Intel Announces 10th Gen Core S-Series CPUs, Led By The Core i9 10900K

[CochainComplex]

You don't need benchmarks, the architecture is unchanged. You can safely predict how fast every part is.

yep, massive binning and thinner dies making the thermal velocity boost of up to 5.3 GHz possible ...

[Jabberwocky]

Yes I have heard about the Z490 prices increase of around 60% over the Z390 https://www.techspot.com/news/84892-...decessors.html ...they never learn it.

p.s. https://www.techspot.com/images2/new...9-image-38.jpg



vs. a 3year old pc ....that is redicoulus marketing

Intel's grasping at straws https://youtu.be/DA6pq4vj4lI?t=1003

[CochainComplex]

Intel's grasping at straws https://youtu.be/DA6pq4vj4lI?t=1003

I was just watching this one - when I have clicked on your link I thought my browser has some bug because the same clip appeared over and over again.

But common Intels singlecore decade is imprinting the illusion of game engines cant do multthreading

Dispite of bad intel CPUs the software guys at intel in the opensource departments are doing a great job e.g.: Clear Linux and iGPU drivers.

[torsionbar28]

Do I really see some reasonable prices for intel CPU's?

Considering it's Yet Another Skylake Refresh (YASR) I should certainly hope so. They're just milking that elderly 14nm process at this point, pure profit for them. Ryzen is faster runs cooler and more advanced all around. The fact that intel is still selling 14 nm and PCIe3 in 2020 makes me wonder why anyone would bother with an intel chip, even at their discounted prices.

[Slartifartblast]

That's one expensive George Forman grill, ironically that's what one used to say about AMD chips, how times have changed.

[GruenSein]

Considering it's Yet Another Skylake Refresh (YASR) I should certainly hope so. They're just milking that elderly 14nm process at this point, pure profit for them. Ryzen is faster runs cooler and more advanced all around. The fact that intel is still selling 14 nm and PCIe3 in 2020 makes me wonder why anyone would bother with an intel chip, even at their discounted prices.

The only corner case I can imagine are very few single threaded workloads for non-pro desktop software (i.e. gaming). And that is mostly academic, too. But what chance do they have apart from pumping 250W into a 125W CPU, which already is a massive increase over the 84W standard we saw a few generations ago? The non-pro desktop is the last market in which power consumption does not matter that much because some kid will have a blast telling his buddies about his new AIO-Cooler and 800W PSU. Everywhere else, power consumption is either expensive (data center: electricity cost, cooling facilities, density issues) or impractical (mobile: lower clocks, more battery drain) or impairs performance (HEDT: fewer cores one can fit in one socket at decent clocks).

[Mario Junior]

"muuhhh gaming"
The actual state of Intel.

[jrdoane]

"muuhhh gaming"
The actual state of Intel.

That's because single core performance is really the only leg Intel can stand on right now. They certainly don't have command when it comes to multi-threaded performance. AMD has that down pat. When push comes to shove, even the single threaded performance isn't that much better. The problem with this strategy is that AMD already has a solution to the scaling problem, you just throw more chiplets at it and yields for relatively small chiplets are going to be pretty good compared to larger dies, so that's a huge advantage. All AMD has to do is improve single threaded performance to kick out the one leg Intel is standing on when it comes to the CPU market.

The bottom line is that Intel will never approach the power efficiency of Ryzen so long as they cling to a 14nm process (regardless of how many pluses you throw after it,) and they'll never get the yields they need to compete with their traditionally large monolithic designs. On that front, AMD appears to be years ahead of Intel. The best thing we see is a Pentium D-esque Xeon chip with 56 cores between two, more or less, identical dies, which shows how far Intel needs to go to catch up with AMD's solution for scaling.

If COVID-19 wasn't a problem, I probably would be building a new tower right now since my 3930k is aging, and I can say that this would not convince me out of buying a Ryzen 3900x.

[r1348]

14nm++++++++++++++++++++++++++++++++++++++++++++++ +++++++++++++++++++++++++

[Mario Junior]

That's because single core performance is really the only leg Intel can stand on right now. They certainly don't have command when it comes to multi-threaded performance. AMD has that down pat. When push comes to shove, even the single threaded performance isn't that much better. The problem with this strategy is that AMD already has a solution to the scaling problem, you just throw more chiplets at it and yields for relatively small chiplets are going to be pretty good compared to larger dies, so that's a huge advantage. All AMD has to do is improve single threaded performance to kick out the one leg Intel is standing on when it comes to the CPU market.

The bottom line is that Intel will never approach the power efficiency of Ryzen so long as they cling to a 14nm process (regardless of how many pluses you throw after it,) and they'll never get the yields they need to compete with their traditionally large monolithic designs. On that front, AMD appears to be years ahead of Intel. The best thing we see is a Pentium D-esque Xeon chip with 56 cores between two, more or less, identical dies, which shows how far Intel needs to go to catch up with AMD's solution for scaling.

If COVID-19 wasn't a problem, I probably would be building a new tower right now since my 3930k is aging, and I can say that this would not convince me out of buying a Ryzen 3900x.

Also, they are releasing Skylake once again. 😂