AMD CPUs, SoC Rumors and Speculations Temp. thread 2

[cemerian]

http://www.bitsandchips.it/52-english-news/6815-speculations-about-zen-after-our-april-s-fool
here you go

 

[8350rocks]

It was well done, and just a reasonable enough better than expectations projection, with accurate A0 sample info and base clocks.

There was not much BS in there to be honest. Intel is bringing a 10 core Broadwell-E to market still...

Per core performance is projected to be strong enough that they do not want to compete with Zen on a core per core basis. That alone should tell people something about what to expect.

 

[juanrga]

http://www.bitsandchips.it/english/52-english-news/6811-first-zen-benchmarks-the-next-amd-uarch-is-a-major-step-forward

Wow...it appears that Zen performance is going to be so strong, Intel will actually go ahead and launch a 10 core Broadwell-E.

The CPU Hash (mostly FPU performance), for the 8 core zen shows to be stronger than a 20 core broadwell Xeon.

Meanwhile, the FMA implementation looks a tad bit suspect with the low ray trace scores...but I am not really concerned.

The ES shows base clock 3.0, which means we will likely see something around 3.5-3.7 all said and done on the top SKU before turbo.

If bits and chips is to be believed...this will be on par with broadwell-E in terms of per core performance...they even outright state so.

As a side note...that...is....actually, better than I had heard. I knew there was a lot of optimism about it. However...those numbers are stronger than I had even heard tossed around. Which is reassuring. Hopefully the other benchmarks that come out will reveal how much of the performance improvement this is across the board. If it is a legitimate 40%...it will likely be broadwell/skylake-ish performance...which would make it quite competitive. Especially if the cost is inline.

It was evidently an April joke from my friend Fottemberg... although some data from the report is real.

Zen microarchitecture is inferior to Intel and IBM designs; therefore, performance will be inferior. Samsung 14LPP node is also inferior to Intel/IBM nodes.

Intel is not releasing a 10-core E because Zen is strong. Intel has been systematically increasing the core count since before Sandy: 4 --> 6 --> 8 and now 10. Intel increases the core count with new nodes, because there is more space in the die.

The Zen FPU unit is clearly inferior. About one half of Haswell/Broadwell and about a quarter of Skylake-E/KNL.

The base clocks will be low because: (i) 14LPP is optimized for sub 3GHz frequencies and (ii) the 95W limit. No way octo-core Zen will hit 3.7/4.1GHz.

 

[8350rocks]

http://www.bitsandchips.it/english/52-english-news/6811-first-zen-benchmarks-the-next-amd-uarch-is-a-major-step-forward

Wow...it appears that Zen performance is going to be so strong, Intel will actually go ahead and launch a 10 core Broadwell-E.

The CPU Hash (mostly FPU performance), for the 8 core zen shows to be stronger than a 20 core broadwell Xeon.

Meanwhile, the FMA implementation looks a tad bit suspect with the low ray trace scores...but I am not really concerned.

The ES shows base clock 3.0, which means we will likely see something around 3.5-3.7 all said and done on the top SKU before turbo.

If bits and chips is to be believed...this will be on par with broadwell-E in terms of per core performance...they even outright state so.

As a side note...that...is....actually, better than I had heard. I knew there was a lot of optimism about it. However...those numbers are stronger than I had even heard tossed around. Which is reassuring. Hopefully the other benchmarks that come out will reveal how much of the performance improvement this is across the board. If it is a legitimate 40%...it will likely be broadwell/skylake-ish performance...which would make it quite competitive. Especially if the cost is inline.

It was evidently an April joke from my friend Fottemberg... although some data from the report is real.

Zen microarchitecture is inferior to Intel and IBM designs; therefore, performance will be inferior. Samsung 14LPP node is also inferior to Intel/IBM nodes.

Intel is not releasing a 10-core E because Zen is strong. Intel has been systematically increasing the core count since before Sandy: 4 --> 6 --> 8 and now 10. Intel increases the core count with new nodes, because there is more space in the die.

The Zen FPU unit is clearly inferior. About one half of Haswell/Broadwell and about a quarter of Skylake-E/KNL.

The base clocks will be low because: (i) 14LPP is optimized for sub 3GHz frequencies and (ii) the 95W limit. No way octo-core Zen will hit 3.7/4.1GHz.

For reference, A0 for Piledriver was 2.8 GHz.

A0 for Zen is 3.0 GHz.

