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

Sad. I like pulling for the underdog, but the Win10 upgrade deadline forced my hand. I bought a Skylake i5 combo from Microcenter. I'm sure I"m not the only one who bought hardware and a cheap win7 key before the deadline. Lots of people won't consider an upgrade for quite a while now. AMD's timing couldn't be much worse.
 
pgaster said:
Sad. I like pulling for the underdog, but the Win10 upgrade deadline forced my hand. I bought a Skylake i5 combo from Microcenter. I'm sure I"m not the only one who bought hardware and a cheap win7 key before the deadline. Lots of people won't consider an upgrade for quite a while now. AMD's timing couldn't be much worse.
Click to expand...

The Win10 key can be transferred to a new build. Basically, I think you can deactivate then reinstall it with your Microsoft account using the Anniversary update.
 
dgz said:
It's no surprise but is it known what's taking so long? I read something about chipset not being ready but I think it was here
Click to expand...

Bristol Ridge uses the same motherboards so it is not the chipset.
 
14nm is hard. Also the new XBOX S is using the 14nm APU right? So they are probably using it as a test bed for the future as I can't imagine everyone with a XBOX One buying 3 consoles in total after Scorpio launches next year. Also AMD learned from the Hector Ruiz years to not contract out too much fab space in advance. For years they paid hundreds of millions to GloFlo for chips that were never run.
 
pgaster said:
Sad. I like pulling for the underdog, but the Win10 upgrade deadline forced my hand. I bought a Skylake i5 combo from Microcenter. I'm sure I"m not the only one who bought hardware and a cheap win7 key before the deadline. Lots of people won't consider an upgrade for quite a while now. AMD's timing couldn't be much worse.
Click to expand...
I've jumped between amd motherboards, cpus, over to intel and back (with win10, not sure if they track my key as an oem key anymore, but it started out on my laptop as a win7 oem install), and all I had to do was call microsoft and enter the code they gave me to reactivate.
 
Saw this on Reddit about the DDR4.

Rambus Announces Silicon-proven R+ DDR4 PHY on GLOBALFOUNDRIES 14nm LPP Process for Networking and Data Center Applications
Rambus Announces Silicon-proven R+ DDR4 PHY on GLOBALFOUNDRIES 14nm LPP Process for Networking and Data Center Applications

AMD will probably use the DDR4 PHY of RAMBUS in Zen.
http://www.bitsandchips.it/52-english-news/7322-amd-will-probably-use-the-ddr4-phy-of-rambus-in-zen

AMD Extends Patent License Agreement with Rambus.
AMD Extends Patent License Agreement with Rambus - Rambus

Good? Bad? What do you'll think?
 
Something I noticed on that Roadmap is that those all have an "AMD PRO" badge, which correct me if I'm wrong but isn't that the FirePro "Workstation" badge that they had been rolling out on their "FirePro APU" chips (socketed and embedded)?

If so, would that roadmap still hold true for the FX line's release schedule? As from what I gathered, they were going to release Zen on Desktop first, instead of Server first, which I'm not entirely sure if they'd consider the FirePro chips to release then or not. Granted, I know that those "Pro" and "Business" chips aren't really any different than their desktop/mobile counterparts, and it's just different clocks and/or memory controller changes.

Not only that, but I'm pretty certain that Lisa Su had even said during.... was it Computex (or one of those trade shows around June)... that the Zen would be available in small numbers for both retail and system builders by the end of this year (2016) Making me wonder if that Roadmap pre-dates the trade show in question. Unless I see/hear otherwise from AMD themselves, I'm sticking with what Lisa stated regarding it coming out around November-December.

However, I've been at the cabin here in Alaska, so have not been keeping that close of tabs on the industry since July 6th when I got here :p (Speaking of which, I can't remember what city Steve lives in, I was going to see if he'd take me for a spin in that Camaro ;) lol)
 
Formula.350 said:
Something I noticed on that Roadmap is that those all have an "AMD PRO" badge, which correct me if I'm wrong but isn't that the FirePro "Workstation" badge that they had been rolling out on their "FirePro APU" chips (socketed and embedded)?

