New Zen information, AM3+ info, APU presentation, and video card information

cageymaru

cageymaru

Fully [H]
Joined
Apr 10, 2003
Messages
22,074
CPU presentation seems to start at the 55:00 mark. I find it interesting that the current FX series is being acknowledged for being VR Ready by Valve and others. I skimmed the video so enjoy watching it along with me. ;) Yes, this is an event from Korea and I saw this information on the AMD Reddit.

 
I can't help but to think that the translator gets things wrong but I wouldn't know because I don't speak Korean :)
1:12:16
"we need to be very careful when how we talk about Zen"

Give me a break please. Stop talking release Zen, no one ever wants to hear about Zen we want to buy it.
 
Last edited:
AMD is definitely being a lot more careful this time around compared to when the 8150 was released. They will probably not give engineering samples to known leakers unless they want them to leak the information. Not allowing Intel a heads up on their processor information is a smart move, just as not letting Nvidia know what is up until after release was also a good idea.
 
Spirit_Retro said:
Pretty sure no one outside of AMD has possession of an engineering sample.
Click to expand...
I'm pretty sure they do :) check what Lisa Su said at Computex 2016 they are giving samples to partners in Q3 :) .
ManofGod said:
AMD is definitely being a lot more careful this time around compared to when the 8150 was released. They will probably not give engineering samples to known leakers unless they want them to leak the information. Not allowing Intel a heads up on their processor information is a smart move, just as not letting Nvidia know what is up until after release was also a good idea.
Click to expand...
I don't mind that not everyone gets something but the idea that you can keep quiet and build the suspense is a joke. We all know it is not going to outperform Intel , the cat is out of the bag. Now what they can keep quiet about is how they are pricing it. But then again as soon as they move things through retail someone is going to spill the beans on that anyway.

So mystifying the whole process is the same as hiding sweets from your kids eventually they will figure it out. But if you are hiding sweets , you go on talking about how you are so cool by hiding them and reminding your kids each and every time you see them ?

I do not like this approach at all beside it being childish, you don't treat your customers like little kids ...
 
Last edited:
And still no release... >(

And nothing concrete that would indicate an impending release.

I really would've expected to see a model list, speeds, pricing, etc by now.


How many years have we been waiting for this?

Either this thing is absolutely marvellous or it's a complete clunker.

And what do they keep ranting on about the old FX line for?

They still hold their own but I wouldn't purchase today without a heavy discount.

VR? Who cares? Reminds the old Nintendo Virtual Boy - is anyone in their right mind going to sit there with that on their head?


AM3+ Zen variants or rebranded FX parts? It is a bit strange to see a new Sabertooth board get released. AMD are the masters at recycling technology.

It's probably gonna be around December isn't it?

I'm just happy that they "seem" to have gotten around to fixing the black screen crashes with the Crimson drivers - only took them since December last year.

I actually get a little amazed each additional year that my Phenom x4 955 keeps on giving. I thought it would've long been on the scrap pile by now - lucky that CPU progress basically stopped.
 
Last edited:
mwnn said:
And still no release... >(

How many years have we been waiting for this?

Either this thing is absolutely marvellous or it's a complete clunker.

<snipped>
Click to expand...

This makes sense to me. And the closer we get to Q4 without some kind of info the more it makes me think it's a dud.

And I hope I'm wrong.
 
In all honesty I'm often tempted to just pickup a Phenom x6 off eBay like 1190T / 1100T and wait for Zen to mature / Intel's next.

It's just the bad pricing considering the age of the chip which holds me back.

If I could only get one for £30/$40 instead of the £80+++ they're going for.

I woudn't have to mess around with a complete overhaul then.

Granted the M4A785TD-V EVO isn't a spring chicken.

Any more money than that and it's Intel i3 / i5 time.


AMD keep talking about DirectX12 and their FX line getting boosted as more cores are utilised.

And we've already seen how the old Phenom's got a boost in Ashes, etc.
 
Last edited:
Spirit_Retro said:
Well we're 17 days into Q3. No back alley leaks?
Nah... there'd be something leaking out within 72 hours.
Click to expand...

