Are the new Ryzen supposed to by any better in overclock? Counterfeit CPU?

[DestroyFX]

Hi there, I do have an Asus B350-Prime plus and I tested 4 CPU on it including that latest 2700X and for that later i'm really not impressed.

1 : R7 1700 (Could do 4Ghz max at 1.42V). It's was a defect one like most/all Ryzen sold for the first six month so AMD replaced it.
2 : R3 1200 (Could do 4Ghz at 1.4V ish, could have probably OC more but it was just for the waiting of RMA of 1.)
3 : R7 1700 from RMA of 1 (Could do 4Ghz at 1.35V, I did not try to up it more than that, 4Ghz was enough for me)
4 : R7 2700X (Can do 4.1Ghz at 1.41V, can't really do more without going at the 1.45V range and I don't like that)

So here are the questions :
a - Is the new Ryzen2 are really better? I mean, it supposed to support higher memory speed but I still have the exact same issue with my TridentZ 3200, XMP cause memory errors and don't work fine with the default 1.35V. I do have to disable XMP and set Voltage to 1.45V and timming to 14-16-16-16-32 to get it right (these are *tighter than the XMP at 16-18-18-18-38*). Same exact setting that was required on my two other R7 1700.

b - The CPU overclock look worse than my old R7 1700 I got back from RMA. As my R7 1700 was doing 4Ghz fine at 1.35V, I'm sure it can do 4.1 at similar voltage, I did not try to up it higher, maybe I should... The new 2700X really like voltage, what's is supposed to be the limit? From what I read I should not touch 1.45? for 24/24//7/7.

c - The fancy turbo stuff of the 2700X on the youtubes review where all getting like 4.3Ghz for one core and over 4Ghz (like 4.1 even) for all core. Mine do ONLY 3.95 for one core and drop to 3.7 ish for all cores.

Final question : Could I have gotten a counterfeit CPU, like a re-branded 1800X? Or it's just really low binning?. I'm considering putting the R7 1700 I got back from RMA in and selling the 2700X.

 

[bigtallanddopey]

To be honest there isn’t a massive difference between the two from what I have seen, especially when manually overclocking. Maybe an extra 200Mhz although you aren’t seeing that.

The turbo boost does seem to be a little better on the newer motherboards. You’re running a budget board anyway from the first ryzen. The new x470 boards do seem to show the better results for turbo boost.

You are experiencing all the reasons I saw no need to upgrade from my 1600. The performance leap wasn’t big enough. Most people have been reporting better ram comparability though. Although again this maybe as much down to your b350 board.

I would return and stick with what you have and save the pennies.

 

[Woomack]

I've seen many results at 4.2GHz but all depends on luck. I guess you can see anything between 4.1GHz and 4.3GHz on ambient temps while most will be close to 4.2GHz.
My 1700X could make 4.05GHz at 1.45V max, 2200G could make 4.0GHz at 1.40V and 1920X can make 4.1GHz at 1.42V.

 

[mackerel]

As my R7 1700 was doing 4Ghz fine at 1.35V, I'm sure it can do 4.1 at similar voltage, I did not try to up it higher, maybe I should...

You can't make that assumption. The 1700 are already close to their limit around 4 GHz. Pushing the next step would require more voltage. How much more, depends on your sample. I didn't have luck, even with silly voltages I can't get mine to bench stable at 4.1.

As for your 2700X, what mobo? What cooling?

 

[DestroyFX]

You can't make that assumption. The 1700 are already close to their limit around 4 GHz. Pushing the next step would require more voltage. How much more, depends on your sample. I didn't have luck, even with silly voltages I can't get mine to bench stable at 4.1.

As for your 2700X, what mobo? What cooling?

The mobo is an Asus B350-Prime plus

The CPU do share 480mm of radiator surface with a Radeon RX480.

 

[mackerel]

Cooling should be ok then, but from memory the 1st gen chipsets didn't allow as good a turbo as the 2nd gen ones. Maybe that could explain the stock performance?

 

[Johan45]

Turbo works the same regardless of chipset. You won't get the "enhanced" XFR on 3xx chipsets. That's the extended turbo function which will boost all cores up to 4.2+. When they say higher memory support that means from 2666 > 2933 Mhz. Your memory is still going to be an issue. I tried Hynix stcks and they were just a PITA. Needed more voltage and much harder to stabilize an overclock.