Engineering samples are usually indicative of process capability to a degree. How quantifiable that is varies by architecture, etc. Having said that...I do not expect the clock speed to hit 4.0 like PD did in lots of chips. However, a 3.7 base clock with 4.0-4.2 turbo could be quite feasible given the fact that it would only be a 700 MHz bump over the A0 engineering sample.

 

[jimmysmitty]

http://www.bitsandchips.it/english/52-english-news/6811-first-zen-benchmarks-the-next-amd-uarch-is-a-major-step-forward

Wow...it appears that Zen performance is going to be so strong, Intel will actually go ahead and launch a 10 core Broadwell-E.

The CPU Hash (mostly FPU performance), for the 8 core zen shows to be stronger than a 20 core broadwell Xeon.

Meanwhile, the FMA implementation looks a tad bit suspect with the low ray trace scores...but I am not really concerned.

The ES shows base clock 3.0, which means we will likely see something around 3.5-3.7 all said and done on the top SKU before turbo.

If bits and chips is to be believed...this will be on par with broadwell-E in terms of per core performance...they even outright state so.

As a side note...that...is....actually, better than I had heard. I knew there was a lot of optimism about it. However...those numbers are stronger than I had even heard tossed around. Which is reassuring. Hopefully the other benchmarks that come out will reveal how much of the performance improvement this is across the board. If it is a legitimate 40%...it will likely be broadwell/skylake-ish performance...which would make it quite competitive. Especially if the cost is inline.

It was evidently an April joke from my friend Fottemberg... although some data from the report is real.

Zen microarchitecture is inferior to Intel and IBM designs; therefore, performance will be inferior. Samsung 14LPP node is also inferior to Intel/IBM nodes.

Intel is not releasing a 10-core E because Zen is strong. Intel has been systematically increasing the core count since before Sandy: 4 --> 6 --> 8 and now 10. Intel increases the core count with new nodes, because there is more space in the die.

The Zen FPU unit is clearly inferior. About one half of Haswell/Broadwell and about a quarter of Skylake-E/KNL.

The base clocks will be low because: (i) 14LPP is optimized for sub 3GHz frequencies and (ii) the 95W limit. No way octo-core Zen will hit 3.7/4.1GHz.

Correct. In fact we knew about the 10 core Broadwell-E part quite a while ago. I am sure Intel will put 12 cores in Skylake-E if they can since that platform is about more than just gamers, it is also their workstation platform.

http://www.bitsandchips.it/english/52-english-news/6811-first-zen-benchmarks-the-next-amd-uarch-is-a-major-step-forward

Wow...it appears that Zen performance is going to be so strong, Intel will actually go ahead and launch a 10 core Broadwell-E.

The CPU Hash (mostly FPU performance), for the 8 core zen shows to be stronger than a 20 core broadwell Xeon.

Meanwhile, the FMA implementation looks a tad bit suspect with the low ray trace scores...but I am not really concerned.

The ES shows base clock 3.0, which means we will likely see something around 3.5-3.7 all said and done on the top SKU before turbo.

If bits and chips is to be believed...this will be on par with broadwell-E in terms of per core performance...they even outright state so.

As a side note...that...is....actually, better than I had heard. I knew there was a lot of optimism about it. However...those numbers are stronger than I had even heard tossed around. Which is reassuring. Hopefully the other benchmarks that come out will reveal how much of the performance improvement this is across the board. If it is a legitimate 40%...it will likely be broadwell/skylake-ish performance...which would make it quite competitive. Especially if the cost is inline.

It was evidently an April joke from my friend Fottemberg... although some data from the report is real.

Zen microarchitecture is inferior to Intel and IBM designs; therefore, performance will be inferior. Samsung 14LPP node is also inferior to Intel/IBM nodes.

Intel is not releasing a 10-core E because Zen is strong. Intel has been systematically increasing the core count since before Sandy: 4 --> 6 --> 8 and now 10. Intel increases the core count with new nodes, because there is more space in the die.

The Zen FPU unit is clearly inferior. About one half of Haswell/Broadwell and about a quarter of Skylake-E/KNL.

The base clocks will be low because: (i) 14LPP is optimized for sub 3GHz frequencies and (ii) the 95W limit. No way octo-core Zen will hit 3.7/4.1GHz.

For reference, A0 for Piledriver was 2.8 GHz.

A0 for Zen is 3.0 GHz.

