• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

NVIDIA GeForce "Ampere" GPUs Built on Samsung 8nm Instead of TSMC 7nm?

Except they didn't, the 1800X was still slower than 6900K across the board, plus Intel had a 10 core CPU available at the time. The only reason they succeed was because of the performance/dollar, proving one again that halo products aren't what drives sales. Most people probably don't even know which are the fastest GPUs/CPUs out there, let alone lettings themselves being influenced somehow because of them. Again, that's how fanboys think.

Are you a troll ?

Ryzen 7 1800X scores 16,248 in PassMark, while Core i7-6900K only 14,478.


 
No, you are, I can't take serious anyone using only PassMark as a legit way to compare CPUs.

So, now I am quite sure you are a troll. How do you dare to compare different platform CPUs for different markets, HEDT vs MSDT, 140-watt vs 95-watt, 700$ vs 300$ ?

Why don't you compare with Core i7-6700K or Core i7-7700K, or with Ryzen Threadripper 1950X?
 
View attachment 161137


What exactly were all those people up until 2017 buying ? AMD stickers ?

Going from 1 AMD CPU sold for every 1 Intel CPU sold in 2006 to 1 AMD CPU sold to every 4 Intel CPUs sold. They might as well have been. AMD was scraping the bottom of the barrel for every sell the could manage so I doubt those were quality sales.
 
So, now I am quite sure you are a troll. How do you dare to compare different platform CPUs for different markets, HEDT vs MSDT, 140-watt vs 95-watt, 700$ vs 300$ ?

Why don't you compare with Core i7-6700K or Core i7-7700K, or with Ryzen Threadripper 1950X?
If smart, Probably using a variety of useful benchmarks and more than one reviewers benchmarks on a wide variety of useful and useable applications rather than a shockingly useless benchmark suite, and especially dodging the use of single data points and source's, trolls. Focus on one winning bench or aspect.
You choose.

No reviewer I saw said the 1800X beat intel In Game's.
 
If smart, Probably using a variety of useful benchmarks and more than one reviewers benchmarks on a wide variety of useful and useable applications rather than a shockingly useless benchmark suite, and especially dodging the use of single data points and source's, trolls. Focus on one winning bench or aspect.
You choose.

No reviewer I saw said the 1800X beat intel In Game's.


There is no such thing a useful benchmark. There are either ones which coincide with Intel's corrupted practices or ones which don't.
Games do not work well on Core i7-6900K since it's not with ring bus.

I don't know why you would repeat gaming when it's not useful unless in niche gaming market.
No one with the right mind buys LGA-2011v3 to play games on it.
 
There is no such thing a useful benchmark. There are either ones which coincide with Intel's corrupted practices or ones which don't.
Games do not work well on Core i7-6900K since it's not with ring bus.

I don't know why you would repeat gaming when it's not useful unless in niche gaming market.
No one with the right mind buys LGA-2011v3 to play games on it.
Re read, I said a variety of useful applications, not one.

And your on crack if you think no benchmark useful, as an engineer and ATM a test engineer, that's the biggest load of balls you ever stated, not that I am or would bother tracking your nonsense.

And game's run fine, you are just leaning on one data point, I'm not.
 
Re read, I said a variety of useful applications, not one.

And your on crack if you think no benchmark useful, as an engineer and ATM a test engineer, that's the biggest load of balls you ever stated, not that I am or would bother tracking your nonsense.

And game's run fine, you are just leaning on one data point, I'm not.


But you lie. You have to compare AM4 with LGA 1151, and TR4 with LGA 2011-v3.
Comparing AM4 with LGA 2011-v3 and claiming AM4 loses is like comparing VW Golf with Mercedes S Maybach and claiming the Mercedes is better which is nonsense because they have very different purposes to begin with.
 
But you lie. You have to compare AM4 with LGA 1151, and TR4 with LGA 2011-v3.
Comparing AM4 with LGA 2011-v3 and claiming AM4 loses is like comparing VW Golf with Mercedes S Maybach and claiming the Mercedes is better which is nonsense because they have very different purposes to begin with.
Talking to you is like chatting to a maybacks boot.

I lie, I said none of that.

you're comparing CPU in a thread about GPU with one shit benchmark, that's what I commented on.

Call me a lier , What.

