Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
AMD Discussion
- Thread starter koyoot
- Start date
- Sort by reaction score
Links? I don't think that I've ever found an AMD motherboard w/ Thunderbolt (and I've looked).
Sure, here is a news article discussing some older ones. They rely on a add in card but this wasn't uncommon for intel systems.
The biggest problem with AMD platform and Thunderbolt is that the cost of the chipset is between 6 and 8 USD, but overall cost add-on to the motherboard is 20-25$. And AMD wanted to have the most, but not for the sake of more expensive platform than Intel.
On the other hand, prepare, that Zen+ Motherboards might have Thunderbolt native. There will be new models released in upcoming years and are already developed. I heard about MSI model with TB3 or 4, but it can just be a rumor.
Like I have said before many times. AMD's marketing team is rubbish. They have genuinely great hardware, that offers phenomenal value, and they should be shouting, and screaming to the world about this, pre-release, yet they have shut up their mouths. Well, they want to hardware speak for itself, but... it not always sell the hardware.
This is the best example what differs AMD from Nvidia: AMD is run by Engineers. Nvidia is run by people from Marketing department. Intel is run by technology.
Nvidia and Intel sell the most hardware, regardless of how good or bad it is. AMD sells small amounts of hardware regardless of how good or bad it is.
On the other hand, prepare, that Zen+ Motherboards might have Thunderbolt native. There will be new models released in upcoming years and are already developed. I heard about MSI model with TB3 or 4, but it can just be a rumor.
Like I have said before many times. AMD's marketing team is rubbish. They have genuinely great hardware, that offers phenomenal value, and they should be shouting, and screaming to the world about this, pre-release, yet they have shut up their mouths. Well, they want to hardware speak for itself, but... it not always sell the hardware.
This is the best example what differs AMD from Nvidia: AMD is run by Engineers. Nvidia is run by people from Marketing department. Intel is run by technology.
Nvidia and Intel sell the most hardware, regardless of how good or bad it is. AMD sells small amounts of hardware regardless of how good or bad it is.
Intel and Nvidia have an easier time selling hardware because it is objectively better. Show me what AMD has that can compete with a pascal based Titan X or a GTX 1080. Show me an AMD CPU on the market that can beat an Intel Core i7 7700K or 6950X.
I don't mean to start a flame war here, but AMD has some catching up to do. They may pull it off with Zen and Vega, but that still remains to be seen.
I do not mean to spin this into Nvidia vs AMD or Intel vs AMD, or Intel vs. Nvidia. Its just what I have written, AMD marketing sucks. If it would not, they would bring competitor for GTX 1080, for example.
This is the Mac Pro forum - show me what AMD has that can compare with a pair of Intel® Xeon® Processor E5-2699 v4 (44 cores/88 threads, 1.5 TiB RAM, 80 PCIe 3.0 lanes...).
Don't compare them with the 3½ year old single socket systems that Apple labels as "Pro".
Remember this?
In the center of the EHP are two CPU clusters, each consisting of four multi-core CPU chiplets stacked on an active interposer base die. On either side of the CPU clusters are a total of four GPU clusters, each consisting of two GPU chiplets on a respective active interposer. Upon each GPU chiplet is a 3D stack of DRAM (e.g., some future generation of JEDEC high-bandwidth memory (HBM) [12]). The DRAM is directly stacked on the GPU chiplets to maximize bandwidth while minimizing memoryrelated data movement energy and total package footprint. CPU computations tend to be more latency sensitive, and so the central placement of the CPU cores reduces NUMA-like effects by keeping the CPU-to-DRAM distance relatively uniform. The interposers underneath the chiplets provide the interconnection network between the chiplets along with other common system functions. Interposers maintain high-bandwidth connectivity among themselves by utilizing wide, shortdistance, point-to-point paths.
The EHP uses eight GPU chiplets. Our initial configuration provisions 32 CUs per chiplet. Each chiplet is projected to provide two teraflops of double-precision computation, for a total of 16 teraflops. The EHP also employs eight CPU chiplets (four cores each), for a total of 32 cores, with greater parallelism through optional simultaneous multi-threading.