If so, would that roadmap still hold true for the FX line's release schedule? As from what I gathered, they were going to release Zen on Desktop first, instead of Server first, which I'm not entirely sure if they'd consider the FirePro chips to release then or not. Granted, I know that those "Pro" and "Business" chips aren't really any different than their desktop/mobile counterparts, and it's just different clocks and/or memory controller changes.

Not only that, but I'm pretty certain that Lisa Su had even said during.... was it Computex (or one of those trade shows around June)... that the Zen would be available in small numbers for both retail and system builders by the end of this year (2016) Making me wonder if that Roadmap pre-dates the trade show in question. Unless I see/hear otherwise from AMD themselves, I'm sticking with what Lisa stated regarding it coming out around November-December.

However, I've been at the cabin here in Alaska, so have not been keeping that close of tabs on the industry since July 6th when I got here :p (Speaking of which, I can't remember what city Steve lives in, I was going to see if he'd take me for a spin in that Camaro ;) lol)
Click to expand...

AMD is now more marketing minded then before so it start to designate stuff more so it would seem more appealing.
You remember correctly about Lisa Su saying that it should launch this year that was what I was hoping as well. I tend to post stuff and link the source for it if there is a good reason to post it here so you can make your own mind up on how you want to interpret it (source credible or not).

cageymaru said:
Saw this on Reddit about the DDR4.

Rambus Announces Silicon-proven R+ DDR4 PHY on GLOBALFOUNDRIES 14nm LPP Process for Networking and Data Center Applications
Rambus Announces Silicon-proven R+ DDR4 PHY on GLOBALFOUNDRIES 14nm LPP Process for Networking and Data Center Applications

AMD will probably use the DDR4 PHY of RAMBUS in Zen.
AMD will probably use the DDR4 PHY of RAMBUS in Zen

AMD Extends Patent License Agreement with Rambus.
AMD Extends Patent License Agreement with Rambus - Rambus

Good? Bad? What do you'll think?
Click to expand...

I had to look up what PHY means. It is just an interface for DDR4. It should be fine :)

That manufacturing deal with Samsung is prolly to start making a boatload of those Zen CPU for server and Desktop for first half of 2017 ?
 
cageymaru said:
Good? Bad? What do you'll think?
Click to expand...

AMD doesn't make their own memory controllers. So they license designs from others. RAMBUS would be a change from Synopsys that they currently use. Good? Bad? More likely same thing.
 
Yeah, that is true, I don't see much change from the memory side of things.
 
cageymaru said:
The New Citavia Blog: Some AMD Zen leaks (ES clocks, PCI info, Rambus DDR PHY, Roadmaps)
Click to expand...

Pardon my ignorance, but can you shed some light on what the highlighted multiples refer to? (I crossed out the HDMI HD Audio since I know that's on the AMD card being used)
Zen_ES_PCI-IDs.png


I mean, I can't imagine they're PCIe lanes, sine just the last x16 entries would total to 112 lanes. A few things I do know is that there are a dedicated amount of lanes that AMD chips have that are not counted as part of the "Available Lanes" that are specified on a chip, which are used for things like SATA, Networking, USB, etc. Similarly, my understanding was that they also tied up (usually) 4 lanes for some of them, but are only PCIe 1.0 speed and so a "x4" could be created from a single PCIe 3.0 lane. That's about the extent of it though... Or are some of these indicating
 
How do you know the audio device is on the cpu? It's probably on the video card.
 
Looks like maybe the maximum number of lanes that could be dedicated to that device, whereas the number outside the parentheses is the number of lanes being used. If I knew what program generated the output I could probably find out for sure.

