AMD CES 2019 Keynote Live Coverage

jimmysmitty

Champion
Moderator
They need to correct Moores Law. It wasn't that the process nodes were getting smaller but that the number of transistors per square inch were doubling. Yes this does tend to correlate with process node shrinks but not always. It has slowed down though due to process node walls being hit and will probably not speed up but slow down more unless they can find an efficient way to stack transistors for CPUs or move beyond transistors and thus Moores law.
 

jimmysmitty

Champion
Moderator


Anandtech showed off a CPU shot of an 8 core with the IO Chip and some numbers showing it just under a 9900K in performance but not much was said about if the 8 core is the top end mainstream or if its just the start.
 

mtracy1991

Upstanding
Nov 6, 2018
353
10
315
14


I read some article and was stated that we wont see any 3000 series processors until middle of 2019
 

salgado18

Distinguished
Feb 12, 2007
521
0
18,990
3
Happy about Ryzen, cautious about Vega.

They finally matched Intel's IPC, using less power, and on a modular design that is easy to escalate. Can Intel have an answer this year? Or will they need a new architecture to get back on top?

Happy Vega is fast, not so happy it is expensive. Cut half the memory and drop the price, we need strong competition against those RTXes.
 

InvalidError

Titan
Moderator

You can't increase transistor count without shrinks since the larger die you would otherwise end up with end up prohibitively expensive due to defect density and the number of die you can get per wafer. No process shrinks mean slow or no transistor budget increases because it isn't economically viable - can't double transistor budgets if the chip would end up costing more than twice as much due to being able to make less than half as many of those double-size chips per wafer.

Every major increase in transistor count for a given product type and price point is directly linked to process shrinks.
 
Aug 7, 2018
4
0
10
0
Well, 3rd gen ryzen still slower than Intel, 1700x was beating Removed out of 6900k in AMD CB benchmarks and blenders, with both know how the reality turned out.

2600x actually BEATS 8700k in CB multithreaded, so do not spread fake news how AMD scales badly there.
 
Aug 7, 2018
4
0
10
0
And that's why they argued with the power consumption, because they are still slower than Intel. Even after 3 years.
 

mtracy1991

Upstanding
Nov 6, 2018
353
10
315
14
MERGED QUESTION
Question from purpleshadowba : "About ZEN 2 and Ryzen 3000x"



Hoping we get pricing and models sooner than later.
 

Specter0420

Distinguished
Apr 8, 2010
72
0
18,630
0
I am only interested in tuned performance (overclocking). Intel has been releasing their chips with 1Ghz+ room for overclocking for over a decade now.
AMD has been releasing chips that are already trying as hard as they can, with room for 200-300Mhz overclocking at best.
If this trend continued then Ryzen 3 is trying as hard as it can and scores numbers just under Intel at stock (from Anandtech), that isn't very promising. Hopefully I am wrong.
 

jimmysmitty

Champion
Moderator


Intel does have a new uArch coming out, Sunny Cove. We will see how the changes they detailed help performance.



I understand that but still the idea is not about process size its about transistor count. And I did state that it will continue to slow down as die shrinks do unless they find a viable way to stack transistors and to deal with the heat issues they tend to present.
 

The 9900K turbos to 5Ghz and a good OC is maybe 5.2Ghz. Plus you need immense cooling to achieve this. So that's 200mhz.
The 8700k turbos to 4.7Ghz and a good OC is around 5Ghz, so that's 300mhz.

If you look at the base clocks then it seems like Intel CPUs OC massively, but the "K" series parts never run anywhere near as low as their base clocks on decent motherboards anyway. You're right to say that Intel CPUs have more OCing headroom (The X series Ryzen CPUs ship with almost no headroom whatsoever), it's nowhere near 1Ghz.
 

shrapnel_indie

Distinguished



Ryzen just beat the i9-9900K by a small margin with the exact same thread and core count... IF the same clocks, IPC is fractionally faster with AMD on that chip... It definitely did it running a cooler temp, using less power.

It's noteworthy that AMD hasn't fully tweaked the design yet, meaning that it will likely extract more performance from the chip before it comes to market. The eight-core 16-thread Ryzen processor was about even with the eight-core 16-thread Core i9-9900K in a Cinebench multi-threaded workload, with the Ryzen score 2,057 to the i9's 2040.

Given that both processors have eight cores and sixteen threads, the benchmark results suggest that AMD is matching Intel's single-threaded performance, at least in this particular benchmark. Cinebench does respond well to AMD's SMT implementation, but the results are impressive nonetheless.

