blackkstar :
As for this marketing slide comparing APU to CPU, the point is irrelevant.
1. It's a marketing slide to promote APUs, of course they're going to be using a situation that favors APUs for their benchmark
2. If 4 SR cores are fast enough to compete with FX 6000 PD models then it's foolish to think that that means the APUs are going to replace CPUs. A CPU only SR chip would also increase in performance by similar ratios that the APU did using SR.
1. Sure but one does not promote one product at the expense of another product unless the second product is going to be discontinued or replaced.
2. But one has to look things overall. Kaveri is desktop/mobile version of Berlin. There is no 6 SR core Warsaw due to lack of demand. Therefore a hypothetical six-core SR FX for the desktop would be much more expensive due to the die being exclusive.
Also APUs can be used for mobile. FX CPUs cannot. Again this adds extra cost.
blackkstar :
But the whole point of Mantle is to further destroy Nvidia from the bottom up, as it'll make lower end APUs comparable to higher end Nvidia parts. Realistically I would think that high end APU with Mantle would be competing with GTX 660 Ti and lower at least, but that's just a random guess.
I don't think so. The new 290 and 290X are going to hurt Nvidia with ordinary games, thanks to excellent performance per dollar ratio.
In my opinion Intel is the target of MANTLE. Increasing FPS from 100 to 130 in a high-end dGPU owned by a 1% of gamers is not going to change sales significantly. Increasing from 46 to 60 in a mobile APU is noticeable.
Recall that main competition for Kaveri is not Haswell but Broadwell, which will introduce a mayor improvement in the iGPU (rumored 40% better graphics than Haswell).
Kaveri + MANTLE is what AMD needs to compete with Intel where all the money is (>90% of gamers).
blackkstar :
Tegra is a disaster that doesn't turn profits and doesn't get design wins. Things are so bad for Nvidia that they have to release their own hardware. This is like AMD releasing Temash, absolutely no one using it, and then AMD making their own tablets and laptops.
Tegra had three problems. First, Nvidia was a GPU company who is now learning to do CPUs and you don't master this in a pair of years. Second, Nvidia is competing against true giants such as Qualcomm, Samsung, Apple... who are hurting even to a big player as Intel. Third, several problems of Tegra have been related to the SoC lacking extras such as a modem, not to the CPU.
Things are changing with Tegra 4 and 4i (includes a modem). They are winning more designs. Tegra 4 is now selling very well in phones and Shield. Last numbers show:
http://www.xbitlabs.com/news/graphics/display/20131108231028_Sold_Out_in_86_Seconds_Nvidia_Reveals_Success_of_Xiaomi_Mi3.html
http://www.androidheadlines.com/2013/11/nvidia-still-android-game-tegra-sales-rise.html
blackkstar :
I'm waiting to hear some sort of news at APU13 about dGPU HSA. If it's even remotely hinted at that it's coming then we know almost for sure that there's going to be a dCPU and dGPU HSA platform and that it'll be coming soon.
I expect HSA dGPU for pairing with APUs. I think that the new 290 and 290X already support unified memory addressing.
blackkstar :
I want to hear from juranga (and others who agree with him) how AMD would plan to drive HSA adoption and to promote software developers of professional applications as well as games when the best they offer is a mid-range part.
AMD plans to drive HSA adoption using the HSA foundation. Both consoles use something close to HSA, therefore games will use it.
gamerk316 :
There is no doubt that APU will be faster than some FX chips in a pure CPU benchmark, but of course, the real strength of a APU is in heterogeneous compute. Why would I use an archaic piece of software that uses only a 20% of the performance of the APU?
Because a single compute unit is slow, and if the software doesn't scale close to linearly, you could easily gain -500% performance this way. And yes, I've seen it happen.
What you are trying to do is compare a CPU to a CPU+GPU. The fact both are on the same die, from a benchmarking perspective, is irrelevant. Give the 8350 even a cheap $50 GPU, and the 8350 pulls ahead again. Farther, that argument means you also need to compare with a i7-3770k and its iGPU, which while not powerful, will pull it ahead when compute tasks are considered.
So please, stop it. For purely CPU workloads, the APU is going to be sub-i3 performance.
Do you read? Because I showed that Kaveri is at i5 level of performance
using only CPU workloads.
Also the slide #13 given by AMD recently compared new A10-6970k APU to Piledriver 6350 and 8350 FX chips using BF4 game (
non-HSA) and the leaked chinese benchmarks compared a Kaveri to Bulldozer and Piledriver FX using
two CPU workloads (integer and floating point).
Finally a single compute unit in Kaveri (900 MHz) will provide about 115 GFLOP. The entire FX-6350 (six cores at 3.9GHz) gives ~187 GFLOP. Six compute units at 900MHz give ~691 GFLOP or
3.7x the performance of the FX-6350.
Facebook
Twitter
Instagram