1st Ryzen CPU Review Leaked

[JustReason]

Damn I almost mistaken you for an intel fanboi until I saw the facts.

he is, believe it. The actual BF1 demo both ran about same with the AMD having lower minimums, AMD getting in the 50s intel 60s for minimums. Both were reported to run 70s. Add lack of boost for AMD then they could likely be dead even.

 

[razor1]

And should we take your word for it, over a guy that was in front of the computers at the events? Keep in mind this was a GPU limited scenario too.

Any case, lets just wait and see what happens at the end, just a week and half away now......

 

[JustReason]

And should we take your word for it, over a guy that was in front of the computers at the events?

Any case, lets just wait and see what happens at the end, just a week and half away now......

I was speaking of what one of the people there reported. I am not wasting my life away trying to pick apart a video that will never be conclusive.

 

[razor1]

ok, then why is Ryzen exhibiting CPU limitations and can you say the other CPU is too or the GPU stopping it? Can't can we? So you conclusion is just as invalid as his lol.

 

[Shintai]

he is, believe it. The actual BF1 demo both ran about same with the AMD having lower minimums, AMD getting in the 50s intel 60s for minimums. Both were reported to run 70s. Add lack of boost for AMD then they could likely be dead even.

Minimums wasn't reported, average FPS was.

 

[razor1]

I was speaking of what one of the people there reported. I am not wasting my life away trying to pick apart a video that will never be conclusive.

Which person talked about mins? Can you link that please or give reference to that?

 

[Child of Wonder]

NERD FIGHT!!!!!!!!!

 

[OrangeKhrush]

http://www.tweaktown.com/guides/748...-intel-skylake-overclocking-guide/index4.html

The graph is easy to understand, you seem to fail understanding it. so lets work though this like its the first day of prep school.

Column 1 is CPU overclocked to 4.7Ghz, each subdivided into DDR 4000/3200/2133

4000 = 117.92
3200 = 117.63
2133 = 117.33

That all eqates to less than 1%

Now lets go to the last column.

Cpu is OC'd to 4Ghz

4000 = 105.38
3200 = 105.53
2133 = 100

The difference is negligable again.

The difference is down to clockspeed not the Memory. Skylake is not bottlenecked at all. The real issue is that Intel are struggling with IPC limitations already. Moores law dictates the IPC brickwall is 7nm and Intel are looking like hitting the limiter at 10nm. We are officially at the point of deminishing returns. CPU design is not leading to significant performance gains going from Haswell to Kaby, the only way to see any gains is in synthetics and that is as usefull as a cock flavoured lollipop to the lay person.

 

[razor1]

The graph is easy to understand, you seem to fail understanding it. so lets work though this like its the first day of prep school.

Column 1 is CPU overclocked to 4.7Ghz, each subdivided into DDR 4000/3200/2133

4000 = 117.92
3200 = 117.63
2133 = 117.33

That all eqates to less than 1%

Now lets go to the last column.

Cpu is OC'd to 4Ghz

4000 = 105.38
3200 = 105.53
2133 = 100

The difference is negligable again.

The difference is down to clockspeed not the Memory. Skylake is not bottlenecked at all. The real issue is that Intel are struggling with IPC limitations already. Moores law dictates the IPC brickwall is 7nm and Intel are looking like hitting the limiter at 10nm. We are officially at the point of deminishing returns. CPU design is not leading to significant performance gains going from Haswell to Kaby, the only way to see any gains is in synthetics and that is as usefull as a cock flavoured lollipop to the lay person.

You mean rows? not columns.... hmm understanding a graph really? come on.

And please look at the last two sets which the CPU isn't over clocked at all.

Yeah 5% change is negligible, first it was 2 % change, then now 5%, then when we have other reviews depending on programs, its huge differences those are negligible too lol. I don't care what you are saying because there are programs that we see 50% games that are 20% memory bottlenecked at 2200mhz let alone 1866 mhz is what you wanted to start off at with Haswell.

Nice to see you aren't reading everything.

 

[OrangeKhrush]

You mean rows? not columns.... hmm understanding a graph really? come on.

And please look at the last two sets which the CPU isn't over clocked at all.

Nice to see you aren't reading everything.

Whether at 4Ghz or 4.7Ghz there is just over or less than 1% difference between DDR 2133 and 4000, great argument you make. It is just about proven everywhere that in Gaming RAM makes almost no difference.

