|Gawd
Joined
Mar 13, 2017
Messages
1,340
1. It's a crappy review on a platform/configuration that no one is likely to actually build.
2. The 2700X probably isn't going to be MUCH faster on a proper configuration (but it will be at least a bit faster).

Both statements can be true.
 

Master_shake_

Fully [H]
Joined
Apr 9, 2012
Messages
17,794
So mentioning that AMD refutes what techradar says about XFR2, mentioning that CPC review agrees with internal tests made by AMD, and asking if the GB mobo is 4+3 is... "attempting to bring AMD down"?

And don't forget that all what I wrote to Linus was correct. :D



I don't think what you imagine. It is obvious to me that they didn't got the hardware from AMD.

No one said that there is "no difference between a A320 and X470". You are just making up things again, instead reading what was written in this thread. I even posted the measured performance variation between A320 and X470. Go figure!



XFR is an automatic overclock feature that pushes clocks above the stock TDP, when there are extra cooling installed. Also your 1700 at stock isn't a 65W chip. The real TDP is 90W.




They don't theorize. They benchmarked performance in different boards, including X470. It is all in page 5 of this thread.

About NDA violations:


the real TDP of the 7700k is 152watts juan.
 

Rvenger

2[H]4U
Joined
Sep 12, 2012
Messages
2,992
So mentioning that AMD refutes what techradar says about XFR2, mentioning that CPC review agrees with internal tests made by AMD, and asking if the GB mobo is 4+3 is... "attempting to bring AMD down"?

And don't forget that all what I wrote to Linus was correct. :D



I don't think what you imagine. It is obvious to me that they didn't got the hardware from AMD.

No one said that there is "no difference between a A320 and X470". You are just making up things again, instead reading what was written in this thread. I even posted the measured performance variation between A320 and X470. Go figure!



XFR is an automatic overclock feature that pushes clocks above the stock TDP, when there are extra cooling installed. Also your 1700 at stock isn't a 65W chip. The real TDP is 90W.




They don't theorize. They benchmarked performance in different boards, including X470. It is all in page 5 of this thread.

About NDA violations:


That's basically what I said Sherlock.
 
Top