AMD CPU speculation... and expert conjecture

[juanrga]

They make more money back than the case would fine them so they will keep doing it. All the tablets with intel CPUs are being funded by intel like the pc back then were, I wonder when the lawsuit for that would resolve.

I think there are two fundamental differences compared to what happened in PCs back then: (i) Now Intel is the underdog in tablet/phone, thus Intel cannot be accused of monopolistic practices. And (ii) Now Intel is not hiding the fact are promoting its chips (we know even the amount given to OEMs per chip); thus it is not illegal and obscure practice.

 

[de5_Roy]

HP ships "Kaveri" 65 Watt APUs in desktop PCs
http://www.cpu-world.com/news_2014/2014061201_HP_ships_Kaveri_65_Watt_APUs_in_desktop_PCs.html
3 new kaveris finally show up in oem pcs. the latest announcements don't show, but another single module (1m/2c) a4 7300 non-k might show up later, with 65w tdp(configurable). assuming 128/256 shaders @720Mhz and 3.4-3.7Ghz clockrate cpu.
http://www.cpu-world.com/CPUs/Bulldozer/AMD-A4-Series%20A4-7300.html

edit:
CYBERPOWERPC Teams Up with Newegg to Launch $299 "Everyman Quad Core" System
http://www.techpowerup.com/202053/cyberpowerpc-teams-up-with-newegg-to-launch-299-everyman-quad-core-system.html

 

[blackkstar]

If AMD really wants MANTLE to become a Standard API they seriously need to hurry, if they manage to get some Games with MANTLE on Linux and open all documents before DX12 arrives, then they may have something... otherwise DX12 will eat MANTLE for breakfast.

Funny how everyone forgets AMD controls maybe 20% of the GPU market? Unless Intel/NVIDIA jump on board, users see no benefit, and the money flows to other APIs instead. AMD can optimize all they want, but if 80% of users see performance boosts only through DX, then which API is going to sell more product?

You also still have the problem where the implementation is tied to GCN, so what happens when AMD moves on to another GPU arch? Now they have the start to make the processing generic, and all those performance gains start to disappear again.

As long as the evolution from GCN is not so drastic it will be easy to maintain.

Also, picture a world where every steam machine carries AMD GPUs to support mantle on linux better than nvidia can support OGL.

If they play their cards right, and developers take the queues and run, AMD could easily gain 10-15% market share in a matter of 6 months...which would take the currently 35-40% market share they enjoy in GPUs and turn it into roughly half or better.

EDIT: You really, really....REALLY lowballed that one. AMD has much better GPU market share than they do CPU market share...

I didn't say dGPU, did I? You have to factor in Intels share here, and thats about 40% of the total market.

You do realize that there is going to be overlap? Everyone who buys a new Intel CPU on mainstream socket or laptop gets an Intel iGPU with it even if they never even touch it. All those 4770k, 4670k, 3770k, 3670k, 2600k, 2500k systems out there have Intel iGPU and it all counts towards their market share.

It's a bit deceptive to include those numbers for something based on gaming rigs, isn't it? Specially since you are removing AMD share in your estimate to account for systems with Intel iGPU when there are clearly systems out there with Intel iGPU with Nvidia and AMD GPUs.

I glanced over Steam Hardware Survey and did some digging to see if SHW counts the Intel iGPU if it's not being used but I couldn't find an answer. Can anyone confirm?

 

[de5_Roy]

You do realize that there is going to be overlap? Everyone who buys a new Intel CPU on mainstream socket or laptop gets an Intel iGPU with it even if they never even touch it. All those 4770k, 4670k, 3770k, 3670k, 2600k, 2500k systems out there have Intel iGPU and it all counts towards their market share.

It's a bit deceptive to include those numbers for something based on gaming rigs, isn't it? Specially since you are removing AMD share in your estimate to account for systems with Intel iGPU when there are clearly systems out there with Intel iGPU with Nvidia and AMD GPUs.

I glanced over Steam Hardware Survey and did some digging to see if SHW counts the Intel iGPU if it's not being used but I couldn't find an answer. Can anyone confirm?