 

[razor1]

Whether at 4Ghz or 4.7Ghz there is just over or less than 1% difference between DDR 2133 and 4000, great argument you make. It is just about proven everywhere that in Gaming RAM makes almost no difference.

1% do you know math?

100 to 105 is 5% difference!

 

[razor1]

wtf is with these noobies lately! got one in the other Zen thread just god damn can't read charts, can't do math?..................................

 

[OrangeKhrush]

1% do you know math?

100 to 105 is 5% difference!

And you think that makes a world of difference?, that is like a frame if lucky.

I mean at 4.7Ghz the difference is 0.60 of a percentage point difference

 

[razor1]

wtf are you taking the overclocked results? Is that the only results you ever look at? I'm not talking with you, brick walls are not worth talking to anymore!


First you try to show Haswell which in the H review has no similar memory, then you start switching your shit around to what ever, I needed to clarify with you what hell you were talking about, then you dismiss 4 different websites that show memory bottlenecks, all using different programs, different clocks etc. so results are different, no one is arguing that, but its definitely there in one way or form.

So no, this is done go suck a lemon or what ever.

 

[OrangeKhrush]

wtf are you taking the overclocked results? Is that the only results you ever look at? I'm not talking with you, brick walls are not worth talking to anymore!


First you try to show Haswell which in the H review has no similar memory, then your start switching you shit around to what ever, I needed to clarify with you what hell you were talking about, then you dismiss 4 different websites that show memory bottlenecks, all using different programs, different clocks etc. so results are different, no one is arguing that, but its definitely there in one way or form.

So no, this is done go suck a lemon or what ever.

How can haswell have similar memory setup, Haswell runs on DDR 3 while Skylake on DDR4, by that token Skylake should have an unprecedented advantage of almost 50% greater bandwidth advantage, instead the 4770 and 4790 easily hold up to a 6700 in just about every gaming benchmark.

It is not a bottleneck, it is the fact that RAM has very little performance advantage, and Techreport, H and Tweaktown show that. At similar clocks RAM plays almost no role in performance.

 

[razor1]

How can haswell have similar memory setup, Haswell runs on DDR 3 while Skylake on DDR4, by that token Skylake should have an unprecedented advantage of almost 50% greater bandwidth advantage, instead the 4770 and 4790 easily hold up to a 6700 in just about every gaming benchmark.

Exactly so wtf did you try to compare them in the first place, do you want me to quote your statements from last page or two? I asked you at that point and you didn't respond to it!

Skylake can use more bandwidth and is bandwidth limited, but its application dependent. Newer applications will use more bandwidth.

 

[razor1]

so 54% increase in app performance when doing something isn't big for you?

a 19% increase in frame rates in a game isn't big for you,

Going to link the graphs

Sorry but its not .3 seconds I'm looking at here

How many FPS difference is there?

That's a pretty big increase in FPS its nowhere near the 5%, its double that.

Well this ones is 19%

Highly App dependent but they all show the same thing, Skylake is Bandwidth bound

Now this is the question I have for you, and how anyone can extrapolate data when the CPU is boosting different under different loads, and what the memory frequency and how it will affect the performance?

Orange + GRAPEFRUIT = UGLI FRUIT

That is what ya get when someone tries to extrapolate data like this.

I'm quoting my self here, see the difference.

This is why I asked you why are you all over the map? With a relevant starting point it can make a huge difference!

PS this is with a 980ti so there is no GPU bottleneck.

 

[OrangeKhrush]

Exactly so wtf did you try to compare them in the first place, do you want me to quote your statements from last page or two? I asked you at that point and you didn't respond to it!

Skylake can use more bandwidth and is bandwidth limited, but its application dependent. Newer applications will use more bandwidth.

The arguments started when Shintai claimed that a Haswell needed to be clocked to 5Ghz to beat a Skylake, he then claimed sarcasm which to me and going on his other assesments of facts hardly seemed like sarcasm but genuine belief.

So I started posting benches to show just how inconsequential Skylake is in gaming over Devils Canyon and an old 4770 in most reviews these CPU's are inside 1-5% of each other.

And lastly how this relates to Zen, is that Zen at least on clock for clock performance has equal or faster than Haswell performance so, for that, it has been about 12 years in the waiting but maybe now Intel will have incentive to give me an upgrade to my 4790 and 4960

 

[OrangeKhrush]

I'm quoting my self here, see the difference.