Engineering samples are usually indicative of process capability to a degree. How quantifiable that is varies by architecture, etc. Having said that...I do not expect the clock speed to hit 4.0 like PD did in lots of chips. However, a 3.7 base clock with 4.0-4.2 turbo could be quite feasible given the fact that it would only be a 700 MHz bump over the A0 engineering sample.

Possibly. It could also be that AMD got better results since Samsungs 14nm is quite mature, not as mature as Intels but it is mature enough that AMD might be hitting higher early stage clock speeds than BD that was the first CPU on the new 32nm process.

I guess we will see.

Even though those benchmarks were a joke I do expect plenty of people to think they are real. Will not be fun dealing with that.

 

[salgado18]

It was evidently an April joke from my friend Fottemberg... although some data from the report is real.

Zen microarchitecture is inferior to Intel and IBM designs; therefore, performance will be inferior. Samsung 14LPP node is also inferior to Intel/IBM nodes.

Intel is not releasing a 10-core E because Zen is strong. Intel has been systematically increasing the core count since before Sandy: 4 --> 6 --> 8 and now 10. Intel increases the core count with new nodes, because there is more space in the die.

The Zen FPU unit is clearly inferior. About one half of Haswell/Broadwell and about a quarter of Skylake-E/KNL.

The base clocks will be low because: (i) 14LPP is optimized for sub 3GHz frequencies and (ii) the 95W limit. No way octo-core Zen will hit 3.7/4.1GHz.

Correct. In fact we knew about the 10 core Broadwell-E part quite a while ago. I am sure Intel will put 12 cores in Skylake-E if they can since that platform is about more than just gamers, it is also their workstation platform.

http://www.bitsandchips.it/english/52-english-news/6811-first-zen-benchmarks-the-next-amd-uarch-is-a-major-step-forward

Wow...it appears that Zen performance is going to be so strong, Intel will actually go ahead and launch a 10 core Broadwell-E.

The CPU Hash (mostly FPU performance), for the 8 core zen shows to be stronger than a 20 core broadwell Xeon.

Meanwhile, the FMA implementation looks a tad bit suspect with the low ray trace scores...but I am not really concerned.

The ES shows base clock 3.0, which means we will likely see something around 3.5-3.7 all said and done on the top SKU before turbo.

If bits and chips is to be believed...this will be on par with broadwell-E in terms of per core performance...they even outright state so.

As a side note...that...is....actually, better than I had heard. I knew there was a lot of optimism about it. However...those numbers are stronger than I had even heard tossed around. Which is reassuring. Hopefully the other benchmarks that come out will reveal how much of the performance improvement this is across the board. If it is a legitimate 40%...it will likely be broadwell/skylake-ish performance...which would make it quite competitive. Especially if the cost is inline.

It was evidently an April joke from my friend Fottemberg... although some data from the report is real.

Zen microarchitecture is inferior to Intel and IBM designs; therefore, performance will be inferior. Samsung 14LPP node is also inferior to Intel/IBM nodes.

Intel is not releasing a 10-core E because Zen is strong. Intel has been systematically increasing the core count since before Sandy: 4 --> 6 --> 8 and now 10. Intel increases the core count with new nodes, because there is more space in the die.

The Zen FPU unit is clearly inferior. About one half of Haswell/Broadwell and about a quarter of Skylake-E/KNL.

The base clocks will be low because: (i) 14LPP is optimized for sub 3GHz frequencies and (ii) the 95W limit. No way octo-core Zen will hit 3.7/4.1GHz.

For reference, A0 for Piledriver was 2.8 GHz.

A0 for Zen is 3.0 GHz.

Engineering samples are usually indicative of process capability to a degree. How quantifiable that is varies by architecture, etc. Having said that...I do not expect the clock speed to hit 4.0 like PD did in lots of chips. However, a 3.7 base clock with 4.0-4.2 turbo could be quite feasible given the fact that it would only be a 700 MHz bump over the A0 engineering sample.

Possibly. It could also be that AMD got better results since Samsungs 14nm is quite mature, not as mature as Intels but it is mature enough that AMD might be hitting higher early stage clock speeds than BD that was the first CPU on the new 32nm process.

I guess we will see.

Even though those benchmarks were a joke I do expect plenty of people to think they are real. Will not be fun dealing with that.

I'm behind juanrga on this. Even Haswell-E, on a 22nm mature node, reaches just 3.0/3.5 GHz with 8 cores. No way AMD will get 3.5+ on a 14nm LPP with 8 newly-developed cores.