Darakian I think that's what he was saying, but either way it may still have a pcie lane dedicated to it--it has to get the audio data to the hdmi controller from the cpu one way or another.
 
In what universe do you need 16 PCIe lanes for an audio device? 0.01 PCIe lane would probably be more than sufficient since honestly how much has computer audio improved since ISA days? We have a pretty hard limitation in wetware relative to everything else a computer can do. Or SATA for that mater, hell most high end SAS RAID cards are 8x. Maybe they're muxing PCIe lanes for some applications?
 
Good change that CPU-Z is not reading certain things correctly what is known however is that the server part does have stupid amount of PCIe lanes.
But that boost clock is something which I find interesting, the 3.2 ghz should be able to do 200-300mhz higher unless the chip is very sensitive with luck have a decent cooler and do that on air ...
 
Key word is may. In other words no confidence. And it may just be 10 units shipped on dec 31st to get a 2016 checkbox.

Its clear you are not going to be able to buy one before 2017.

Kaby Lake U and Y models have shipped for a month now to compare. Yet you cant buy a single one yet due to the chain.
 
Darakian said:
How do you know the audio device is on the cpu? It's probably on the video card.
Click to expand...
Nobu said:
Darakian I think that's what he was saying, but either way it may still have a pcie lane dedicated to it--it has to get the audio data to the hdmi controller from the cpu one way or another.
Click to expand...
Indeed. What I meant by "AMD card" was the "VGA Controller x1 (x16) AMD Cedar Pro (Radeon HD 5450/Radeon HD 6350)"

My reasoning behin implying that the HDMI Audio listed was tied to that card is based on 2 things:
1) I use AMD myself (for the entire build), as well as HDMI video+audio from my R9 390, as my 'monitor' is my 46" Samsung Full HDTV (which, then, I use its line-out to my speakers), therefore, I know that Windows will display entries like that for HDMIt audio 'device'.
2) It actually says "High Def Audio x1 (x16) AMD Cedar/Park HDMI Audio", which seems rather conclusive that they are tied together.

Furthermore, but prefaced with this disclaimer that I am not entirely certain how to read [understand] everything on that Device ID dump, but all the numbers on the LEFT side of each entry seem to correlate with each other in some way... For example the Graphics is [5 - 00 - 0] and the HDMI Audio below it is [5 - 00 - 1] which seems to indicate that they are tied together and using the same interface, slot, or hard-wired PCIe lane/s.

All in all that might have somewhat answered my initial question about what all those (x#) mean, and that you only add up those multiples from the individual 'devices' inside the brackets on the left. So any that start with the same first-number are all utilizing the same number of lanes in the parenthesis. Which again, for example, means that the HDMI Audio is not using 16 Lanes itself, only that it's sharing the 16 Lanes that the Graphics Card is allocated.

As far as the leading number goes just before what I'm speculating as being the Lanes count, I assumed it was the number of devices in the system, given there's 1 graphics card using 16 Lanes. Similarly, indicating there are 4x USB 3.1 ports, utilizing (I presume) 4 Lanes; however, those 4 Lanes being shared by the 4 SATA ports (or pairs of ports, so 8 total?) and the first PCI Bridge, as they all say [1 - 00 - #]


-Dragon- said:
In what universe do you need 16 PCIe lanes for an audio device? 0.01 PCIe lane would probably be more than sufficient since honestly how much has computer audio improved since ISA days? We have a pretty hard limitation in wetware relative to everything else a computer can do. Or SATA for that mater, hell most high end SAS RAID cards are 8x. Maybe they're muxing PCIe lanes for some applications?
Click to expand...