This is why I asked you why are you all over the map? With a relevant starting point it can make a huge difference!

PS this is with a 980ti so there is no GPU bottleneck.

That is great and all but we were talking about gaming, the only platform where RAM matters and I said that much was in professional or synthetic applications.

 

[razor1]

The arguments started when Shintai claimed that a Haswell needed to be clocked to 5Ghz to beat a Skylake, he then claimed sarcasm which to me and going on his other assesments of facts hardly seemed like sarcasm but genuine belief.

So I started posting benches to show just how inconsequential Skylake is in gaming over Devils Canyon and an old 4770 in most reviews these CPU's are inside 1-5% of each other.

And lastly how this relates to Zen, is that Zen at least on clock for clock performance has equal or faster than Haswell performance so, for that, it has been about 12 years in the waiting but maybe now Intel will have incentive to give me an upgrade to my 4790 and 4960

he was being sarcastic, and yet you kept on going with that! he didn't claim it, I stated it, that is the way I read it, I laughed when I read his statement. Because anyone knows that getting haswell to 5 ghz is pretty much impossible in regular situation.

 

[razor1]

That is great and all but we were talking about gaming, the only platform where RAM matters and I said that much was in professional or synthetic applications.

There are two games there too, I guess you like to read half stuff.........

I can give you two more games on top of that at 9% and 5% too.

 

[OrangeKhrush]

There are two games there too, I guess you like to read half stuff.........

I can give you two more games on top of that at 9% and 5% too.

No there were more, mostly Crysis, Farcry, Tombraider, Witcher 3, Civilisations, Project cars (lol)

 

[CCLMD]

Is this really going on multiple pages with people arguing over a 0-3% increase with RAM speed as significant?

What was the overall argument again? That RAM isn't a bottleneck in what context? That the Ryzen CPU will have better performance than indicated once adding super RAM? I don't even remember.

This has become a bit esoteric at this point... RAM is not going to make a massive difference. The fact that you can spend 2 pages on a thread linking 50 different graphs to support a 3-5% difference is like, a non-issue.

Things that are actually things on the internet end up with actual pages dedicated to them. Notice that you don't see this idea of RAM bottlenecking Haswell+ all over the internet. This seems one of the only threads talking about it in the last two years.

 

[Brackle]

I mean its an ES sample clocked at 3.15. We know stock clocks will start at 3.4ghz and boost to 3.7-3.8ghz. I would seriously wait for a good reviews.

I mean who knows if they havent had a new bios microcode. Or even a new stepping for ZEN. Just my 0.02c.

 

[razor1]

Is this really going on multiple pages with people arguing over a 0-3% increase with RAM speed as significant?

What was the overall argument again? That RAM isn't a bottleneck in what context? That the Ryzen CPU will have better performance than indicated once adding super RAM? I don't even remember.

This has become a bit esoteric at this point... RAM is not going to make a massive difference. The fact that you can spend 2 pages on a thread linking 50 different graphs to support a 3-5% difference is like, a non-issue.

Things that are actually things on the internet end up with actual pages dedicated to them. Notice that you don't see this idea of RAM bottlenecking Haswell+ all over the internet. This seems one of the only threads talking about it in the last two years.

Some graphs show much more than that, that was the point Skylake is sensitive to bandwidth (app dependent), Haswell we have no clue, haven't seen any overclocked ram results from it..... and too hard to overclock those ram units that fit Haswell too.

I mean its an ES sample clocked at 3.15. We know stock clocks will start at 3.4ghz and boost to 3.7-3.8ghz. I would seriously wait for a good reviews.

I mean who knows if they haven't had a new bios microcode. Or even a new stepping for ZEN. Just my 0.02c.

new stepping highly unlikely at this point, it will take 6 months to validate minimum so if anything the last stepping they would have done would be early June time right around the time they showed the first Zen presentation.

But yeah clock speeds should be at 3.4 with boosts going higher.