I never heard of this logic , do engineering samples come with Auto Leak feature pre installed ?
I have already seen a person post about Zen engineering sample which sounded credible few months back.

I doubt that they hand engineering samples out like candy and the NDA covering them prolly is another form where people think twice about leaking anything .....
 
Pieter3dnow said:
I never heard of this logic , do engineering samples come with Auto Leak feature pre installed ?
I have already seen a person post about Zen engineering sample which sounded credible few months back.

I doubt that they hand engineering samples out like candy and the NDA covering them prolly is another form where people think twice about leaking anything .....
Click to expand...

There's leaks for almost any product in technology. Since we've not seen anything substantial, my experience leads me to believe it's not sampling, or vendors do not have engineering samples to build boards for.

For instance: Bulldozer benchmarks were leaked by several sites in advance of an October release. And they were accurate- the chip sucked at release.

How's that for logic? Or do I have to retell the history of pre release leaks that I would have assumed you would have known?
 
Spirit_Retro said:
There's leaks for almost any product in technology. Since we've not seen anything substantial, my experience leads me to believe it's not sampling, or vendors do not have engineering samples to build boards for.
For instance: Bulldozer benchmarks were leaked by several sites in advance of an October release. And they were accurate- the chip sucked at release.
How's that for logic? Or do I have to retell the history of pre release leaks that I would have assumed you would have known?
Click to expand...

Some people learn from their mistakes. What you are saying that even tho it is 6 years later everything will stay the same , which I find a little bit troubling ...
 
Pieter3dnow said:
Some people learn from their mistakes.
Click to expand...

Are you saying I learn from my mistakes? Which isn't really logical since I've not made any yet (that you know of).

Or are you saying AMD learns from their mistakes? In which case your sentence is not appropriately constructed since AMD is a company not a person.

Don't bother answering- I really do not care.

What you are saying that even tho it is 6 years later everything will stay the same , which I find a little bit troubling ...
Click to expand...

No I didn't say that. Here's what was said:

If there are no leaks- no working silicon is outside of AMD <SIC>. Because 95% of the time prior to a tech product release- there are leaks.
 
Spirit_Retro said:
Are you saying I learn from my mistakes? Which isn't really logical since I've not made any yet (that you know of).
Or are you saying AMD learns from their mistakes? In which case your sentence is not appropriately constructed since AMD is a company not a person.
Don't bother answering- I really do not care.
No I didn't say that. Here's what was said:
If there are no leaks- no working silicon is outside of AMD <SIC>. Because 95% of the time prior to a tech product release- there are leaks.
Click to expand...
What you say is not what Lisa Su says , samples for partners in Q3 . That you assume leaks is your problem you expect leaks because that is what happened last time. Last time was 6 years ago.
Now you start to generalize it, leaks will happen because they are happening everywhere ? Good god man give it a rest ...
 
Pieter3dnow said:
What you say is not what Lisa Su says , samples for partners in Q3 . That you assume leaks is your problem you expect leaks because that is what happened last time. Last time was 6 years ago.
Now you start to generalize it, leaks will happen because they are happening everywhere ? Good god man give it a rest ...
Click to expand...

I realize English is a second language for you, or you're a troll.

But we're 17 days into Q3. I'm sure everyone realizes that- except perhaps you.
 
Last edited:
Latest Zen Rumors (note, this is not from me, but something I grabbed from elsewhere):

Hi all, I've got some info about Zen ES stage. I thought you would like to see it after the many off-topic Polaris posts.

Zen ES is at the moment in revision A0 - it might not be a suprise.

L2/L3 variations: 2/8 MB, 4/16 MB, 8/32 MB, 12/64 MB, 16/64 MB
(512kb L2/core, 8MB/4 cores)

Core counts are: 4c/8t, 8c/16t, 16c/32t, 32c/64t. As it seems now there won't be a 6c/12t at the launch, there will be only complete core complexes. Later AMD might release a 6/12 version, will see.

AMD's working on 2 kind of packages: AM4 and SP3. Later there might be a SP4 package of course.