Spoiler

i don't know about how much nvidia's igp have marketshare in pcs. amd has onboard igps as well as igpus in apus in various price ranges. intel has igpus in various price ranges but doesn't have discreet gfx presence like amd and nvidia. afaik intel's biggest slice of gpu marketshare comes not from the unlocked procssors but from igps like intel gma hd, igpus like hd gfx (sku without numeric number 6-10 eus), 2000, 2500, 4400, 4600 in oem desktops, 3k, 4k, 4400, 4600 in laptop cpus and to a lesser extent, hd5000, 5100, iris and iris pro in high end laptops and some apple macs (hd3k, 4k, 5k and iris/pro). amd's shares exist in igps in -g chipsets e.g. 760g, various apus and socs at various price ranges, all kinds of discreet gfx cards. nvidia's come from discreet gfx only (mobile and desktop, consumer and pro/workstation). nvidia's arm socs are usually excluded in counting pc gpu marketshare. some of intel's ulp socs with hd gfx igpus are counted excluding socs with powervr igpus. despite massive gpu marketshare, intel doesn't seem to be influencial in gfx applications such as gaming. most isvs prefer amd and nvidia.

igpus like hd 3k and 4k do have commanding presence due to their presence in mainstream laptops. amd and nvidia take very distant second and 3rd positions but enjoy much stronger influence over isv.
if they wanted, both amd and nvidia can leverage their isv relationships over intel's marketshare. this is what amd is doing, except amd isn't playing to their full potential. gcn (-based gpus) does have very small market presence. gcn's biggest presence is in console socs which are closed, fixed, non-pc devices. in pcs, gcn's biggest share is in kabini soc based pcs, followed by in some laptops and in diy pcs (as apus and/or dgpus). gcn is not majorly present in laptop apus nor in dt apus (in both those areas older llano and trinity/richland apus reign). so far, mantle doesn't show any advantage in kabini. it'll take over a year or two for gcn to gain a solid standing in pc market. if last quarter was any indication, mantle didn't help amd gain any marketshare. amd will have to act smarter from here on out.

 

[etayorius]

I don`t know about most of you, but i WOULD NEVER touch and Intel IGPU, Maybe... MAYBE if it were for the most basic stuff and there was no other choice, i mean sending emails, web browsing and maybe watching videos.

For Gaming? not me... ever.

 

[de5_Roy]

maybe watching videos.

sorry, if you're gonna watch videos with intel igpu, make sure it's a haswell cpu. older igpus cause headache-inducing stuttering during video playback. even vsync and driver update won't help much.

For Gaming? not me... ever.

this is why kaveri/sr's delay and slow spreading is hurtful for amd. had kaveri come out in time in 2012-2013, we wouldn't be wondering how mantle would stand up by itself.

 

[gamerk316]

If AMD really wants MANTLE to become a Standard API they seriously need to hurry, if they manage to get some Games with MANTLE on Linux and open all documents before DX12 arrives, then they may have something... otherwise DX12 will eat MANTLE for breakfast.

Funny how everyone forgets AMD controls maybe 20% of the GPU market? Unless Intel/NVIDIA jump on board, users see no benefit, and the money flows to other APIs instead. AMD can optimize all they want, but if 80% of users see performance boosts only through DX, then which API is going to sell more product?

You also still have the problem where the implementation is tied to GCN, so what happens when AMD moves on to another GPU arch? Now they have the start to make the processing generic, and all those performance gains start to disappear again.

As long as the evolution from GCN is not so drastic it will be easy to maintain.

Also, picture a world where every steam machine carries AMD GPUs to support mantle on linux better than nvidia can support OGL.

If they play their cards right, and developers take the queues and run, AMD could easily gain 10-15% market share in a matter of 6 months...which would take the currently 35-40% market share they enjoy in GPUs and turn it into roughly half or better.

EDIT: You really, really....REALLY lowballed that one. AMD has much better GPU market share than they do CPU market share...

I didn't say dGPU, did I? You have to factor in Intels share here, and thats about 40% of the total market.

You do realize that there is going to be overlap? Everyone who buys a new Intel CPU on mainstream socket or laptop gets an Intel iGPU with it even if they never even touch it. All those 4770k, 4670k, 3770k, 3670k, 2600k, 2500k systems out there have Intel iGPU and it all counts towards their market share.

It's a bit deceptive to include those numbers for something based on gaming rigs, isn't it? Specially since you are removing AMD share in your estimate to account for systems with Intel iGPU when there are clearly systems out there with Intel iGPU with Nvidia and AMD GPUs.

I glanced over Steam Hardware Survey and did some digging to see if SHW counts the Intel iGPU if it's not being used but I couldn't find an answer. Can anyone confirm?

http://store.steampowered.com/hwsurvey/videocard/

About 15% total, but closer to 30% for DX10.x class GPU's which indicates the group that don't go crazy upgrading their PC's. That's a decent slice, and remember: Steam biases toward gamers. Intel needs to be "good enough" for the general user.