PS microcode and bios would have been done a long time ago, at this point it should just be tweaks. There was an Intel presentation on this when they finish up microcode and bios prior to actual silicon being manufactured, and all they do after that is tweak em. (this is the same with GPU drivers and bios for them too)

 

[Gideon]

Some graphs show much more than that, that was the point Skylake is sensitive to bandwidth (app dependent), Haswell we have no clue, haven't seen any overclocked ram results from it..... and too hard to overclock those ram units that fit Haswell too.




new stepping highly unlikely at this point, it will take 6 months to validate minimum so if anything the last stepping they would have done would be early June time right around the time they showed the first Zen presentation.

But yeah clock speeds should be at 3.4 with boosts going higher.

PS microcode and bios would have been done a long time ago, at this point it should just be tweaks. There was an Intel presentation on this when they finish up microcode and bios prior to actual silicon being manufactured, and all they do after that is tweak em. (this is the same with GPU drivers and bios for them too)

No offense Razor but you have no clue on what level of refinement that Zen chip was at. Your applying what you know about previous launches and applying that logic to Zen to justify what your trying to portray as facts. For all we know it was one of the first Zen es chips and has had no microcode updates done to it or is up to date.. we have no clue there. No proof is presented in the article just what they wrote and we have no clue on the status of the board used, could be a very early board as well or the latest one with new bios refinement, once again we have no clue. You are assuming too many things and treating them as facts, we know too little to call it a fact. Hell were not even sure when the chip will launch, just first quarter. That article is fun to speculate but no one can take much fact from it, since all we had was their word. hell the current CEO of AMD has said they are not done tweaking the chip yet, so I doubt the microcode is finalized yet.

 

[razor1]

No offense Razor but you have no clue on what level of refinement that Zen chip was at. Your applying what you know about previous launches and applying that logic to Zen to justify what your trying to portray as facts. For all we know it was one of the first Zen es chips and has had no microcode updates done to it or is up to date.. we have no clue there. No proof is presented in the article just what they wrote and we have no clue on the status of the board used, could be a very early board as well or the latest one with new bios refinement, once again we have no clue. You are assuming too many things and treating them as facts, we know too little to call it a fact. Hell were not even sure when the chip will launch, just first quarter. That article is fun to speculate but no one can take much fact from it, since all we had was their word. hell the current CEO of AMD has said they are not done tweaking the chip yet, so I doubt the microcode is finalized yet.

I'm applying my knowledge of how EE and Chip design has been done for years now. They aren't going to be changing things just because you or someone else thinks they can. There is a reason why they can't do it, Validation for CPU's is a much longer process and then you have motherboard manufactures that need final samples months prior to release of the chip. And they have their things to do too.

I didn't say finalized, I stated they are tweaking that is all.

They don't have the time to tweak the chip now, if they are launching in two weeks. Even in August when they stated they were still tweaking things, I highly doubt they were talking about the silicon, they were talking about microcode and bios, that is all they can do from that point onward.

So lets just say they do another respin of Zen, right in August? Hmm that gives them how many months to get back from the fab? usually it takes 1/2 a Q for them to get it back, then lets say they expedite validation.... 3 months instead of the normal 6 months to a year? You are already way past the point of launch in early Q1 of 2017, its now mid 2017 launch all this without the motherboard manufacturing getting the final product, so motherboards are going to be coming out when?

How can this time line fit with what you guys are stating? It is impossible to fit, so no respins after June OK?

 

[noko]

I'm applying my knowledge of how EE and Chip design has been done for years now. They aren't going to be changing things just because you or someone else thinks they can. There is a reason why they can't do it, Validation for CPU's is a much longer process and then you have motherboard manufactures that need final samples months prior to release of the chip. And they have their things to do too.

I didn't say finalized, I stated they are tweaking that is all.

They don't have the time to tweak the chip now, if they are launching in two weeks. Even in August when they stated they were still tweaking things, I highly doubt they were talking about the silicon, they were talking about microcode and bios, that is all they can do from that point onward.

So lets just say they do another respin of Zen, right in August? Hmm that gives them how many months to get back from the fab? usually it takes 1/2 a Q for them to get it back, then lets say they expedite validation.... 3 months instead of the normal 6 months to a year? You are already way past the point of launch in early Q1 of 2017, its now mid 2017 launch all this without the motherboard manufacturing getting the final product, so motherboards are going to be coming out when?

How can this time line fit with what you guys are stating? It is impossible to fit, so no respins after June OK?