 

[mackerel]

I was using "turbo" to refer to anything above base speed regardless of the exact mechanism. Could this explain part c in the original post?

 

[DestroyFX]

I was using "turbo" to refer to anything above base speed regardless of the exact mechanism. Could this explain part c in the original post?

Part c, I was initially wondering if I got a counterfeit as the turbo look more like Ryzen 1800X range and I got the CPU for an abnormally low price for Mexico. But it might be chipset related for the XFR part.. I was just expecting better turbo and clock.

@Johan45 : For the memory, I do have Hynix too, They are rated for 1.35V 16-18-18-18-38 with it's XMP but i'm OK with 1.45V 14-16-16-16-32, I use these settings since April 2017. It's like if it's hard to go up in clock only (need moar memory voltage ans SoC Voltage for 3200 but then I can just put tighter timing once I got the 3200 working.

 

[LRG5]

I got XFR.2 working on my x370 Taich. check it out and this is standard Bios P-460 and works.







DestroyFX is that mother Board rated for the power need for a Ryzen7 2700x ?

 

[DestroyFX]

I got XFR.2 working on my x370 Taich. check it out and this is standard Bios P-460 and works.




View attachment 198581


DestroyFX is that mother Board rated for the power need for a Ryzen7 2700x ?

View attachment 198583

Thanks, I took note and found these settings in my B350 BIOS so I will test if I decide to return to stock+turbo.. They did also mention on the different review that XFR2 was actually working on x370 Board.

For the Maximum voltage by Asus, i'm not sure if peoples should use that as a reference, it say 1.3V Max for Soc Ambiant. It's kill the Ryzen Soc Voltage. As for the others, 1.8 & 1.05, I did not touch them on my side, yet.

 

[Woomack]

Officially more than 1.2V SOC is not safe but have you heard about anyone killing CPU while setting more ? I was checking my CPU up to 1.35V but I haven't seen any difference above 1.17V or something near. Even at auto, motherboard sets ~1.14V what is enough for stable 3733 in quad channel. Well it's on TR but IMC and almost everything else is the same as in other Ryzens.

 

[DestroyFX]

Officially more than 1.2V SOC is not safe but have you heard about anyone killing CPU while setting more ? I was checking my CPU up to 1.35V but I haven't seen any difference above 1.17V or something near. Even at auto, motherboard sets ~1.14V what is enough for stable 3733 in quad channel. Well it's on TR but IMC and almost everything else is the same as in other Ryzens.

I have read on Reddit about more than one dead Ryzen chip with only 1.2x V SOC (I remember about two dead chips). I don't remember the specific oc and other settings, i did look at that like 11 month ago..

On my 1700's, setting SOC to 1.15 did help a little for memory stability.
On my 2700x, putting anything but default (1.1V) for SOC produce errors in memtest.

 

[Woomack]

So it's not 100% confirmed that SOC killed these processors or something else. I'm not recommending higher SOC as for most users it won't help in anything and since someone tested that can damage CPU then I guess it can at least degrade IMC over time.
Now I run 4x8GB@3600/Micron IC with SOC at auto what gives about 1.14V and it's stable for couple of days. Before I was testing 4x8GB Samsung @3733 for couple of days at 1.2V SOC ( I just set 1.2V manually but later noticed that auto so lower voltage is also stable ). I set 1.35V when I was checking if I can set 4000 on my board but regardless of voltages or timings I was always getting d5 error so I back to 3600-3866 tests.

 

[storm-chaser]

Hi there, I do have an Asus B350-Prime plus and I tested 4 CPU on it including that latest 2700X and for that later i'm really not impressed.

Final question : Could I have gotten a counterfeit CPU, like a re-branded 1800X? Or it's just really low binning?. I'm considering putting the R7 1700 I got back from RMA in and selling the 2700X.

Counterfeit CPU? Not in a million years. I've seen this myth make the rounds on overclocking sites and computer related forums as of late and it seems like more and more its gaining traction and destroying buyer confidence in the over seas market. I can tell you for a fact nobody in China is making counterfeit CPUs. Yes, there are some retail box CPUs sold on Amazon coming from China that contain "dummy" processors that are literally hunks of junk, but that's a totally different situation entirely. The retail box scam typically revolves around using an old Celeron or LGA 775 chip with Ryzen or i7 etching contained in a nearly perfect counterfeit AMD or Intel retail box. They package the chip to make it appear correct in the images to the unsuspecting buyer. And the gig is up as soon as the end user open's the box. Obviously, an LGA 775 chip isn't going to fit in a 1331-pin lidded micro-PGA AM4 package. Whole different scam entirely versus counterfeiting the actual CPU, with all its intricate working components. And this is nearly impossible when you start to realize what goes in to the actual process of manufacturing a CPU. Counterfeiting a CPU and actually getting it to fool the end user would be akin to counterfeiting Windows 10 (without the source code) and getting it to actually pass the Windows Validation as genuine and have it update drivers across diverse hardware. Not gonna happen. I know this is a hardware vs software comparison but it still highlights the complexities that would stymie even the most ingenuous fraudster.

