naruto-story Chapter 517 The new card actually didn't want to use 14nm.
Chapter 517 The new card actually didn't want to use 14nm.
Good evening, welcome to the new era of GeForce.
Old Huang's words awakened the audience, and the clients also told them that if they really had grievances, they would wait until after the press conference to express them...
However, he didn't rush to report the parameters. Instead, he walked to the screen in the center of the stage, tapped it lightly, and a line graph appeared on the screen.
That was a graph showing improvements in GPU performance.
The chart showed the actual performance improvement of graphics cards from the past five generations, which made many audience members grit their teeth in frustration.
"I've seen everyone's feedback, and they are indeed very dissatisfied with this improvement. So today, Lao Huang is going to... personally tear off the shackles binding all of us."
Many viewers were once again wondering, what is Lao Huang up to today?
"It's not that we're seeking change, but as everyone can see."
How much has GPU performance improved in the past 5 years?
After saying that, he picked up two more cards and compared the parameters on the paper. It looked like an improvement of more than 50%, but the actual net improvement rate was...
"That's right, everyone read it correctly, it's 15%."
What about power consumption? It only decreases by 5% per year.
After saying that, he shook his head, his tone tinged with regret, showing no sign that he had done these things.
"That's not right."
Players spend more money but get smaller and smaller improvements.
The industry is stuck in a rut, making incremental progress, but we don't want to wait any longer.
After saying this, he clenched his fist, and in that instant, the line graph on the screen shattered, replaced by a schematic diagram of a completely new architecture.
"Damn, Lao Huang is... so ruthless he even curses himself!"
"How dare he bring up these things... Didn't Amd learn them from him?"
"Upstairs, I won't allow you to insult Hwang In-hoon, at least not until the price is revealed. [doge emoji]"
"It seems that Lao Huang has realized his mistake [doge emoji]"
"Hey guys upstairs, stop messing around. Old Huang isn't going to change. His strange behavior right now, don't you think he's brewing something big?"
"Don't worry, I'll send my regards to his family when the time comes!"
After the new architecture was unveiled, Huang waited a while to allow his master to process this unusual press conference and prepare himself mentally.
As you all probably know from the leaks, our generation of products will see a comprehensive price increase.
But we don't make money through dishonest means.
Therefore, we are bringing you the brand-new Maxwell Pro architecture.
"Pah! You, Old Huang, dare to talk about not making money through dishonest means?"
"I can say that, but you, Old Huang, can't."
"What? A completely new architecture? I heard the price went up, so I just bought last year's card, and now you're telling me you're not releasing incremental upgrades?"
The live stream chat immediately started showering Huang with insults, as they had a feeling that the performance upgrade this time would be significant...
But the so-called price increase is just to clear the inventory faster.
So it wasn't just the chat in the live stream that started sending greetings; many viewers in the audience also joined in.
But Lao Huang was completely unaffected! After all, this is only the beginning. More and more people will be criticizing him later.
"This is a design that starts from scratch."
This includes entirely new stream processor units, a completely new cache architecture, and entirely new power control logic.
At this point, Lao Huang deliberately paused for a moment, allowing all the competitors who were switching screens to learn a thing or two.
"Most importantly, we have a unique advantage!"
This is the world's first consumer-grade GPU manufactured using a 14nm process.
This isn't like Intel's first-generation process; our performance is very stable.
After hearing that it was confirmed to be 14nm, some of the abusive consumers immediately shut up.
After all, the mainstream process in the industry is still 28nm.
Although Intel claims to have solved the 14nm process, it has been unable to mass-produce it, and it is just a PPT process.
Nvidia, on the other hand, has already started mass production, putting it far ahead of the industry.
"Actually, when we were deciding on the process, many people on the team asked me why we were going to use 14 nanometers."
After all, no one has solved the problem yet, so it's not impossible to try again next year.
By next year, the audience seemed to feel that Hwang In-hoon on stage had changed again, becoming the familiar Hwang they knew...
This statement is full of implications of swordsmanship.
"To be honest, I was tempted at first too."
There's nothing we can do about it; our competitors are too weak. We can still sell a lot even with only minor upgrades.
However, after I submitted this report to our General Manager Shen, he rejected it outright.
"What the heck? I actually saw regret in his eyes."
"You weren't the only one upstairs; I saw it too."
"It's clear that Huang really doesn't want to use 14nm."
The director's close-up shot conveyed that regretful expression throughout the entire live broadcast room.
"Hey, director, you can't do this to people like this, it's just a little bit..."
On stage, Hwang In-hoon made a joke, but Zhao Ming in the audience looked on with admiration. As expected, Hwang is a professional when it comes to interpersonal relationships.
Not only did he take all the blame, but he also made all the viewers grateful to Shen Fei, so this wave of graphics card price increases should be able to pass easily.
"Of course, the reason that I can accept it wholeheartedly is not a big deal, it's just that we have more 14nm process production lines."
If we want to use the 28nm process... there's no production capacity. Even if we use 22nm, we'll have to wait in line.
After saying that, Lao Huang put on an invincible expression again: "Wafer fabs are too awesome, there's nothing we can do about it."
"Damn! I used to think Lao Huang might start his own business, but now it seems he finally understands the allure of the entire industry chain."
"Who would want to go out when they can enjoy the shade under the big tree of Chuanyin?"
On stage, Lao Huang stopped joking and changed his attitude.
"That was just a small incident. Let me tell you what the 14nm process has brought to our generation of GPUs."
As soon as he finished speaking, a brand new line of PowerPoint slides appeared on the big screen:
Double the performance!
Power consumption halved!
"That's right, this is the most powerful force brought about by the improvement in manufacturing process."
A smaller process can fit more transistors into the same area.
After he finished speaking, a detailed architecture analysis immediately appeared on the screen.
The new Mac45 Pro architecture doubles the efficiency of its stream processors, delivering 90% better performance for the same number of stream processors compared to the Kepler architecture.
In addition, the brand-new display memory controller improves memory bandwidth utilization by 40%...
With multiple improvements and the increased density brought by the 14nm process, more computing units can be accommodated in the same chip area.
After seeing these improvements, the viewers in the live stream could no longer contain their amazement and exclaimed in amazement.
"Did they really squeeze out a whole tube of toothpaste?"
"Upstairs, don't be so sure. After all, it's a completely new process, so the yield rate is bound to be an issue!"
The technologies shown above can probably only be fully utilized on the top-of-the-line model!