What I got from it is final clock speed, turbo speeds and probably the built in OCing or boost speeds and overall balance as in voltage control of the chip. How much can be earned from that has to be seen. Isn't it already been put out 3.4ghz+? Plus if binned maybe a faster version as well? As for Respin, why would you respin a successful chip? Seems like it is hitting clock speeds and power envelope just fine? Now what was demonstrated was that not really the commercial chip even though still considered an engineering sample? I think so. Zen chip is ready (except maybe some fine tuning for memory, speeds etc), now everything else may need work.

 

[Gideon]

I'm applying my knowledge of how EE and Chip design has been done for years now. They aren't going to be changing things just because you or someone else thinks they can. There is a reason why they can't do it, Validation for CPU's is a much longer process and then you have motherboard manufactures that need final samples months prior to release of the chip. And they have their things to do too.

I didn't say finalized, I stated they are tweaking that is all.

They don't have the time to tweak the chip now, if they are launching in two weeks. Even in August when they stated they were still tweaking things, I highly doubt they were talking about the silicon, they were talking about microcode and bios, that is all they can do from that point onward.

So lets just say they do another respin of Zen, right in August? Hmm that gives them how many months to get back from the fab? usually it takes 1/2 a Q for them to get it back, then lets say they expedite validation.... 3 months instead of the normal 6 months to a year? You are already way past the point of launch in early Q1 of 2017, its now mid 2017 launch all this without the motherboard manufacturing getting the final product, so motherboards are going to be coming out when?

How can this time line fit with what you guys are stating? It is impossible to fit, so no respins after June OK?

I never said they need to respin it, but what I am saying is you have no idea if that was the first ES Zen made or a respin and a newer ES. What I am telling you is you have no idea what generation ES chip that was, you are going with the theory it was the last ES sample they made and thus what they approved for mass production. It's a great piece to speculate on but thats about all it's worth, heck the reviewer even said he didnt think it was ready for prime time, which could mean its a very early ES sample he had. Just don't see how it's giving us many facts to go with.

 

[NamelessPFG]

Assuming these benches are accurate, what I'm seeing here is that Ryzen's going to need some serious overclocking headroom to hang with the Haswell and Skylake quads we're used to, and that won't be easy with twice the cores to draw power and dump heat.

It is, however, a far more compelling option for workloads that would normally warrant 6/8-core Intel HEDT chips with their huge price premiums. AMD might have a chance.

As for gaming, though? Call me when someone's got some kinda DCS World/PlanetSide 2/ArmA III/Squad/insert heavily CPU-limited game here benchmarks, the kind that really highlight how weak Sandy Bridge is IPC-wise compared to Haswell and Skylake right now, even when overclocked. Those will really make or break Ryzen as an option for a gaming build.

 

[razor1]

I never said they need to respin it, but what I am saying is you have no idea if that was the first ES Zen made or a respin and a newer ES. What I am telling you is you have no idea what generation ES chip that was, you are going with the theory it was the last ES sample they made and thus what they approved for mass production. It's a great piece to speculate on but thats about all it's worth, heck the reviewer even said he didnt think it was ready for prime time, which could mean its a very early ES sample he had. Just don't see how it's giving us many facts to go with.

Ah yes the ES sample they used to show off, there could be another respin if after they got that one back they immediately sent it it in for another respin in June.

ES chips are fully functional, and pretty much very close to retail chips. Now for a reviewer to get their hands on a ES it would have to come from the motherboard manufacturer, cause it didn't come form AMD. ES samples are sent to motherboard manufactures for them to get working on their boards.

What I got from it is final clock speed, turbo speeds and probably the built in OCing or boost speeds and overall balance as in voltage control of the chip. How much can be earned from that has to be seen. Isn't it already been put out 3.4ghz+? Plus if binned maybe a faster version as well? As for Respin, why would you respin a successful chip? Seems like it is hitting clock speeds and power envelope just fine? Now what was demonstrated was that not really the commercial chip even though still considered an engineering sample? I think so. Zen chip is ready (except maybe some fine tuning for memory, speeds etc), now everything else may need work.

Now the rest of the stuff low clocks, stability, most likely is a chip set and motherboard driver issues, not really the chip itself. At least not this late in the game. These take longer to get ready than the microcode and bios of the CPU

 

[OrangeKhrush]

Assuming these benches are accurate, what I'm seeing here is that Ryzen's going to need some serious overclocking headroom to hang with the Haswell and Skylake quads we're used to, and that won't be easy with twice the cores to draw power and dump heat.