 

[LRG5]

check the rating, some of the motherboard VRM can not handle 2700x, iT pull a lot of power, More than a 1700!!!! And the setting for XFR2.0 maybe different for that MB.

For BIOS 470 PPT 200000 , TDC 11400 , FDC 16800.

 

[DestroyFX]

Counterfeit CPU? Not in a million years. I've seen this myth make the rounds on overclocking sites and computer related forums as of late and it seems like more and more its gaining traction and destroying buyer confidence in the over seas market. I can tell you for a fact nobody in China is making counterfeit CPUs. Yes, there are some retail box CPUs sold on Amazon coming from China that contain "dummy" processors that are literally hunks of junk, but that's a totally different situation entirely. The retail box scam typically revolves around using an old Celeron or LGA 775 chip with Ryzen or i7 etching contained in a nearly perfect counterfeit AMD or Intel retail box. They package the chip to make it appear correct in the images to the unsuspecting buyer. And the gig is up as soon as the end user open's the box. Obviously, an LGA 775 chip isn't going to fit in a 1331-pin lidded micro-PGA AM4 package. Whole different scam entirely versus counterfeiting the actual CPU, with all its intricate working components. And this is nearly impossible when you start to realize what goes in to the actual process of manufacturing a CPU. Counterfeiting a CPU and actually getting it to fool the end user would be akin to counterfeiting Windows 10 (without the source code) and getting it to actually pass the Windows Validation as genuine and have it update drivers across diverse hardware. Not gonna happen. I know this is a hardware vs software comparison but it still highlights the complexities that would stymie even the most ingenuous fraudster.

Counterfeiting CPU don't necessarily mean to manufacture CPU, It could mean to sell CPU as a specific retail model while not supposed to be sold at all.
In the china market you could buy Kaby Lake and skylake CPU for 75$, they where sold as prototype/reject/manufacturing sample on aliexpress and they do work fine and are not counterfeit because they are not sold label as the final product in a nice retail box. The only difference compared to the final Intel Kaby Lake / Skylake are the clock, max overclock and turbo frequency. What do you think append to early units and prototypes? do you really think they are all scrapped? Unfortunately, the price of these CPUs skyrocked because peoples got too noisy about them.

While I think I do have a genuine product but issue with the board or really got unlucky at the silicon lottery, it have two of the three symptoms of these CPU sold on aliexpress, Slower turbo and crappy overclockability, but mine got a proper retail box and that would make the difference with a CPU "Sold without a box/cooler as manufacturing sample" and would make mine a counterfeit product assuming it escaped the factory early or without passing all the tests.

Also they don't have to come from China. If a Mexican computer assembler decide to open Ryzen boxes from under to avoid damaging the seal, and then put these CPU in the computer they sell.. Then after they take CPU they got cheaper from the same factory or aliexpress (like manufacturing sample with lower turbo frequency), and put them in the boxes they did open earlier to sell them back on the local market as retail product, they are in fact selling counterfeit units even if they all came from AMD and from the same fab.

I did order mine on launch day + 1, I got it from the Mexican craiglist equivalent, Mercadolibre and it was cheaper than all the unit sold in the actual computer shops (Usually they are more expensive on mercadolibre than in the shops). But it does not necessarily mean it's counterfeit, I know, it could be simply someone who got one and did not have enough money to eat so try to get rid of it asap.... It could have also felt from the truck (stolen).

 

[LRG5]

All power setting at stock, Ram set 3466 with XFR2.0 On. and Air cooling.