4 variants of ES Zen are available at the moment:
AM4 8 cores with 95W TDP
AM4 4 cores with 65W TDP
SP3 24 cores with 150W TDP
SP3 32 cores with 180W TDP

The most exciting part is core clock. The 8c/95W variant's base clock is 2.8GHz, all core boost is 3.05GHz and maximum boost is 3.2GHz.
The 4c/65W part's clock is the same. (I would expect 3.5GHz base clock for a retail 4c/95W variant.)
Idle clock is exciting as well. AM4 versions can lower the clock to 550 MHz in idle which is a very nice level from an AMD CPU. Idle wattage is 5W for 8c version and 2.5W for 4c version.
The SP3 versions have even lower idle clock: it's only 400MHz. Regarding the boost clocks the 32c/180W version has a 2.9GHz boost clock and the 24c/150W version has a 2.75GHz boost clock.

AFAIK Intel has no answer for the 32c/64t Zen variant, so it could be a great win for AMD on the server market.

What AMD is doing different in case of Zen is the purpose of the CPU. AMD doesn't build it for the future but for the present. Maybe it won't be strong in AVX and FMA but an average user won't realize it because the average user won't use the AVX and FMA capabilities. Average Joe will see that Zen is as fast as Haswell on Intel's side (if AMD won't screw it up) and it's enough for everything what Average Joe does on his PC.

On server side Haswell IPC and high core count with a reasonable price will convince the customers to give AMD a try. And I hope AMD can grab the chance after all the years of failures.​
 
If that rumor is even remotely true, I think again it is going to come down to cost, and for enthusiasts OC ability.

Skylake offered very little gain over Haswell; all it really did is allow newer tech on an intel system.


However, being still a "generation" behind, I think the only way Zen rocks the house, provided it all goes smoothly, is if they end up 1/3 of Intel's prices or thereabouts, like the old days. Then you'll see a new war on our hands.
 
Zion Halcyon said:
If that rumor is even remotely true, I think again it is going to come down to cost, and for enthusiasts OC ability.
Skylake offered very little gain over Haswell; all it really did is allow newer tech on an intel system.
However, being still a "generation" behind, I think the only way Zen rocks the house, provided it all goes smoothly, is if they end up 1/3 of Intel's prices or thereabouts, like the old days. Then you'll see a new war on our hands.
Click to expand...

That quote is from anandtech forums and the person that posted it made a new account ;) . So it is hard to see if this is true or not. Baffled on the 4 cores tho :)
 
Pieter3dnow said:
That quote is from anandtech forums and the person that posted it made a new account ;) . So it is hard to see if this is true or not. Baffled on the 4 cores tho :)
Click to expand...

Yeah, I actually just found the first source a while ago - some dude named AMD Polaris. So yeah, I guess HUGE grain of salt, lol. Although it does fit with the original Zen rumors of it competing with Haswell, before the rumor mill went all wonky with over the top predictions. Not unlike what happened with Polaris, so consider it a lesson learned on my part.
 
Zion Halcyon said:
Latest Zen Rumors (note, this is not from me, but something I grabbed from elsewhere):

Hi all, I've got some info about Zen ES stage. I thought you would like to see it after the many off-topic Polaris posts.

Zen ES is at the moment in revision A0 - it might not be a suprise.

L2/L3 variations: 2/8 MB, 4/16 MB, 8/32 MB, 12/64 MB, 16/64 MB
(512kb L2/core, 8MB/4 cores)

Core counts are: 4c/8t, 8c/16t, 16c/32t, 32c/64t. As it seems now there won't be a 6c/12t at the launch, there will be only complete core complexes. Later AMD might release a 6/12 version, will see.

AMD's working on 2 kind of packages: AM4 and SP3. Later there might be a SP4 package of course.

4 variants of ES Zen are available at the moment:
AM4 8 cores with 95W TDP
AM4 4 cores with 65W TDP
SP3 24 cores with 150W TDP
SP3 32 cores with 180W TDP

The most exciting part is core clock. The 8c/95W variant's base clock is 2.8GHz, all core boost is 3.05GHz and maximum boost is 3.2GHz.
The 4c/65W part's clock is the same. (I would expect 3.5GHz base clock for a retail 4c/95W variant.)
Idle clock is exciting as well. AM4 versions can lower the clock to 550 MHz in idle which is a very nice level from an AMD CPU. Idle wattage is 5W for 8c version and 2.5W for 4c version.
The SP3 versions have even lower idle clock: it's only 400MHz. Regarding the boost clocks the 32c/180W version has a 2.9GHz boost clock and the 24c/150W version has a 2.75GHz boost clock.