It is, however, a far more compelling option for workloads that would normally warrant 6/8-core Intel HEDT chips with their huge price premiums. AMD might have a chance.

As for gaming, though? Call me when someone's got some kinda DCS World/PlanetSide 2/ArmA III/Squad/insert heavily CPU-limited game here benchmarks, the kind that really highlight how weak Sandy Bridge is IPC-wise compared to Haswell and Skylake right now, even when overclocked. Those will really make or break Ryzen as an option for a gaming build.

I think you are underestimating the IPC, I think the 1180+ score posted by the Japanese guy was base clock IPC, making it possibly faster than Haswell at very least equal.

Low clocks and 200mhz turbo headroom is not great for gaming, it is the reason why my 4460 is not used for gaming, why I use a 4790 as a gaming system. The higher clock spead and wider turbo are great. If it can game with Skylake i5's already and be unoptimised (ie: low native IMC on ES is 2133, beta motherboards, flimsy turbo that sometimes doesn't even boost) I think it is far more promising than most want to believe. Caution is fine but out right negitivity isn't. Why would AMD waste 5 years of development to develop a Sandy bridge processor when the stake of the company is on it? AMD did what intel had to do with Pentium 4 and D, roll out your rubbish architecture but on the side work on the new one and fix the mistakes made. Intel did that with Conroe, AMD has done that before with Hammer, they will do it again with Ryzen.

Another indicator for me is AMD are not stalling this CPU, they seem very confident in its performance but they are not excessively releasing market fud like the BD era marketing. They are very secretive on this, but when need to they are quite confident.

 

[Shintai]

I think you are underestimating the IPC, I think the 1180+ score posted by the Japanese guy was base clock IPC, making it possibly faster than Haswell at very least equal.

And I think you are overestimating the IPC. Before launch the hype train will have put it way ahead of Skylake too!

 

[Rvenger]

Once this CPU releases and it DOES have somewhere between Haswell and Skylake IPC then we will be seeing a bunch of cherry picked benchmarks of what Ryzen doesn't do well. Lets make those goal posts float.

 

[Shintai]

Once this CPU releases and it DOES have somewhere between Haswell and Skylake IPC then we will be seeing a bunch of cherry picked benchmarks of what Ryzen doesn't do well. Lets make those goal posts float.

That would be a welcome change from the cheery picks we usually see the other way around. Then we dont have to see 7zip and WinRAR+gaming and what else have been presented over time as relevant benchmarks.

 

[OrangeKhrush]

A highly unoptimised Zen on a motherboard that is known to not boost and doesn't work off the latest ES chips yet performs inside 10% of a 6900K with optimal boost clocks around 3.5Ghz and as high as 3.7Ghz. There is still lots to come out of this. I think I am more optimistic than a couple months ago.

 

[SighTurtle]

A highly unoptimised Zen on a motherboard that is known to not boost and doesn't work off the latest ES chips yet performs inside 10% of a 6900K with optimal boost clocks around 3.5Ghz and as high as 3.7Ghz. There is still lots to come out of this. I think I am more optimistic than a couple months ago.

Who said it would boost that high? All I've seen from AMD is 3.4+.

 

[Rvenger]

That would be a welcome change from the cheery picks we usually see the other way around. Then we dont have to see 7zip and WinRAR+gaming and what else have been presented over time as relevant benchmarks.

I was going to say 7 Zip. LOL

A highly unoptimised Zen on a motherboard that is known to not boost and doesn't work off the latest ES chips yet performs inside 10% of a 6900K with optimal boost clocks around 3.5Ghz and as high as 3.7Ghz. There is still lots to come out of this. I think I am more optimistic than a couple months ago.

I wouldn't be coming to conclusions just yet. We don't need overhype to derail the hype train.. Just wait for the final specs.

 

[noko]

The best scenario is AMD is a little bit smarter and showed even benchmark scores with a held back chip to foster excitement - discussion - arguments that keeps AMD more in the headlines.