Just to touch on this quickly, since I more or less covered it above... The audio doesn't need 16, and they never use 16 either. Those 16 are only being listed on that device because it's integrated into the graphics card... which is what is using 16 Lanes. Most dedicated sound cards these days that are PCIe are a x1 device, which yes, even that single lane at PCIe 1.0 specs would (I suspect) be enough or even uncompressed DTS Surround. So yea, no worries, the "HDMI Audio" is not allocating 16 Lanes for its own use lol Simply is sharing those 16 from the GPU :)
 
My (poorly made) point was I don't think those are referring to PCIe lanes since none of that stuff could possibly use 16 of them, unless zen has 16 lanes (or 16 lane equivalent instead of intel's 4 lane equivalent) to the AMD's PCH equivalent or south bridge or whatever it's called these days. SATA, xHCI (USB controller), audio are all things that might be part of a chipset.
 
AM4 with Zen and Bristol Ridge is an x16 slot (x8 on Bristol Ridge) for a discrete GPU essentially. And then an x4, just as Intel for the chipset. Its not without reason the former name of AM4 was FM3.

amd-soc-and-promontory-chipset-io-coverage-1024x724.jpg

AMD_Bristol_Ridge_Architecture.jpg

AMD-Summit-Ridge.jpg
 
AOTS CPU benchmark.

Summit Ridge 8C/16T ES - 2.8-3.2 GHz (2017 Zen)
Average: 58.9 FPS
Normal batch: 65.8 FPS
Medium batch: 62.8 FPS
Heavy batch: 50.5 FPS

Ashes of the Singularity

Core i7-980 6C/12T - 3.33 GHz (2010 Westmere)
Average: 58.6 FPS
Normal batch: 65.2 FPS
Medium batch: 59.7 FPS
Heavy batch: 52.3 FPS

Ashes of the Singularity

Core i7-6700K 4C/8T - 4.0 GHz (2015 Skylake)
Average: 107.3 FPS
Normal batch: 125.7 FPS
Medium batch: 113.8 FPS
Heavy batch: 89.2 FPS

Ashes of the Singularity
 
Last edited:
Shintai said:
AOTS CPU benchmark.

Summit Ridge 8C/16T ES - 2.8-3.2 GHz (2017 Zen)
Average: 58.9 FPS
Normal batch: 65.8 FPS
Medium batch: 62.8 FPS
Heavy batch: 50.5 FPS

Ashes of the Singularity

Core i7-980 6C/12T - 3.33 GHz (2010 Westmere)
Average: 58.6 FPS
Normal batch: 65.2 FPS
Medium batch: 59.7 FPS
Heavy batch: 52.3 FPS

Ashes of the Singularity

Core i7-6700K 4C/8T - 4.0 GHz (2015 Skylake)
Average: 107.3 FPS
Normal batch: 125.7 FPS
Medium batch: 113.8 FPS
Heavy batch: 89.2 FPS

Ashes of the Singularity
Click to expand...
I don't think summit ridge is Zen, it is Carrizo.
 
Shintai said:
AOTS CPU benchmark.

Summit Ridge 8C/16T ES - 2.8-3.2 GHz (2017 Zen)
Average: 58.9 FPS
Normal batch: 65.8 FPS
Medium batch: 62.8 FPS
Heavy batch: 50.5 FPS

Ashes of the Singularity

Core i7-980 6C/12T - 3.33 GHz (2010 Westmere)
Average: 58.6 FPS
Normal batch: 65.2 FPS
Medium batch: 59.7 FPS
Heavy batch: 52.3 FPS

Ashes of the Singularity

Core i7-6700K 4C/8T - 4.0 GHz (2015 Skylake)
Average: 107.3 FPS
Normal batch: 125.7 FPS
Medium batch: 113.8 FPS
Heavy batch: 89.2 FPS

Ashes of the Singularity
Click to expand...
Is there a FX-8350 or better yet FX-9590 version of that? I mean I'm a self admitted AMD hater due to what I perceive as a long series of serious management blunders and missteps over the years combined with a few outright lies ("Fury is an overclockers dream" LOLOLOL also the BD shilling fiasco with JF-AMD) leaving the PC market in the current state of zero competition for intel and NVidia at the high end, and pretty much constantly shits on their fan base by years of sub-par products, but if this is true this isn't good for anybody. I mean a 40% IPC boost is great but when it comes with a 25% drop in clock speed that's practically no gain at all. I'm wondering if AMDs initial implementation of SMT will be as bad as intel's in the P4 series where you were at times better turning it off and they ended up just taking that feature out for a few generations to get it working right again.
 
