AMD A8-8650 APU: A quad-core, single-threaded APU at 3.2 GHz / 3.8 GHz on a configurable TDP. It gets 4 MB L2 cache and 760 MHz Radeon R7 GPU with 384 Sea Islands GCN cores. AMD A10-8850K APU: This flagship part uses 4 cores, 4 threads, 3.7 GHz / 4.1 GHz clocks, Radeon R7 Sea Islands graphics (512 streaming cores, 856 MHz), 4 MB L2 cache and 95W TDP. It should get a price of $149.

What? Quad-core single threaded, this sounds like 1 thread for 4 cores..?

So, since this states it is Excavator Based, can we expect HSA 1.0 in there? I hope so. Price is good. TDP is good and this time they have inside those power control things, so CPU cores will not get choked once iGPU is used. This time around A10-8850k will be extra good for price.

What? Quad-core single threaded, this sounds like 1 thread for 4 cores..?

No it means 1 thread per core...

Anytime I read that AMD is releasing smth to compete with Intel/Nvidia it makes me laugh. Let's just hope they can really compete this year with something. AMD we need competition to Intel/Nvidia to drive their prices down!

No it means 1 thread per core...

Then why the difference in wording:

AMD A8-8650 APU: A quad-core, single-threaded APU at 3.2 GHz / 3.8 GHz on a configurable TDP. It gets 4 MB L2 cache and 760 MHz Radeon R7 GPU with 384 Sea Islands GCN cores. AMD A10-8850K APU: This flagship part uses 4 cores, 4 threads, 3.7 GHz / 4.1 GHz clocks, Radeon R7 Sea Islands graphics (512 streaming cores, 856 MHz), 4 MB L2 cache and 95W TDP. It should get a price of $149.

Then why the difference in wording:

It is just two ways of saying the same thing. Single threading just means the threads do not implement something like hyperthreading.

I understand they want to keep costs down, but APU's really need more then DDR3 to stay current. They should at least have one platform with DDR4 memory support.

Anytime I read that AMD is releasing smth to compete with Intel/Nvidia it makes me laugh. Let's just hope they can really compete this year with something. AMD we need competition to Intel/Nvidia to drive their prices down!

Anytime I read something like this I laugh and wonder under what rock people sometimes hide...;) nVidia makes nothing comparable to AMD's APUs (which is why AMD keeps getting the game console--xBone/PS4--contracts and nVidia keeps losing them), and AMD's APU IGP performance clobbers Intel's by several hundred percent. Yes, that's right: AMD's IGP performance is way out in front of Intel's and has been for years. In discrete gpus I am not aware that nVidia has any advantage over AMD at the moment--nVidia's Titan X is a bad joke, imo. What a waste of time and money. AMD competes very successfully with both of those companies, actually. I can't imagine why you'd think otherwise. In fact, the only place where AMD loses these days is high-end x86 cpus--AMD doesn't compete there because AMD doesn't make a ~$350 consumer cpu at the moment. But AMD's current x86 cpus, in the ~$200 & under price range are actually very competitive with Intel's cpus in the same price range. What's hurting AMD at the moment, and it can't be denied that AMD is hurting, is the fact that they've currently got nothing high-end x86 to sell, and they have no up-to-date core logic chipset to sell with it. You have to go to AMD's APUs to get PCIe3.x support, etc. So you can say that AMD is stumbling at the moment, but you certainly cannot say they are not competitive with Intel and nVidia, besting both companies in several areas as they do.

What? Quad-core single threaded, this sounds like 1 thread for 4 cores..?

No man, this means it wiil be a quad core, but just one single graphic core (single gpu). Is says specifically that it will have 2 clusters / 4 cores. vbetts 4/28/2015 9:11 am EST - Removed the rude part. Please do not be rude to other members.

I understand they want to keep costs down, but APU's really need more then DDR3 to stay current. They should at least have one platform with DDR4 memory support.

I would like to see APUs with faster memory, but I think that issue is not in decision to stay with DDR3. Issue is that AMD's DDR3 memory controller transfer data considerably slower while running same frequency memory in comparison to intel. Even carrizo with 2133MHz ddr3 copies data from one region of memory to another slower than older i5 with 1600MHz. If AMD can solve this, then they would have no need to go for DDR4 with APUs even if they get like 6cores/768SP.

Their Ax platform on FM2 and FM2+ have considerably better memory controllers than what they have on AM3+. There really has not been a new chipset for memory controllers on AM3 ever, just a few code changes and additions built off of prior controllers on the cpu and on the motherboard as well. Their APU's are the only platforms that they've introduced new controllers.

Yes, and it is still area where AMD has room to improve. Maybe they did with this generation, who knows?

Anytime I read that AMD is releasing smth to compete with Intel/Nvidia it makes me laugh. Let's just hope they can really compete this year with something. AMD we need competition to Intel/Nvidia to drive their prices down!

Yeah, because sitting there laughing at them is really going to help. Unless you have a 120+Hz and/or a 4K monitor, chances are, AMD has been good enough for your gaming and daily (non-workstation) needs for a long while. I've been supporting them not because I think they're better but because they need the money and intel doesn't. I also support ARM whenever I can too. I believe that if you want to see change, you have to take matters into your own hands.

Yes, and it is still area where AMD has room to improve. Maybe they did with this generation, who knows?