AFAIK Intel has no answer for the 32c/64t Zen variant, so it could be a great win for AMD on the server market.

What AMD is doing different in case of Zen is the purpose of the CPU. AMD doesn't build it for the future but for the present. Maybe it won't be strong in AVX and FMA but an average user won't realize it because the average user won't use the AVX and FMA capabilities. Average Joe will see that Zen is as fast as Haswell on Intel's side (if AMD won't screw it up) and it's enough for everything what Average Joe does on his PC.

On server side Haswell IPC and high core count with a reasonable price will convince the customers to give AMD a try. And I hope AMD can grab the chance after all the years of failures.​
Click to expand...

Zion Halcyon said:
If that rumor is even remotely true, I think again it is going to come down to cost, and for enthusiasts OC ability.

Skylake offered very little gain over Haswell; all it really did is allow newer tech on an intel system.


However, being still a "generation" behind, I think the only way Zen rocks the house, provided it all goes smoothly, is if they end up 1/3 of Intel's prices or thereabouts, like the old days. Then you'll see a new war on our hands.
Click to expand...

If the first post I quoted is on the money, then that fits well with what AMD has said about IPC compared to Excavator. It's somewhat telling that the turbo frequency and base clocks are slightly below that of Intel's Haswell and newer offerings. This tells us that the architecture is slightly less efficient. How well it overclocks is anyone's guess, but my guess is that it will require more power to hit higher clocks than Intel's parts do. This brings cooling, power draw and total overclockability into question. Even if you can't quite hit the same numbers it might be a win for AMD provided the pricing is right.
 
if they end up 1/3 of Intel's prices or thereabouts, like the old days.
Click to expand...

I have been watching AMD for over 20 years (and was an AMD fanboy for 1/2 of that) and I don't remember prices ever being 1/3 of what Intel had for something I would call a comparable product. Maybe you are talking about the early 1990s and before. I guess I did have an overdrive 586 100 cpu back in the early 1990s before I went to dual Pentiums. I can't remember the price difference back that far.
 
Last edited:
Zion Halcyon said:
The most exciting part is core clock. The 8c/95W variant's base clock is 2.8GHz, all core boost is 3.05GHz and maximum boost is 3.2GHz.
The 4c/65W part's clock is the same. (I would expect 3.5GHz base clock for a retail 4c/95W variant.)
Click to expand...
How is this exciting? This is terrible news. A 40% IPC bump means nothing with a 25%+ drop in clock speed. 1.4x IPC * 75% clock speed = 5% performance boost over existing chips. 5%, that's it! You can already get a 8c/16t 3.5GHz stock intel part with higher IPC. It's like AMD didn't learn a god damn thing with the whole BD fiasco. Here's a fucking clue AMD, I know people harped on IPC but that's not the real problem, the real problem is single threaded performance, and if this rumor holds true they didn't do jack shit to address that! All they did was changed what comes after "AMD sucks at" from "IPC" to "frequency"! This is just "moar cores!" all over again.
 
-Dragon- said:
How is this exciting? This is terrible news. A 40% IPC bump means nothing with a 25%+ drop in clock speed. 1.4x IPC * 75% clock speed = 5% performance boost over existing chips. 5%, that's it! You can already get a 8c/16t 3.5GHz stock intel part with higher IPC. It's like AMD didn't learn a god damn thing with the whole BD fiasco. Here's a fucking clue AMD, I know people harped on IPC but that's not the real problem, the real problem is single threaded performance, and if this rumor holds true they didn't do jack shit to address that! All they did was changed what comes after "AMD sucks at" from "IPC" to "frequency"! This is just "moar cores!" all over again.
Click to expand...

Don't shoot the messenger. I merelthink requoted it.
 
