Therefore, Intel would directly cut off the circuit through software or hardware, preventing that part of the circuit from running.
Subsequently, this eight-core processor would become a seven-core processor, but Intel doesn't have a seven-core processor model.
So, they can only cut off the second physical core again through software or hardware, turning the seven-core processor into a six-core Core i5 processor to sell.
Taiwan Novel Network β π π π πΊ π . πΌ π π
This is also why CPU processors have high, medium, and low-end models.
The reason is that you are buying CPU processors that should have been scrapped for others.
Of course, compared to Intel, Nvidia's 'knife skills' are naturally more precise.
Because Intel only produces three configurations, while Nvidia produces even more models and configurations.
When Intel's eight-core processor is damaged, they will honestly and proactively disable a normal core again, turning a good seven-core CPU into a six-core CPU.
On the other hand, Nvidia does things completely differently. They will only add a 'Ti' to the end of the graphics card name and then sell the seven-core processor.
Therefore, players will be dumbfounded to find that Nvidia's graphics card models only have five models: GTX1030, GTX1050, GTX1060, GTX1070, and GTX1080.
But in reality, if you count them, including GTX1050Ti, GTX1070Ti, and GTX1080Ti, Nvidia has a total of eight graphics card models!
Therefore, people on the internet in the previous life exclaimed that Old Huang's 'knife skills' were good, truly not wasting anything.
Of course, there are also disadvantages.
Because the more graphics card models there are, the more times game developers need to optimize for that graphics card.
The workload of optimizing three graphics cards versus optimizing eight graphics cards is different, so many game developers will scold Nvidia.
However, as long as Nvidia's graphics cards sell well, it's useless no matter how much game developers scold them.
After all, their graphics cards have a high market share. Optimizing this one graphics card is equivalent to optimizing graphics cards with a large market share.
So, game developers have no choice but to honestly prioritize optimizing Nvidia's graphics cards amidst curses.
Who told Nvidia's graphics cards to have high sales and a high graphics card market share?
At this time, Geek Technology Company produced 5 configurations of graphics cards, obviously learning from the experience of Intel and Nvidia.
Therefore, the low yield rate of graphics cards doesn't matter much. Anyway, as long as it's not completely unable to run,
it can be turned into mid-range and low-end graphics cards and sold.
As for what are the disadvantages besides the super large chip area, which leads to super high chip costs?
The disadvantage is that the power consumption is relatively high. The power consumption of the GTX580 graphics card is only 250W, but Geek Technology Company's reaches an astonishing 350W.
Such high power consumption naturally leads to serious heat generation, which can be seen from the Phoenix graphics card having three oversized fans.
However, although these two disadvantages are relatively obvious, they are insignificant flaws in the eyes of players.
Because they value performance improvement more. High power consumption is just a matter of paying more electricity bills, and large cooling fans are just a matter of louder fan noise.
As for the problem that the chip area is relatively large, which means the graphics card price will definitely not be cheap.
If they can't afford the Phoenix 1080 graphics card, can't they afford the Phoenix 1030 and Phoenix 1050 graphics cards?
,
So, these costs and disadvantages are insignificant in the eyes of players, which Nvidia CEO Old Huang naturally knows, and his face immediately becomes quite gloomy.
Of course, at this time, Old Huang's focus is not only on hardware parameters but also on other aspects.
"It's one thing for the hardware parameters to be so high, but it even integrates an NPU neural network processor and frame generation technology!"
In the VIPεΈ, Nvidia CEO Old Huang said with a gloomy face.
First of all, the hardware parameters of this Phoenix 1080 graphics card have completely surpassed the GTX680 graphics card that Nvidia is developing and preparing to officially release in March next year.
After all, the GTX680 graphics card they are developing only has 1536 stream processors,
and only 128 texture units and 32 raster units.
Therefore, in terms of hardware core parameters, the GTX680 graphics card that is about to be developed is directly defeated by the Phoenix 1080 graphics card.
Moreover, what makes Nvidia CEO Old Huang even more uneasy is another thing, that thing is the NPU neural network processor and frame generation technology!
Integrating an AI chip into the graphics card is a crazy idea!
Old Huang had thought about doing this before, but he didn't really do it.
Because Nvidia is in an absolute dominant position in the market and doesn't need to do this.
So, the GTX580 chip simply didn't plan to put an AI chip into the graphics card.
Therefore, the GTX680 graphics card that Nvidia is developing naturally does not support frame generation technology!
"As you can see, this is our Phoenix 1080 graphics card.
It adopts a 28nm manufacturing process and has billion transistors.
Therefore, it has the world's largest graphics card chip area, with a chip area of ββan astonishing 594 square millimeters.
With so many transistors, what can it bring us?
It brings super computing power and super game graphics!"
When Gao Nian said this, the screen behind him suddenly changed, and floating-point arithmetic data appeared.
The peak computing power displayed on this floating-point arithmetic data is an astonishing 11TFL0PS!
"This is impossible!"
Nvidia CEO Old Huang couldn't help but stand up, looking at the projection screen in front of him with disbelief.
Because this floating-point computing performance is far ahead of Nvidia's GTX680 graphics card.
Outsiders don't know, but in fact, as the CEO of Nvidia, Old Huang knows how much floating-point computing power their GTX680 graphics card is developing.
The theoretical peak of the floating-point computing power of the GTX680 they are developing is only TFL0PS, while Geek Technology Company has directly achieved 11TFLOPS.
This is equivalent to the floating-point computing power of the Phoenix 1080 graphics card being times that of the GTX680 graphics card, which is really embarrassing for Nvidia!
"As you can see, our Phoenix 1080 graphics card has an astonishing 11TFL0PS of floating-point computing power.
This computing power is quite amazing, because the Tianhe-1 supercomputer we released in 2010 only has a peak computing power of 4700TFL0PS.
If we don't consider losses and other issues, theoretically we only need to use 427 Phoenix 1080 graphics cards,
then we can have the computing power of the Tianhe-1 supercomputer."
"Oh my god!"
"My god!"
"Awesome, it only takes 427 Phoenix 1080 graphics cards to have the computing power of the Tianhe-1 supercomputer!"
"I understand how powerful the Phoenix graphics card is!"
"It turns out that the Phoenix 1080 graphics card is also a small supercomputer."
"Awesome, Geek Technology Company is really awesome."
β
The people at the scene shouted excitedly, cheering for the power of Geek Technology Company.
After all, according to Gao Nian's theoretical conversion, this is equivalent to the Phoenix 1080 graphics card having 1/427th of the computing power of the Tianhe-1 supercomputer.
Thinking of this, how can consumers not be excited and excited?