Last edited:
I dont think they had an option as such, 14LPP cant compete with 32nm SOI in terms of clocks. They simply traded one thing for another with this beefed cat core style, its still 4 core clusters as well. Remember AOTS is very threaded, so its a tad unfair compare directly with the FX as well. Since the FX gets more module penalties and is 4/8, while Zen is 8/16. But you can find the results and compare.

Its all about R&D size. While its no guarantee to get what you pay for, you certainly dont get what you didn't pay for.
 
Shintai said:
AOTS CPU benchmark.

Summit Ridge 8C/16T ES - 2.8-3.2 GHz (2017 Zen)
Average: 58.9 FPS
Normal batch: 65.8 FPS
Medium batch: 62.8 FPS
Heavy batch: 50.5 FPS

Ashes of the Singularity

Core i7-980 6C/12T - 3.33 GHz (2010 Westmere)
Average: 58.6 FPS
Normal batch: 65.2 FPS
Medium batch: 59.7 FPS
Heavy batch: 52.3 FPS

Ashes of the Singularity

Core i7-6700K 4C/8T - 4.0 GHz (2015 Skylake)
Average: 107.3 FPS
Normal batch: 125.7 FPS
Medium batch: 113.8 FPS
Heavy batch: 89.2 FPS

Ashes of the Singularity
Click to expand...

Interesting considering I distinctly remember getting a lower score with a 6700k at 4.5Ghz and a 980 Ti that ran at 1.4GHz out of the box.
 
Well they lost my sale. I've waited 4 years, time for an upgrade this christmas.
 
I find it interesting that the current FX series is being acknowledged for being VR Ready by Valve and others

Half Life 3 Confirmed :p
 
AotS is very much not a good representation of CPU performance in anything other than AotS. Although it would be nice to get some numbers in, I wonder when those are coming.
 
Shintai said:
AOTS CPU benchmark.

Summit Ridge 8C/16T ES - 2.8-3.2 GHz (2017 Zen)
Average: 58.9 FPS
Normal batch: 65.8 FPS
Medium batch: 62.8 FPS
Heavy batch: 50.5 FPS

Ashes of the Singularity

Core i7-980 6C/12T - 3.33 GHz (2010 Westmere)
Average: 58.6 FPS
Normal batch: 65.2 FPS
Medium batch: 59.7 FPS
Heavy batch: 52.3 FPS

Ashes of the Singularity

Core i7-6700K 4C/8T - 4.0 GHz (2015 Skylake)
Average: 107.3 FPS
Normal batch: 125.7 FPS
Medium batch: 113.8 FPS
Heavy batch: 89.2 FPS

Ashes of the Singularity
Click to expand...

If you can give me the exact configuration for the benchmark I'll run it on a 980/9590 setup.
 
Quix said:
AotS is very much not a good representation of CPU performance in anything other than AotS. Although it would be nice to get some numbers in, I wonder when those are coming.
Click to expand...

isn't AotS the benchmark tool used by AMD users by default to show "their massive superior CPU performance under DX12?" I thought it was that way..
 
Araxie said:
isn't AotS the benchmark tool used by AMD users by default to show "their massive superior CPU performance under DX12?" I thought it was that way..
Click to expand...

CPU? Hell no, you mean GPU. CPU....I believe everyone acknowledges Intel is king. If I'm correct, AotS uses a lot of features that are expected in DX12, but whether they are used by other games is obviously not known yet, CPU/GPU.
 
You must log in or register to reply here.
Back
Top