 

[de5_Roy]

Andrew Feldman slips away during AMD reorganization
http://vr-zone.com/articles/andrew-feldman-slips-away-amd-reorganization/79250.html
SeaMicro founder and AMD’s server boss says it’s time to take a break, and the server unit is being left in good hands.

 

[-Fran-]

http://store.steampowered.com/hwsurvey/videocard/

About 15% total, but closer to 30% for DX10.x class GPU's which indicates the group that don't go crazy upgrading their PC's. That's a decent slice, and remember: Steam biases toward gamers. Intel needs to be "good enough" for the general user.

Notebooks.

I have a notebook with Steam in it; I game with it (GTX675M ~ GTX580M), but the Steam survey told me I had a HD4000.

I don't think you're wrong either, but those numbers "for gamers" are not really truthful, because Steam can't get the survey correct most of the time in notebooks. I don't know how big is that pie, but oh well.

Cheers!

 

[etayorius]

I can`t remember where i read something about MANTLE would allow different Radeons to CrossFireX, say a 260X and a 290X, i can`t seem to find where i read this but i remember reading it 3 times because i was wondering if had misreading the article... Can anyone Confirm or Deny this?

 

[juanrga]

I can`t remember where i read something about MANTLE would allow different Radeons to CrossFireX, say a 260X and a 290X, i can`t seem to find where i read this but i remember reading it 3 times because i was wondering if had misreading the article... Can anyone Confirm or Deny this?

MANTLE supports assymetric multi-GPU systems, but the real goal behind it is the use of APU+dGPU, where for instance the GPU may handle the rendering load and offload all post-processing to the APU

http://gearnuke.com/in-depth-look-at-amd-mantle/

 

[juanrga]

AMD organizational change to adapt to new goals

http://www.anandtech.com/show/8174/amd-reorganizes-business-units-names-dr-lisa-su-as-coo

 

[jdwii]

I don`t know about most of you, but i WOULD NEVER touch and Intel IGPU, Maybe... MAYBE if it were for the most basic stuff and there was no other choice, i mean sending emails, web browsing and maybe watching videos.

For Gaming? not me... ever.

Me either yet i would never touch a A10 7850K either i would rather build a Pentium dual core and 250X machine for the same amount of money i built several 400$ builds with a pentium and 250x and in gaming in creamed the A10 and in general apps was around 20% slower when all cores where used. For 300$ i can build this same machine without a Video card and have them buy one down the line. Or they can take that same 400$ machine and upgrade to a I7 on the same board without much on an TDP increase. Sorry but that wins on my custom builds.

 

[jdwii]

Some more test's i'm doing on my 8350fx i found something sad about windows scheduler and yes i am running windows 8.1 with all the updates.
Windows scheduler vs setting Affinity to 4 cores(0,2,4,6) 1 core per module Clocked at 4.3Ghz power savings and turbo off
Affinity set 9655 Kilo Nodes per second
Windows 9384 Kilo Nodes per second
Although a Small difference of 3% but i'll test other benchmarks as well
CineBench
Windows 4.44
Affinity set 4.53
2% Performance Hit
Wprime
Windows 12.223 sec
Affinity set 12.047 sec
1.5% Performance Hit
About a 3% performance hit on GTA4 to when not manually setting the cores to 0,2,4,6
Also under handbrake i found no performance hit when sharing between modules so far the only program.
Hand brake Tested 3 times same video 175MB AVI to MP4
1.53Min setting affinity 0,1,2,3
1.54Min setting affinity 0,2,4,6

 

[8350rocks]

Back from Austin...LOTS of things I cannot talk about, but some things that I can...

1.) K12 is intended specifically for servers, and the single platform boards are intended for servers as well.

2.) AMD is not done in the BIG server space.

3.) Next gen HEDT is being reworked from scratch, it will be x86, and they are saying some pretty interesting things. It will not be an APU, it has an internal code name that has not yet been released so I cannot tell you what it is called, though the comments I got were along the lines of..."Well, who is to say the same die we use for an APU has to have GCN cores at all...?" (Obviously cannot say more...)

4.) Carrizo is going to be a NICE performance increase, in more ways than one...I cannot give details beyond that...though expect large gains in GPU and large gains in CPU performance for multiple reasons.

 

[anxiousinfusion]

So K12 has nothing to do in the HEDT space? I would have thought that to be their reworked architecture to scale across multiple form factors as so many articles claim.

Also, how do we know we can believe you?

 

[jdwii]

Arm in high-end no only in juan's mind i agree with 8350fxrocks findings!

 

[blackkstar]

If AMD really wants MANTLE to become a Standard API they seriously need to hurry, if they manage to get some Games with MANTLE on Linux and open all documents before DX12 arrives, then they may have something... otherwise DX12 will eat MANTLE for breakfast.