 

[Alaric]

Counterfeit CPU? Not in a million years. I've seen this myth make the rounds on overclocking sites and computer related forums as of late and it seems like more and more its gaining traction and destroying buyer confidence in the over seas market. I can tell you for a fact nobody in China is making counterfeit CPUs. Yes, there are some retail box CPUs sold on Amazon coming from China that contain "dummy" processors that are literally hunks of junk, but that's a totally different situation entirely. The retail box scam typically revolves around using an old Celeron or LGA 775 chip with Ryzen or i7 etching contained in a nearly perfect counterfeit AMD or Intel retail box. They package the chip to make it appear correct in the images to the unsuspecting buyer. And the gig is up as soon as the end user open's the box. Obviously, an LGA 775 chip isn't going to fit in a 1331-pin lidded micro-PGA AM4 package. Whole different scam entirely versus counterfeiting the actual CPU, with all its intricate working components. And this is nearly impossible when you start to realize what goes in to the actual process of manufacturing a CPU. Counterfeiting a CPU and actually getting it to fool the end user would be akin to counterfeiting Windows 10 (without the source code) and getting it to actually pass the Windows Validation as genuine and have it update drivers across diverse hardware. Not gonna happen. I know this is a hardware vs software comparison but it still highlights the complexities that would stymie even the most ingenuous fraudster.

Or silk screening i7 info on an old i3...

 

[storm-chaser]

Or silk screening i7 info on an old i3...

Sure. Someone could do that with an old I3 and make a few bucks and it would be considered "counterfeit" under the normal definition. I'm just hearing more and more about how counterfeit CPUs run hot and don't overclock as well as their genuine counterpart, implying and propagating the myth that there are in fact back alley factories in China with tooling processes capable of engineering a late model CPU from the ground up and having them essentially run and "pass" for good on the average, unsuspecting buyer. CPU fraud does happen but never to the depth of actually producing a complete knockoff chip in house. These fake chips (if they even exist in the real world) would never and could never deceive programs like HWInfo, CPU-Z or AIDA64. To put it another way: Even if you have been defrauded you are still running a genuine Intel or AMD processor, it just may not be the high end model that you thought you were getting. Use aforementioned programs to determine exactly what chip is under the hood.

Counterfeiting CPU don't necessarily mean to manufacture CPU, It could mean to sell CPU as a specific retail model while not supposed to be sold at all.
In the china market you could buy Kaby Lake and skylake CPU for 75$, they where sold as prototype/reject/manufacturing sample on aliexpress and they do work fine and are not counterfeit because they are not sold label as the final product in a nice retail box. The only difference compared to the final Intel Kaby Lake / Skylake are the clock, max overclock and turbo frequency. What do you think append to early units and prototypes? do you really think they are all scrapped? Unfortunately, the price of these CPUs skyrocked because peoples got too noisy about them.

While I think I do have a genuine product but issue with the board or really got unlucky at the silicon lottery, it have two of the three symptoms of these CPU sold on aliexpress, Slower turbo and crappy overclockability, but mine got a proper retail box and that would make the difference with a CPU "Sold without a box/cooler as manufacturing sample" and would make mine a counterfeit product assuming it escaped the factory early or without passing all the tests.

Also they don't have to come from China. If a Mexican computer assembler decide to open Ryzen boxes from under to avoid damaging the seal, and then put these CPU in the computer they sell.. Then after they take CPU they got cheaper from the same factory or aliexpress (like manufacturing sample with lower turbo frequency), and put them in the boxes they did open earlier to sell them back on the local market as retail product, they are in fact selling counterfeit units even if they all came from AMD and from the same fab.

I did order mine on launch day + 1, I got it from the Mexican craiglist equivalent, Mercadolibre and it was cheaper than all the unit sold in the actual computer shops (Usually they are more expensive on mercadolibre than in the shops). But it does not necessarily mean it's counterfeit, I know, it could be simply someone who got one and did not have enough money to eat so try to get rid of it asap.... It could have also felt from the truck (stolen).

Sure, there are plenty of examples of fraudulent activity relative to CPU sales. You described a few of the underhanded methods used in CPU counterfeiting in your post above. Just for some reason, the use of the term "counterfeiting" in the CPU industry breeds hysteria and In my opinion is not specific enough to address the main methods of fraud that go on in the sale of CPUs. Nothing against you or your claim at all, sorry if I cam off as harsh. I was just speaking in generalizations. However, if your CPU was an engineering sample you would be able to discover that using CPUz. As for an early prototype? CPU stepping determination is generally the best way to see where your specific CPU falls in terms of the production process. Remember, fraudsters may be able to deceive, but only from the outside. From the inside you will be able to read the chip for exactly what it is. Sorry of that was confusing I did my best.