Heterogeneous System Architecture (HSA) compatibility is one of the major design goals of the APU. HSA provides a system architecture where all computing elements (CPU, GPU, and possibly other accelerators) share a unified coherent virtual address space. These features are supported by AMD’s Radeon Open Compute platform (ROCm) to improve the programmability of such heterogeneous systems. We have created novel mechanisms like the QuickRelease synchronization mechanism [14] and heterogeneous race free memory models (HRF) [15]–[17] to reduce synchronization overhead between GPU threads, and heterogeneous system coherence (HSC) [18] to transparently manage coherence between CPU and GPU caches.
In the center of the EHP are two CPU clusters, each consisting of four multi-core CPU chiplets stacked on an active interposer base die. On either side of the CPU clusters are a total of four GPU clusters, each consisting of two GPU chiplets on a respective active interposer. Upon each GPU chiplet is a 3D stack of DRAM (e.g., some future generation of JEDEC high-bandwidth memory (HBM) [12]). The DRAM is directly stacked on the GPU chiplets to maximize bandwidth while minimizing memoryrelated data movement energy and total package footprint. CPU computations tend to be more latency sensitive, and so the central placement of the CPU cores reduces NUMA-like effects by keeping the CPU-to-DRAM distance relatively uniform. The interposers underneath the chiplets provide the interconnection network between the chiplets along with other common system functions. Interposers maintain high-bandwidth connectivity among themselves by utilizing wide, shortdistance, point-to-point paths.
The EHP uses eight GPU chiplets. Our initial configuration provisions 32 CUs per chiplet. Each chiplet is projected to provide two teraflops of double-precision computation, for a total of 16 teraflops. The EHP also employs eight CPU chiplets (four cores each), for a total of 32 cores, with greater parallelism through optional simultaneous multi-threading.
Heterogeneous System Architecture (HSA) compatibility is one of the major design goals of the APU. HSA provides a system architecture where all computing elements (CPU, GPU, and possibly other accelerators) share a unified coherent virtual address space. These features are supported by AMD’s Radeon Open Compute platform (ROCm) to improve the programmability of such heterogeneous systems. We have created novel mechanisms like the QuickRelease synchronization mechanism [14] and heterogeneous race free memory models (HRF) [15]–[17] to reduce synchronization overhead between GPU threads, and heterogeneous system coherence (HSC) [18] to transparently manage coherence between CPU and GPU caches.
The answers for this you will get in Q2, this year.
I'm not sure what marketing has to do with designing high performance GPUs...
I must be misreading this. You quoted Geekbench 3 single/multi scores of around 984/16,000 for Naples.
At https://browser.geekbench.com/geekbench3/multicore?page=5 it shows E5-2699 v4 results of 3588/84,909.
That seems far slower than the Xeon, and even if it's dual Xeon to single Naples the AMD system is less than half the speed.
I don't care about those Naples benchmarks. They were not made close to launch.
AFAICT, Asus never got Intel certification to support Thunderbolt bolt for those AM3+ mobos and it never shipped. All Asus ThunderboltEX add-in cards only support Intel mobos.
It's simple.
Aside from releasing Summit Ridge next week, AMD will release Naples in Q2.
Naples is similar to Summit Ridge and might be used in the next Mac Pro.
AMD is also releasing Raven Ridge in H2 and that might be used in the Mac Mini and the MacBook Pro.
We have first official leak from review, from China. Yes, those numbers are finally official:
This CPU is supposed to cost around 400$.
https://videocardz.com/66182/amd-radeon-7-1700x-pictured-and-tested
This CPU is supposed to cost around 400$.
https://videocardz.com/66182/amd-radeon-7-1700x-pictured-and-tested
At Nvidia, Marketers dictate to Engineers what they want to target. What to embrace, to spin in the marketing material. Thats how Nvidia works. Thats how their brand works.
It is more difficult if you have also been wounded by anticompetitive behavior.
Man, if only AMD marketing people figured out that they had to tell their engineers to design high performance GPUs, then that would solve all their problems. /sarcasm
AMD had a competitive CPU back in the Athlon 64 days. Intel had nothing to do with bulldozer being crap.
I know that, but AMD was also crippled by a lot of crap they did not create themselves.
I don't actually think your sarcasm is correct in the context of AMD, and how its run.
https://videocardz.com/66204/amd-ryzen-supports-ddr4-3600-mhz-memory
I wish the X300 MoBos will support at least 3200 MHz DDR4 memory.
I wish the X300 MoBos will support at least 3200 MHz DDR4 memory.
Aren't you suspicious that they're only focusing on a single application benchmark - and one that's wildly variable in the field?