-Dragon- said:
How is this exciting? This is terrible news. A 40% IPC bump means nothing with a 25%+ drop in clock speed. 1.4x IPC * 75% clock speed = 5% performance boost over existing chips. 5%, that's it! You can already get a 8c/16t 3.5GHz stock intel part with higher IPC. It's like AMD didn't learn a god damn thing with the whole BD fiasco. Here's a fucking clue AMD, I know people harped on IPC but that's not the real problem, the real problem is single threaded performance, and if this rumor holds true they didn't do jack shit to address that! All they did was changed what comes after "AMD sucks at" from "IPC" to "frequency"! This is just "moar cores!" all over again.
Click to expand...

Actually DX12 and Vulkan would work well with more cores.
But explain to me how AMD within a reasonable budget can surpass their IPC (instructions per cycle (yes single thread performance))problems without having a budget for it ?

"Everyone" wants AMD to make something that can beat Intel, that is not the point the point is that AMD is trying to do this and if it was easy AMD would never have struggled at it.
Zion Halcyon said:
Don't shoot the messenger. I merelthink requoted it.
Click to expand...
No no no we clearly established it, it is your fault :)
 
Pieter3dnow said:
"Everyone" wants AMD to make something that can beat Intel, that is not the point the point is that AMD is trying to do this and if it was easy AMD would never have struggled at it.
Click to expand...

It seems like the new plan is release products that are wallet friendly and at least give decent performance. It seems like that is what they're doing with their new graphics cards, why would the CPU side be different?
 
scojer said:
It seems like the new plan is release products that are wallet friendly and at least give decent performance. It seems like that is what they're doing with their new graphics cards, why would the CPU side be different?
Click to expand...

Because, they already made it clear that the RX480 is their mid range card. I seriously doubt their newest processor on the top end is only going to be good enough to be mid range.
 
drescherjm said:
I have been watching AMD for over 20 years (and was an AMD fanboy for 1/2 of that) and I don't remember prices ever being 1/3 of what Intel had for something I would call a comparable product. Maybe you are talking about the early 1990s and before. I guess I did have an overdrive 586 100 cpu back in the early 1990s before I went to dual Pentiums. I can't remember the price difference back that far.
Click to expand...

I prefer AMD and will buy them only if that is an option. However, you have reminded me of the Intel Pentium Overdrive chip I had back then, those were the days.
 
Zen may simply not clock well at all. In those that hopes on more than the current 2.8Ghz clock. 14LPP is part of that problem.

Hdgkv0F.png


But the uarch is also part of it. 40% IPC is easy to reach as such, high IPC and high clock, now that's the challenge.

Also we have to see actual benches, AMD only need to hit 40% IPC in 1 workload to claim its true. 40% IPC may not translate into 40% performance at all either. If AMD had such a great product, they would be showcasing it already.

4C/8T at 65W for 2.8Ghz/3.2Ghz. To compare Skylake reaches 2.8Ghz/3.4Ghz at 35W, including IGP.
 
flatty said:
anyway, seems some in consumer advantage... but this would be in a perfect world lol, in the real world, guess who is the winner? :)
Click to expand...
I would say we need something real rather then soft spoken words and other hints on Zen
scojer said:
OR, show their next CPU earlier than they anticipated.
Click to expand...
I doubt that would work, AMD needs to create some demand for a new platform , we already seen that AM4 is a new platform but has no boards and no chips . Desktop parts Bristol Ridge then Summit Ridge (Zen) but they are all missing....
Shintai said:
Zen may simply not clock well at all. In those that hopes on more than the current 2.8Ghz clock. 14LPP is part of that problem.
But the uarch is also part of it. 40% IPC is easy to reach as such, high IPC and high clock, now that's the challenge.
Also we have to see actual benches, AMD only need to hit 40% IPC in 1 workload to claim its true. 40% IPC may not translate into 40% performance at all either.
Click to expand...
I think I have heard this one before. here is something which will make a lot of sense :)
Lets make a new cpu on a new process then select the process on which we can absolutely not gain (which could work if the architectural changes are ground breaking) any ground on Intel because we can not bump enough hertz into the process.
How plausible does that sound ?

What AMD wants is a 8 core 16 threads chip at 95 Watt and within those perimeters trying to do the best they can.
 