• Followed by a launch that blows away previous benches with a more than expected clock speed and a good performance in more benches then not.
• AMD does not need to cherry pick benchmarks yet again - that would be a losing strategy. Common to AMD in the past.
• What that really means is I hope HardOCP procures a sample and does an accurate review (the norm) followed by discussions of RyZen.
• I hope AMD is smarter and talk through the reviewers as in like Kyle gets answers direct from AMD and of course AIBs (no bypassing from AMD) to foster the discussions online and build those discussion groups (AnAndTech, HardOCP, TechReport, YouTube etc).
• This way Ryzen through numerous sources will be relayed with tens of thousands of participants joining in many different formatted discussions and video's.
• This would also shield AMD as well since mistakes (they will happen) can be corrected without the ever AMD deceived us, lied to us etc. accusations.
• Once the dust settles, then AMD shows up on the forums, Newegg, BestBuy, YouTube etc. with an already established educated group to answer questions directly (not before) to show utter support of their products hopefully bringing along smiling very happy customers showing their enthusiasm especially on the YouTube type format.
  • Nothing sells better then smiling enthusiastic owners broadcasting their upmost fantastic experience with a product (not smiling AMD managers, except maybe in the background showing support and gratitude for their customers)

We just are waiting on the real McCoy now, the whole package. I just hope what ever AMD does they smartly delivers that package using resources way beyond what they can do by themselves.

 

[OrangeKhrush]

Who said it would boost that high? All I've seen from AMD is 3.4+.

I didn't say Ryzen, the i7 6900K boosts standard between 3.5 and 3.7Ghz, a clock bump of Ryzen to a base of at least 3.4Ghz should bring down the 10-12% difference by a decent margin. The simple fact is, at 3.1 and with the Ageas firmware A0 boards not allowing proper Turbo frequencies, a 200mhz boost is pointless. The fact is, this chip is competeting rather well with even skylake and it is still not even retail standard yet.

I was going to say 7 Zip. LOL



I wouldn't be coming to conclusions just yet. We don't need overhype to derail the hype train.. Just wait for the final specs.

It is not our hype but CanardPC are very trustworthy, and they are the people that do the validations for CPUID/CPU-Z so I think I trust them more than most.

AMD have done next to no hype on this CPU at all, in fact it almost secretive.

The best scenario is AMD is a little bit smarter and showed even benchmark scores with a held back chip to foster excitement - discussion - arguments that keeps AMD more in the headlines.

• Followed by a launch that blows away previous benches with a more than expected clock speed and a good performance in more benches then not.
• AMD does not need to cherry pick benchmarks yet again - that would be a losing strategy. Common to AMD in the past.
• What that really means is I hope HardOCP needs to procure a sample and does an accurate review (the norm) followed by discussions of RyZen.
• I hope AMD is smarter and talk through the reviewers as in like Kyle gets answers direct from AMD and of course AIBs (no bypassing from AMD) to foster the discussions online and build those discussion groups (AnAndTech, HardOCP, TechReport, YouTube etc).
• This way Ryzen through numerous sources will be relayed with tens of thousands of participants joining in many different formatted discussions and video's.
• This would also shield AMD as well since mistakes (they will happen) can be corrected without the ever AMD deceived us, lied to us etc. accusations.
• Once the dust settles, then AMD themselves really show up on the forums, Newegg, BestBuy, YouTube etc. with an already established educated group to answer questions directly (not before) to show utter support of their products hopefully bringing along smiling very happy customers showing upmost enthusiasm especially on the YouTube type format.
  • Nothing sells better then smiling enthusiastic owners broadcasting their upmost fantastic experience with a product (not smiling AMD managers, except maybe in the background showing support and gratitude for their customers)

We just are waiting on the real McCoy now, the whole package. I just hope what ever AMD does they smartly delivers that package using resources way beyond what they can do by themselves.

My feeling two months ago was that best case or at least the target was Haswell, talking to people with privy to ES samples adn going on the stategic leaks, there is genuine sentiment that Zen is a very strong architecture. I get that hype can be a set up, but unlike bulldozer and the 500 delays, marketing rubbish and what not that went on for a year, Zen is anything but that. AMD have ensured that even the ES chips are sandbagged. Not a single ES chip feature the XFR, SenseMI technology, no finished retail boards yet with teh full support, no clocks yet and it still managed to be inside 10% to a 6900K. It has been forever since I had an AMD CPU, and given how "boring" intel has become with useless peripherals to mask less than optimal improvements, I want something new and a little gung ho.

I am one of those people that wants a strong AMD, the market is better for it, that is great for everyone, to have choices.