However, I do believe they will get there with 4 cores.

 

[Vogner16]

Looks like juans early estimate of 3.5 was pretty close

 

[gamerk316]

Wow, we got people falling for obvious fake benchmarks already? Come on guys, it's not even 3 months before release yet.

 

[jimmysmitty]

Wow, we got people falling for obvious fake benchmarks already? Come on guys, it's not even 3 months before release yet.

This is how it goes and will continue to go for the next couple of months. I expect some sites to start ramping up their news articles with multiple rumors per week until we have an official release of information through one of the major tech sites.

 

[jed]

Wow, we got people falling for obvious fake benchmarks already? Come on guys, it's not even 3 months before release yet.

You know this is how it start off, rumors become facts and then the product is released over-hyped
and short on perceived performance, with egg all over ones face. this goes on every time AMD release
a product, and it's getting very old.

This is where it will start this time.


"There was not much BS in there to be honest. Intel is bringing a 10 core Broadwell-E to market still...

Per core performance is projected to be strong enough that they do not want to compete with Zen on a core per core basis. That alone should tell people something about what to expect."

 

[juanrga]

http://www.bitsandchips.it/english/52-english-news/6811-first-zen-benchmarks-the-next-amd-uarch-is-a-major-step-forward

Wow...it appears that Zen performance is going to be so strong, Intel will actually go ahead and launch a 10 core Broadwell-E.

The CPU Hash (mostly FPU performance), for the 8 core zen shows to be stronger than a 20 core broadwell Xeon.

Meanwhile, the FMA implementation looks a tad bit suspect with the low ray trace scores...but I am not really concerned.

The ES shows base clock 3.0, which means we will likely see something around 3.5-3.7 all said and done on the top SKU before turbo.

If bits and chips is to be believed...this will be on par with broadwell-E in terms of per core performance...they even outright state so.

As a side note...that...is....actually, better than I had heard. I knew there was a lot of optimism about it. However...those numbers are stronger than I had even heard tossed around. Which is reassuring. Hopefully the other benchmarks that come out will reveal how much of the performance improvement this is across the board. If it is a legitimate 40%...it will likely be broadwell/skylake-ish performance...which would make it quite competitive. Especially if the cost is inline.

It was evidently an April joke from my friend Fottemberg... although some data from the report is real.

Zen microarchitecture is inferior to Intel and IBM designs; therefore, performance will be inferior. Samsung 14LPP node is also inferior to Intel/IBM nodes.

Intel is not releasing a 10-core E because Zen is strong. Intel has been systematically increasing the core count since before Sandy: 4 --> 6 --> 8 and now 10. Intel increases the core count with new nodes, because there is more space in the die.

The Zen FPU unit is clearly inferior. About one half of Haswell/Broadwell and about a quarter of Skylake-E/KNL.

The base clocks will be low because: (i) 14LPP is optimized for sub 3GHz frequencies and (ii) the 95W limit. No way octo-core Zen will hit 3.7/4.1GHz.

For reference, A0 for Piledriver was 2.8 GHz.

A0 for Zen is 3.0 GHz.

Engineering samples are usually indicative of process capability to a degree. How quantifiable that is varies by architecture, etc. Having said that...I do not expect the clock speed to hit 4.0 like PD did in lots of chips. However, a 3.7 base clock with 4.0-4.2 turbo could be quite feasible given the fact that it would only be a 700 MHz bump over the A0 engineering sample.

Engineering samples depend on a lot of factors, including how mature is the process node used or how accurate are the design/implementation tools used by the engineers. Quoting base frequencies of ES on different process nodes on so large time span means little.

What matters to know frequencies of final silicon is: (i) microarchitecture details, (ii) process node characteristics, (iii) platform TDP.

Points (i)+(ii)+(iii) together clearly suggest lower frequencies.

 

[juanrga]

Even though those benchmarks were a joke I do expect plenty of people to think they are real. Will not be fun dealing with that.

You can find 'professional' site news already reporting the fake Zen benchmarks like genuine, somewhat like they reported the fake Zen slides the past year.

 

[8350rocks]

Looks like juans early estimate of 3.5 was pretty close

Lots of people expected in that range...to be quite honest...