How's about we skip this tat and get back on topic.

8nm Ampere or 7nm on Tsmc, given Nvidia tried to low ball Tsmc I think this could be valid, and could possibly have ramifications.
 
Talking to you is like chatting to a maybacks boot.

I lie, I said none of that.

you're comparing CPU in a thread about GPU with one shit benchmark, that's what I commented on.

Call me a lier , What.

How's about we skip this tat and get back on topic.

8nm Ampere or 7nm on Tsmc, given Nvidia tried to low ball Tsmc I think this could be valid, and could possibly have ramifications.


You didn't have to comment anything, whatsoever.

The point is and still stands - AMD needs a real product in order to compete, something that AMD doesn't have currently.
Their financial situation was so dire that 3 years ago they were almost bankrupt, their stock was $2.
Now, it's higher than $50.

Because of the halo created by the superior to Intel's microarchitectures Zen architecture.


Everything else argues by you and the other guy is pure trolling.
 
You didn't have to comment anything, whatsoever.

The point is and still stands - AMD needs a real product in order to compete, something that AMD doesn't have currently.
Their financial situation was so dire that 3 years ago they were almost bankrupt, their stock was $2.
Now, it's higher than $50.

Because of the halo created by the superior to Intel's microarchitectures Zen architecture.


Everything else argues by you and the other guy is pure trolling.
But, But what does that have to do with Nvidias Ampere being built on 8nm Samsung nodes, do tell.

Halo , ill go with that, still effin off topic nonsense.
 
But, But what does that have to do with Nvidias Ampere being built on 8nm Samsung nodes, do tell.

Ask @BoboOOZ

He started with the claim that companies don't need "best" products in order to make money. Which isn't and can never be true.

This will be my last post on this subject, because I think we discussed it enough.
But basically I think you idealize what is competition in a free market. Companies do not try to make the best product that they can make. Companies try to make the most money they can make. They do that usually by doing the cheapest product possible but which is still better than the competition, or at least competitive with the price of the competition.
Or, as they say, when you're chased by the bear, you don't have to be faster than the bear, you just have to be faster than your friend ;).
 
So all this trolling is my fault?
Deeply sorry, I take it all back.
Anyway, despite the off-topic nature, a few interesting points were made.
 
Seven times slower doesn't mean the feature is missing, every other higher end Kepler card could do FP64. Again, the only thing that supposedly made this a "prosumer" card is that it was faster and had more memory. Great, but it didn't have any "pro" feature. Only the Titan V can be argued to have been a proper workstation card, Tensor Cores, HBM. All the others ? Meh.

Yeah, 7 times faster is a pro feature, just like the quadros having that fast of performance was a pro feature. It makes the difference between the cards basically not being usable for the tasks because it's so slow to being entirely usable. The Titans had two customers, people that wanted to show off their epeen and people that wanted to use them for rendering and didn't want to fork out a bunch of money for a Quadro. Actual gamers weren't buying Titans because they weren't meant for gamers. Heck, that's why most websites that reviewed gaming graphics cards didn't even bother reviewing most of them.

Hardly anyone renders anything using FP64 outside some rare instances and the complexity of your simulation is independent from the precision used. "Incredibly important" is an overstatement, take it from someone that actually wrote some numerical simulations for GPUs.

A large enough number do it to make the Titans viable choices at that price. I mean, if they didn't sell at that price, if they weren't worth the money, it's a completely wasted product. Yet, they sold out.

Want me to ruin your idea that these Titans are not workstation cards even further ? The current RTX Titan just as every other Turing card has the same 1:32 FP64 core ratio. So explain to me again, is that a prosumer card or not since you think FP64 is incredibly important ?

Yep, and the reason for that is they ate far too much into the Quadro sales. The modern Titans are pretty pointless other than epeen, I'll totally admit that. But you were talking about the original Titan. Past that, AMD hasn't really competed with the high end. Like I said, you had to go back 7 years to find an example. So even nVidia's actual high end gaming cards(not the fringe halo products) have been allowed to run out of control price wise.

And it isn't like AMD isn't without their overpriced halo products. $1,500 Radeon Pro Duo ring any bells?
 
Well, it seems the trolling is still ongoing.

Nevermind, I just discovered the "ignore" feature, pure gold :cool:
 
Back
Top