Possibly matching Intel's single-threaded performance is a watershed moment for AMD, as that type of workload has long been one of the few areas where the Ryzen processors lagged behind Intel's models. But that isn't all. Su also pointed out that the denser 7nm node allowed the Ryzen processor to consume less power, which ultimately equates to heat, than the Core i9-9900K.
 

AMD have had the Vega 20 GPU in the works for some time, they've publicly said it would never be released as a gaming part. Either that was AMD misdirection, or something changed.

Vega 20 (now announced as Vega VII) was always going to be roughly competitive with the 1080ti, but with a >300mm2 die on a brand new 7nm process, 16GB of HBM2 and a 300W power draw, it was always going to be really, really expensive to produce. It always made sense for AMD in professional markets with (relatively) low volume but high margins.

I'm guessing that AMD never intended to release Vega 20 to the gaming market because they assumed they'd never be able to price such an expensive card to effectively compete with Nvidia through 2019. AMD were probably thinking: who in their right mind would pay $700 for 1080ti performance in 2019?! Remember in August 2018, it wasn't difficult at all to find a 1080ti for under $600.
Aaaaand then Nvidia launches the 2080 card which offers worse price to performance (on street prices anyway) and shows their intention to charge $700 for 1080ti performance through the bulk 2019.
AMD think: well we have a card sitting here that offers 1080ti performance and we can still make a profit @ $700... let's release it! The lack of partner cards is a decent indication that this is a recent decision.

I would love AMD to release these at a steep price cut, but I genuinely don't think they could make a profit any other way.

On a separate note: RE memory, does anyone know if the 4096bit memory bus is actually possible with just 8GB of HMB2? Perhaps their decision to go with a high (possibly excessive!?) 16GB is because it's required for the bandwidth. They could do 8GB, but that would mean dropping to the 2048bit interface which it seems was a very significant bottleneck for Vega 10 (56 & 64).
 

drivinfast247

Notable
May 29, 2018
1,279
12
1,065
85
I am only interested in tuned performance (overclocking). Intel has been releasing their chips with 1Ghz+ room for overclocking for over a decade now.
AMD has been releasing chips that are already trying as hard as they can, with room for 200-300Mhz overclocking at best.
If this trend continued then Ryzen 3 is trying as hard as it can and scores numbers just under Intel at stock (from Anandtech), that isn't very promising. Hopefully I am wrong.
Please show proof of a 1ghz Intel OC. And not on LN.
 

InvalidError

Titan
Moderator

The gap between single-core boost and sustainable OC is nowhere near 1GHz, more like 200-300MHz without throwing ridiculous cooling and VRM at it.
 

joeblowsmynose

Distinguished
Jun 14, 2011
91
2
18,635
0


The AMD chip was faster than the 9900k and used 30% less power. Lisa held up a delidded 8 core Ryzen three and interestingly the 8 core "chiplet" is offset to one side and there is just enough room exactly to add another 8 core chiplet inside the package. They wouldn't have offset the chiplet if they didn't have plans to fill the void that this left, with either more cores or iGPU.
 

Specter0420

Distinguished
Apr 8, 2010
72
0
18,630
0
My i7-920 (1st Gen) ran at 2.6 Ghz ALL CORES constant MAX, stock. I overclocked it to 3.8Ghz, ALL CORES, ALL THE TIME. That is a 1.2Ghz overclock. You deceitful little turds and your boost clock BS. That max is for one core, sometimes, for a limited amount of time. I kept this system for 10 years, it powered a 1060 6GB and ran fine in all VR games and most VR flight sims at reduced settings into 2018. It took about 8 years before OCed AMD could beat it.
My 8086K runs at 4.0Ghz, ACM, stock. I overclocked it to 5.2Ghz (keep her at 5.0 though), that is a 1.2 Ghz overclock.
This is all on air cooling, any liquid mixed with computer dust will become conductive and I keep my systems for a long time.
Amd's Ryzen 7 2700x runs at 3.7Ghz ACM (so 3.5-3.6Ghz compared to Intel's IPC, lol) and OC's ACAT on air to what? 4.0-4.1?
Intel = 1Ghz+ OC
AMD = less than 500mhz OC
Apples to apples, what I said holds true. Intel is jogging at stock, AMD is already gasping for air at stock.
 

ASK THE COMMUNITY