If that bit about the 5w/2.5w idle power consumption is true, I might have to consider replacing my 2500k in my file/plex/minecraft server, as the load is usually very low, and I've been looking around for solutions like this to keep things cool and sipping power.
 
Shintai said:
Zen may simply not clock well at all. In those that hopes on more than the current 2.8Ghz clock. 14LPP is part of that problem.


But the uarch is also part of it. 40% IPC is easy to reach as such, high IPC and high clock, now that's the challenge.

Also we have to see actual benches, AMD only need to hit 40% IPC in 1 workload to claim its true. 40% IPC may not translate into 40% performance at all either. If AMD had such a great product, they would be showcasing it already.

4C/8T at 65W for 2.8Ghz/3.2Ghz. To compare Skylake reaches 2.8Ghz/3.4Ghz at 35W, including IGP.
Click to expand...

The design is vastly different between cpu and gpu. Generally speaking most gpu's have much lower clock rates due to these variables.
A. Core counts are high on gpus resulting in larger variance of speed between cores.
B. The die size on gpu's is much larger compared to cpu
C. cpus have a Less complex design, no need for multiple gddr memory controllers, etc etc allows for extra die space to be used to increase contact and interconnect size allowing for more current to be used before electromigration becomes an issue resulting in thermal runaway or worse.

If the rumored numbers are correct
4c/8t zen running at 3.5ghz base with a 65 watt tdp. The I7-6700 has a 3.4ghz base and 4ghz turbo in a 65 watt TDP. Not sure where you got 2.8ghz/3.4ghz at 35 w tdp.

If the rumors are true for the 8c/16t then they are more efficient than broadwell-e in terms of clock speed and TDP.
6900k 8c/16t- 3.2ghz base/ 3.7ghz turbo 140 watt TDP = 22.9 mhz per watt TDP base clock, 26.4mhz per watt TDP boost clock
Zen 8c/16t- 2.8ghz base/ 3.2ghz turbo 95 watt TDP= 29.5 mhz per watt TDP base clock, 33.7 mhz per watt TDP boost clock

Now it doesn't mean it's faster, but if the rumors are true it is more efficient per clock than Broadwell-E, and on par with Skylake. Again TDP is different than actual power consumption, but they are usually pretty close, and I am in no way predicting Zen actual performance.

I suspect Zen will be very competitive with Broadwell-E/ Haswell-E. Skylake will likely be a tad faster, but not by much and AMD will likely give you more cores for the $.
 
ebduncan said:
The design is vastly different between cpu and gpu. Generally speaking most gpu's have much lower clock rates due to these variables.
A. Core counts are high on gpus resulting in larger variance of speed between cores.
B. The die size on gpu's is much larger compared to cpu
C. cpus have a Less complex design, no need for multiple gddr memory controllers, etc etc allows for extra die space to be used to increase contact and interconnect size allowing for more current to be used before electromigration becomes an issue resulting in thermal runaway or worse.
Click to expand...

The CPU is the hard one to make. Not the GPU. AMD doesn't make their own memory controllers anymore either, its all done by Synopsys. Its also a heavy synthethized design, and that's not positive. We already got multiple examples of this.

ebduncan said:
If the rumored numbers are correct
4c/8t zen running at 3.5ghz base with a 65 watt tdp. The I7-6700 has a 3.4ghz base and 4ghz turbo in a 65 watt TDP. Not sure where you got 2.8ghz/3.4ghz at 35 w tdp.
Click to expand...

i7 6700T. Not to mention Intel is already in front in terms of IPC with almost 100%. IPC doesn't always translate into direct performance. And who knows what the 40% claim is based on. They only need to hit it in one single thing to make it valid.

ebduncan said:
I suspect Zen will be very competitive with Broadwell-E/ Haswell-E. Skylake will likely be a tad faster, but not by much and AMD will likely give you more cores for the $.
Click to expand...

Zen is designed for servers. If it cant beat Skylake-E/EP (It is a 2017 product) in perf/watt. It will fail on TCO alone, as in pretty much 0 sales. And more cores/$ we already tried didn't we. Everyone didn't buy FX CPUs for that.
 
Last edited:
You must log in or register to reply here.
Back
Top