Funny how everyone forgets AMD controls maybe 20% of the GPU market? Unless Intel/NVIDIA jump on board, users see no benefit, and the money flows to other APIs instead. AMD can optimize all they want, but if 80% of users see performance boosts only through DX, then which API is going to sell more product?

You also still have the problem where the implementation is tied to GCN, so what happens when AMD moves on to another GPU arch? Now they have the start to make the processing generic, and all those performance gains start to disappear again.

As long as the evolution from GCN is not so drastic it will be easy to maintain.

Also, picture a world where every steam machine carries AMD GPUs to support mantle on linux better than nvidia can support OGL.

If they play their cards right, and developers take the queues and run, AMD could easily gain 10-15% market share in a matter of 6 months...which would take the currently 35-40% market share they enjoy in GPUs and turn it into roughly half or better.

EDIT: You really, really....REALLY lowballed that one. AMD has much better GPU market share than they do CPU market share...

I didn't say dGPU, did I? You have to factor in Intels share here, and thats about 40% of the total market.

You do realize that there is going to be overlap? Everyone who buys a new Intel CPU on mainstream socket or laptop gets an Intel iGPU with it even if they never even touch it. All those 4770k, 4670k, 3770k, 3670k, 2600k, 2500k systems out there have Intel iGPU and it all counts towards their market share.

It's a bit deceptive to include those numbers for something based on gaming rigs, isn't it? Specially since you are removing AMD share in your estimate to account for systems with Intel iGPU when there are clearly systems out there with Intel iGPU with Nvidia and AMD GPUs.

I glanced over Steam Hardware Survey and did some digging to see if SHW counts the Intel iGPU if it's not being used but I couldn't find an answer. Can anyone confirm?

http://store.steampowered.com/hwsurvey/videocard/

About 15% total, but closer to 30% for DX10.x class GPU's which indicates the group that don't go crazy upgrading their PC's. That's a decent slice, and remember: Steam biases toward gamers. Intel needs to be "good enough" for the general user.

I mentioned that point because you were writing off AMD market share to make a point about Mantle not having a decent install base.

I do not know how Intel making iGPUs that can play Farmville has anything to do with Mantle. Am I missing something? Are we at a point where Facebook flash games need Mantle?

The bottom line is that you misrepresented AMD market share.

FX8350 that is some good news. If you had a friend who liked to beat on his FX 8350, how long would you think it would be a wise idea to aim to keep that CPU alive before he expects a decent replacement? What would your words of wisdom be to them? Beat it up or make it last?

 

[etayorius]

AMD organizational change to adapt to new goals

http://www.anandtech.com/show/8174/amd-reorganizes-business-units-names-dr-lisa-su-as-coo

Dang, i was starting to doubt my own self... i thought i may have dreamed about it or got it wrong... thanks.

 

[8350rocks]

So K12 has nothing to do in the HEDT space? I would have thought that to be their reworked architecture to scale across multiple form factors as so many articles claim.

Also, how do we know we can believe you?

K12 is ARM cores ONLY. Internal codename for the x86 project is actually a different name. AMD does not anticipate ARM to enter consumer market space for DT PCs anytime soon.

@juanrga:

Because I know you are going to roll out the slide...that is a reference to ULP SERVERS.

 

[8350rocks]

If AMD really wants MANTLE to become a Standard API they seriously need to hurry, if they manage to get some Games with MANTLE on Linux and open all documents before DX12 arrives, then they may have something... otherwise DX12 will eat MANTLE for breakfast.

Funny how everyone forgets AMD controls maybe 20% of the GPU market? Unless Intel/NVIDIA jump on board, users see no benefit, and the money flows to other APIs instead. AMD can optimize all they want, but if 80% of users see performance boosts only through DX, then which API is going to sell more product?

You also still have the problem where the implementation is tied to GCN, so what happens when AMD moves on to another GPU arch? Now they have the start to make the processing generic, and all those performance gains start to disappear again.

As long as the evolution from GCN is not so drastic it will be easy to maintain.

Also, picture a world where every steam machine carries AMD GPUs to support mantle on linux better than nvidia can support OGL.

If they play their cards right, and developers take the queues and run, AMD could easily gain 10-15% market share in a matter of 6 months...which would take the currently 35-40% market share they enjoy in GPUs and turn it into roughly half or better.

EDIT: You really, really....REALLY lowballed that one. AMD has much better GPU market share than they do CPU market share...

I didn't say dGPU, did I? You have to factor in Intels share here, and thats about 40% of the total market.