 

[8350rocks]

http://www.bitsandchips.it/english/52-english-news/6811-first-zen-benchmarks-the-next-amd-uarch-is-a-major-step-forward

Wow...it appears that Zen performance is going to be so strong, Intel will actually go ahead and launch a 10 core Broadwell-E.

The CPU Hash (mostly FPU performance), for the 8 core zen shows to be stronger than a 20 core broadwell Xeon.

Meanwhile, the FMA implementation looks a tad bit suspect with the low ray trace scores...but I am not really concerned.

The ES shows base clock 3.0, which means we will likely see something around 3.5-3.7 all said and done on the top SKU before turbo.

If bits and chips is to be believed...this will be on par with broadwell-E in terms of per core performance...they even outright state so.

As a side note...that...is....actually, better than I had heard. I knew there was a lot of optimism about it. However...those numbers are stronger than I had even heard tossed around. Which is reassuring. Hopefully the other benchmarks that come out will reveal how much of the performance improvement this is across the board. If it is a legitimate 40%...it will likely be broadwell/skylake-ish performance...which would make it quite competitive. Especially if the cost is inline.

It was evidently an April joke from my friend Fottemberg... although some data from the report is real.

Zen microarchitecture is inferior to Intel and IBM designs; therefore, performance will be inferior. Samsung 14LPP node is also inferior to Intel/IBM nodes.

Intel is not releasing a 10-core E because Zen is strong. Intel has been systematically increasing the core count since before Sandy: 4 --> 6 --> 8 and now 10. Intel increases the core count with new nodes, because there is more space in the die.

The Zen FPU unit is clearly inferior. About one half of Haswell/Broadwell and about a quarter of Skylake-E/KNL.

The base clocks will be low because: (i) 14LPP is optimized for sub 3GHz frequencies and (ii) the 95W limit. No way octo-core Zen will hit 3.7/4.1GHz.

For reference, A0 for Piledriver was 2.8 GHz.

A0 for Zen is 3.0 GHz.

Engineering samples are usually indicative of process capability to a degree. How quantifiable that is varies by architecture, etc. Having said that...I do not expect the clock speed to hit 4.0 like PD did in lots of chips. However, a 3.7 base clock with 4.0-4.2 turbo could be quite feasible given the fact that it would only be a 700 MHz bump over the A0 engineering sample.

Engineering samples depend on a lot of factors, including how mature is the process node used or how accurate are the design/implementation tools used by the engineers. Quoting base frequencies of ES on different process nodes on so large time span means little.

What matters to know frequencies of final silicon is: (i) microarchitecture details, (ii) process node characteristics, (iii) platform TDP.

Points (i)+(ii)+(iii) together clearly suggest lower frequencies.

~100W TDP suggests nothing in the way of low clocks juan...what was the TDP of Devil's Canyon? 88W?

 

[juanrga]

Looks like juans early estimate of 3.5 was pretty close

Lots of people expected in that range...to be quite honest...

Lately, I can see several people expecting ~3.5GHz, but two years ago, many of them rejected my claims and predicted "4.1GHz" and even higher base clocks. That was the same people that predicted "AVX512", "2x256 bit FMAC", "CMT", and other stuff for Zen and failed.

 

[juanrga]

http://www.bitsandchips.it/english/52-english-news/6811-first-zen-benchmarks-the-next-amd-uarch-is-a-major-step-forward

Wow...it appears that Zen performance is going to be so strong, Intel will actually go ahead and launch a 10 core Broadwell-E.

The CPU Hash (mostly FPU performance), for the 8 core zen shows to be stronger than a 20 core broadwell Xeon.

Meanwhile, the FMA implementation looks a tad bit suspect with the low ray trace scores...but I am not really concerned.

The ES shows base clock 3.0, which means we will likely see something around 3.5-3.7 all said and done on the top SKU before turbo.

If bits and chips is to be believed...this will be on par with broadwell-E in terms of per core performance...they even outright state so.

As a side note...that...is....actually, better than I had heard. I knew there was a lot of optimism about it. However...those numbers are stronger than I had even heard tossed around. Which is reassuring. Hopefully the other benchmarks that come out will reveal how much of the performance improvement this is across the board. If it is a legitimate 40%...it will likely be broadwell/skylake-ish performance...which would make it quite competitive. Especially if the cost is inline.

It was evidently an April joke from my friend Fottemberg... although some data from the report is real.

Zen microarchitecture is inferior to Intel and IBM designs; therefore, performance will be inferior. Samsung 14LPP node is also inferior to Intel/IBM nodes.