Kevari scales pretty well to memory frequencies, a lot better than Richland or Trinity at least. They need to get their sata controllers in line too!

Kevari scales pretty well to memory frequencies, a lot better than Richland or Trinity at least. They need to get their sata controllers in line too!

Yes, I agree. But I really wonder how much better would kaveri be if they could even match this: http://browser.primatelabs.com/geekbench3/compare/1998382?baseline=2433637 You have faster kaveri APU on one side and on other side you have my system, where I intentionally lowered memory speed to 1333MHz CL9/9/9/27 1T. And that what makes me wonder is memory bench at bottom. Edit: And as usual, bench reads my CPU clock wrong, it sits at 4.5GHz. So performance wise Kaveri is not bad at all if one can extract similar performance as one benched here. (which is hard due to TDP lock, and reason why I think carrizo will show nice boost)

Quote:

Anytime I read something like this I laugh and wonder under what rock people sometimes hide...;) nVidia makes nothing comparable to AMD's APUs (which is why AMD keeps getting the game console--xBone/PS4--contracts and nVidia keeps losing them), and AMD's APU IGP performance clobbers Intel's by several hundred percent. Yes, that's right: AMD's IGP performance is way out in front of Intel's and has been for years. In discrete gpus I am not aware that nVidia has any advantage over AMD at the moment--nVidia's Titan X is a bad joke, imo. What a waste of time and money. AMD competes very successfully with both of those companies, actually. I can't imagine why you'd think otherwise. In fact, the only place where AMD loses these days is high-end x86 cpus--AMD doesn't compete there because AMD doesn't make a ~$350 consumer cpu at the moment. But AMD's current x86 cpus, in the ~$200 & under price range are actually very competitive with Intel's cpus in the same price range. What's hurting AMD at the moment, and it can't be denied that AMD is hurting, is the fact that they've currently got nothing high-end x86 to sell, and they have no up-to-date core logic chipset to sell with it. You have to go to AMD's APUs to get PCIe3.x support, etc. So you can say that AMD is stumbling at the moment, but you certainly cannot say they are not competitive with Intel and nVidia, besting both companies in several areas as they do.

End Quote: Sorry but You don't know what you are talking about. '...AMD keeps...' Keeps getting what? They were chosen for two consoles of the same generation. So what have they 'kept'? As for 'In fact, the only place where AMD loses these days is high-end x86 cpus'. First that isn't a fact... it isn't even correct. AMD are losing all over the place. http://www.bit-tech.net/news/hardware/2015/04/20/amd-q1-2015-loss/1 Yes, they are doing OK in certain markets but most areas of their business are still losing money despite large turn over and other areas are making huge losses, hence the need for another restructuring. EDIT Quote box doesn't seem to be coming up in standard comment window.

Granted it was ATI before AMD, ATI/AMD has had the Gamecube, Wii, Wii U, and 360 in the console segments. Now they are in the cpu side of console gaming.

Has Amd stated which node they will be using for this apu?

Anytime I read something like this I laugh and wonder under what rock people sometimes hide...;) nVidia makes nothing comparable to AMD's APUs (which is why AMD keeps getting the game console--xBone/PS4--contracts and nVidia keeps losing them), and AMD's APU IGP performance clobbers Intel's by several hundred percent. Yes, that's right: AMD's IGP performance is way out in front of Intel's and has been for years. In discrete gpus I am not aware that nVidia has any advantage over AMD at the moment--nVidia's Titan X is a bad joke, imo. What a waste of time and money. AMD competes very successfully with both of those companies, actually. I can't imagine why you'd think otherwise. In fact, the only place where AMD loses these days is high-end x86 cpus--AMD doesn't compete there because AMD doesn't make a ~$350 consumer cpu at the moment. But AMD's current x86 cpus, in the ~$200 & under price range are actually very competitive with Intel's cpus in the same price range. What's hurting AMD at the moment, and it can't be denied that AMD is hurting, is the fact that they've currently got nothing high-end x86 to sell, and they have no up-to-date core logic chipset to sell with it. You have to go to AMD's APUs to get PCIe3.x support, etc. So you can say that AMD is stumbling at the moment, but you certainly cannot say they are not competitive with Intel and nVidia, besting both companies in several areas as they do.

AMD doesn't have the fastest GPU, their power consumption is off the charts, their current GPU is months late and months away, their APUs are only competitive in a place where middling graphics performance is desirable (so, nowhere) their CPUs are way way behind Intel at every single pricepoint with the exception of some very specific workloads at very specific segments of the market. No one even knows what an Opteron is anymore and the one place they could maybe have an impact, desktop CPUs, they can't even compete because, super ****ing ironically, they don't have a GPU in it, so you have to get a discrete one. If you're ok with that, then you can have a chipset from 2011. Everything they're doing right now is slower than their competitors for more power. Everything. With the sole exception of desktop APUs, which offer useless performance in a useless segment.

Has Amd stated which node they will be using for this apu?

Lisa Su stated that they want to go 20nm before going for 14nm FF next year, but I believe this is 28nm because 20nm proved to be pita and probably one of reasons for AMD not releasing stuff sooner and burning more money. Slides for carrizo mentioned move from CPU friendly process where iGPU burned too much energy to GPU friendly+CPU friendly. I do not consider 20nm to be GPU friendly.