You do realize that there is going to be overlap? Everyone who buys a new Intel CPU on mainstream socket or laptop gets an Intel iGPU with it even if they never even touch it. All those 4770k, 4670k, 3770k, 3670k, 2600k, 2500k systems out there have Intel iGPU and it all counts towards their market share.

It's a bit deceptive to include those numbers for something based on gaming rigs, isn't it? Specially since you are removing AMD share in your estimate to account for systems with Intel iGPU when there are clearly systems out there with Intel iGPU with Nvidia and AMD GPUs.

I glanced over Steam Hardware Survey and did some digging to see if SHW counts the Intel iGPU if it's not being used but I couldn't find an answer. Can anyone confirm?

http://store.steampowered.com/hwsurvey/videocard/

About 15% total, but closer to 30% for DX10.x class GPU's which indicates the group that don't go crazy upgrading their PC's. That's a decent slice, and remember: Steam biases toward gamers. Intel needs to be "good enough" for the general user.

I mentioned that point because you were writing off AMD market share to make a point about Mantle not having a decent install base.

I do not know how Intel making iGPUs that can play Farmville has anything to do with Mantle. Am I missing something? Are we at a point where Facebook flash games need Mantle?

The bottom line is that you misrepresented AMD market share.

FX8350 that is some good news. If you had a friend who liked to beat on his FX 8350, how long would you think it would be a wise idea to aim to keep that CPU alive before he expects a decent replacement? What would your words of wisdom be to them? Beat it up or make it last?

Current FX lineup will be supported through the end of the year for sure...

Carrizo may seriously offer a solid replacement, even in an APU (seriously...news of this soon)

As for dCPU, it will be after carrizo (it will be dCPU, too, wish I could say more...) but performance from the new uarch will harken to the glory days...(they let Jim Keller have all the good toys...cannot say more...)

Also by the time HEDT dCPU hits, fab issues will have been resolved...there is another company they are looking at possibly partnering with if things get rough with GloFo and/or TSMC. They will have more options at that node.

EDIT: As a side note...the message about packaging and product on the 295X2 was loud and clear...noise and quality are going to be high priorities from here on out with GPUs.

 

[anxiousinfusion]

4.) Carrizo is going to be a NICE performance increase, in more ways than one...I cannot give details beyond that...though expect large gains in GPU and large gains in CPU performance for multiple reasons.

Carrizo may seriously offer a solid replacement, even in an APU (seriously...news of this soon)

I'll bite. What can you tell us? From what I gather, the only reasons would be the new CPU architecture, HSA additions and DDR4 support.

 

[de5_Roy]

@8350rocks: gotta say, nicely placed info. just enough gap for "but that's not what i said before." later. ;D
btw, can you tell me if a10 7850k's pricing is also under n.d.a.? i wanted to know why amd priced the apu slightly south of $200 (which retailers brought closer to $200 at launch) until the most recent price drop. i want to match your findings with my hypotheses. i am certain that even amd knew that 7850k isn't worth paying that much premium over richland 6800k without the established hsa ecosystem, in 7850k's target market and price range. at least some of these are gonna come out anyway in their next earnings call.

as for the new info, my newbiespeculationizer (nspz) machine says the new x86 core could be a heterogenous multiprocessor uarch with 4-8 big cores with 4-8 small cores like puma+ or cheetah (i can't see a57/k12 in that without massive, incredibly reworked iommu). those take up really little place. or gcn 2.0 cores instead of gcn. 8350rocks never said which gcn iteration.

i think carrizo's (excavator) promised upgrades require at least some sort of cache-reworking. but the real improvement might be in the igpu as usual.

amd should take calxeda and nvidia as cautionary tales because arm's big casino isn't all fun and games. while arm will always win, calxeda closed down and nvidia almost got squeezed out of consumer smartphone/tablet market. nvidia's project debacle.. err... i mean denver is still m.i.a. we know jetson has 4x a15 cores! :lol:

edit:

Steamboy is a handheld Steam Machine
http://www.nextpowerup.com/news/9692/steamboy-is-a-handheld-steam-machine.html
mullins? i don't think intel has a gaming i/gpu.

 

[esrever]

The x86 core will most definitely be in an APU. K12 will be an APU too. Nothing announced so far will feature both ARM and x86 cores. Any HEDT space AMD will enter will be an HSA accelerated one.

 

[8350rocks]

So K12 has nothing to do in the HEDT space? I would have thought that to be their reworked architecture to scale across multiple form factors as so many articles claim.

Also, how do we know we can believe you?

Just FYI, you can follow me on twitter, was posting pics of the trip...@jbazzrea.