Intel is not releasing a 10-core E because Zen is strong. Intel has been systematically increasing the core count since before Sandy: 4 --> 6 --> 8 and now 10. Intel increases the core count with new nodes, because there is more space in the die.

The Zen FPU unit is clearly inferior. About one half of Haswell/Broadwell and about a quarter of Skylake-E/KNL.

The base clocks will be low because: (i) 14LPP is optimized for sub 3GHz frequencies and (ii) the 95W limit. No way octo-core Zen will hit 3.7/4.1GHz.

For reference, A0 for Piledriver was 2.8 GHz.

A0 for Zen is 3.0 GHz.

Engineering samples are usually indicative of process capability to a degree. How quantifiable that is varies by architecture, etc. Having said that...I do not expect the clock speed to hit 4.0 like PD did in lots of chips. However, a 3.7 base clock with 4.0-4.2 turbo could be quite feasible given the fact that it would only be a 700 MHz bump over the A0 engineering sample.

Engineering samples depend on a lot of factors, including how mature is the process node used or how accurate are the design/implementation tools used by the engineers. Quoting base frequencies of ES on different process nodes on so large time span means little.

What matters to know frequencies of final silicon is: (i) microarchitecture details, (ii) process node characteristics, (iii) platform TDP.

Points (i)+(ii)+(iii) together clearly suggest lower frequencies.

~100W TDP suggests nothing in the way of low clocks juan...what was the TDP of Devil's Canyon? 88W?

Apart from ignoring two of my three points, you also ignore that I was discussing the clocks for an "octo-core", and the 88W Devil's Canyon is a quad-core, not to mention it uses a different 22nm process than Zen...

 

[8350rocks]

http://www.bitsandchips.it/english/52-english-news/6811-first-zen-benchmarks-the-next-amd-uarch-is-a-major-step-forward

Wow...it appears that Zen performance is going to be so strong, Intel will actually go ahead and launch a 10 core Broadwell-E.

The CPU Hash (mostly FPU performance), for the 8 core zen shows to be stronger than a 20 core broadwell Xeon.

Meanwhile, the FMA implementation looks a tad bit suspect with the low ray trace scores...but I am not really concerned.

The ES shows base clock 3.0, which means we will likely see something around 3.5-3.7 all said and done on the top SKU before turbo.

If bits and chips is to be believed...this will be on par with broadwell-E in terms of per core performance...they even outright state so.

As a side note...that...is....actually, better than I had heard. I knew there was a lot of optimism about it. However...those numbers are stronger than I had even heard tossed around. Which is reassuring. Hopefully the other benchmarks that come out will reveal how much of the performance improvement this is across the board. If it is a legitimate 40%...it will likely be broadwell/skylake-ish performance...which would make it quite competitive. Especially if the cost is inline.

It was evidently an April joke from my friend Fottemberg... although some data from the report is real.

Zen microarchitecture is inferior to Intel and IBM designs; therefore, performance will be inferior. Samsung 14LPP node is also inferior to Intel/IBM nodes.

Intel is not releasing a 10-core E because Zen is strong. Intel has been systematically increasing the core count since before Sandy: 4 --> 6 --> 8 and now 10. Intel increases the core count with new nodes, because there is more space in the die.

The Zen FPU unit is clearly inferior. About one half of Haswell/Broadwell and about a quarter of Skylake-E/KNL.

The base clocks will be low because: (i) 14LPP is optimized for sub 3GHz frequencies and (ii) the 95W limit. No way octo-core Zen will hit 3.7/4.1GHz.

For reference, A0 for Piledriver was 2.8 GHz.

A0 for Zen is 3.0 GHz.

Engineering samples are usually indicative of process capability to a degree. How quantifiable that is varies by architecture, etc. Having said that...I do not expect the clock speed to hit 4.0 like PD did in lots of chips. However, a 3.7 base clock with 4.0-4.2 turbo could be quite feasible given the fact that it would only be a 700 MHz bump over the A0 engineering sample.

Engineering samples depend on a lot of factors, including how mature is the process node used or how accurate are the design/implementation tools used by the engineers. Quoting base frequencies of ES on different process nodes on so large time span means little.

What matters to know frequencies of final silicon is: (i) microarchitecture details, (ii) process node characteristics, (iii) platform TDP.

Points (i)+(ii)+(iii) together clearly suggest lower frequencies.

~100W TDP suggests nothing in the way of low clocks juan...what was the TDP of Devil's Canyon? 88W?

Apart from ignoring two of my three points, you also ignore that I was discussing the clocks for an "octo-core", and the 88W Devil's Canyon is a quad-core, not to mention it uses a different 22nm process than Zen...

There are lots of different octo-core processors...some have clocks significantly higher than others.

While I understand your desire to speak from an Intel perspective...AMD is clearly not intel...and therefore your observations based on Intel technology are purely anecdotal.

 

[8350rocks]

CERN conference for FastForward2 provides AMD generated specs for future Zen Commercial products: http://vrworld.com/2016/02/12/cern-confirms-amd-zen-high-end-specifications/

 

[-Fran-]

Interesting read. IF that thing is accurate, then I don't think what they mentioned in that CERN talk is about consumer grade stuff. The big hint is "MCM". I *really* don't think AM4 is going to support MCM and not 8 channels. I mean, it would be amazing, but that would skyrocket the price to a level where it would not make it viable for consumers.

It is still nice info though. At least it shows AMD is gonna up the stakes with Zen. Plus that APU sounds like one big mean piece of silicon.

Cheers!

 

[8350rocks]

Interesting read. IF that thing is accurate, then I don't think what they mentioned in that CERN talk is about consumer grade stuff. The big hint is "MCM". I *really* don't think AM4 is going to support MCM and not 8 channels. I mean, it would be amazing, but that would skyrocket the price to a level where it would not make it viable for consumers.

It is still nice info though. At least it shows AMD is gonna up the stakes with Zen. Plus that APU sounds like one big mean piece of silicon.

Cheers!

Yeah, certainly HPC/server class commercial products, but still interesting...

 

[juanrga]

There are lots of different octo-core processors...some have clocks significantly higher than others.

Evidently not all octo-cores are the same and that is why FX-9590 hits 5GHz but Zen octo-core will not.

While I understand your desire to speak from an Intel perspective...AMD is clearly not intel...and therefore your observations based on Intel technology are purely anecdotal.

It is difficult to understand from where you got that false idea. Re-read my posts and check my points (i), (ii), and (iii). I am talking from the perspective of Samsung 14LPP, AM4 platform, and Zen microarchitecture.

 

[juanrga]

CERN conference for FastForward2 provides AMD generated specs for future Zen Commercial products: http://vrworld.com/2016/02/12/cern-confirms-amd-zen-high-end-specifications/

It would be a good thing to split the genuine information found in some CERN slides [1] from the personal speculations of Theo Valich, and this from the plain wrong stuff that he writes (which is a lot of).


[1] Information has been posted here before. E.g. in one of my very early posts I wrote Zen server processors would top at 32 cores.

 

[juanrga]

The legendary overclocker the Stilt has shared his thoughts about the overclocking capabilities of Zen:

Despite I don't expect first generation Zen CPUs to have basically any margin left for overclocking, I certainly hope AMD hasn't made too grave compromises when designing the socket infrastructure.

 

[jimmysmitty]

CERN conference for FastForward2 provides AMD generated specs for future Zen Commercial products: http://vrworld.com/2016/02/12/cern-confirms-amd-zen-high-end-specifications/

Weird to see names of past posters in the comments on that site.

It looks good, just need to know more. Theo has had a history though and I am not sure if it is 100%.

I guess we will know more in a few months.

 

[cdrkf]

The legendary overclocker the Stilt has shared his thoughts about the overclocking capabilities of Zen:

Despite I don't expect first generation Zen CPUs to have basically any margin left for overclocking, I certainly hope AMD hasn't made too grave compromises when designing the socket infrastructure.

I think that sounds sensible- Zen is a first gen new arch, first gen on new process tech and first gen on new platform. That is a *massive* risk, and incidentally was one of the things that contributed to Phenom I falling flat (even though fundamentally it was a pretty good design, which thankfully they got sorted with Phenom II).

This also happened with Bulldozer, with Piledriver improving many of the failings (although obviously with the entire dozer' derived line there are some core issues that can't really be tweaked out).

That is the one big positive with Zen- the info we are seeing points to a more fundamentally sound design from the start. Still lots of things that could go wrong, however even if it doesn't go to plan we are looking at a Phenom type situation rather than a Bulldozer one.

Here's hoping AMD can get a good hit on the first try this time though! Third times the charm and all that 😛