Intel 10th Gen Core "Comet Lake-S" IGP-Disabled Processor Lineup Detailed

erek

[H]F Junkie
Joined
Dec 19, 2005
Messages
10,875
Who's all going with Intel this round?

"According to a company slide leaked to the web by InformaticaCero, there are at least three each of "F" and "KF" SKUs in the works. The lineup includes the 10-core/20-thread i9-10900KF and i9-10900F; the 8-core/16-thread i7-10700KF and i7-10700F; and the 6-core/12-thread i5-10600KF and i5-10600F. Clock speeds and cache sizes of these chips are identical to their corresponding non-F SKUs (eg: i7-10700KF clock-speeds being identical to those of the i7-10700K). Provided they're sold at slightly lower prices, the lack of an iGPU doesn't affect target buyers of these chips - PC gamers or creative professionals who use graphics cards and don't need an iGPU. Competing Ryzen processors lack iGPUs by design. Intel is expected to debut its 10th generation Core "Comet Lake-S" processors in April."

https://www.techpowerup.com/264043/...lake-s-igp-disabled-processor-lineup-detailed
 
Taking potshots at AMD for not including IGP, when AMD's Renoir parts will be available on desktop 65w chips shortly :D

There's a good reason AMD doubled the core count over Raven Ridge, well Intel prefers to trickle -out more cores like some fickle water spout.

Renoir 8-core should compete well with i7 10700
 
Last edited:
  • Like
Reactions: N4CR
like this
I just hope that the new socket 1200 doesn't fully isolate power between the CPU and GPU on the chip. Previous F parts still pulled all there power just from the CPU side of things. Power consumption was the same which meant that the CPU side was simply moving more amperage off of the existing lines instead of pulling current from those fromt the GPU side of things.
 
Shame there isn't a way to use the GPU core as a added processor if it's not being used to display graphics instead of being dead weight.
 
Last edited:
  • Like
Reactions: erek
like this
Intel has had 40 lanes since Coffee Lake (8th-gen Core).

Right, it's just an imaginary number. - they keep adding more lanes, while the back-end stays 4-lanes.

AMD was the only CPU maker to actually add 4 more lanes directly connect to the CPU (with Zen 2.) It just requires a new motherboard. They also added PCIe 4.0 support (something it's going to take Intel another year to get working correctly)
 
Right, it's just an imaginary number. - they keep adding more lanes, while the back-end stays 4-lanes.

AMD was the only CPU maker to actually add 4 more lanes directly connect to the CPU (with Zen 2.) It just requires a new motherboard. They also added PCIe 4.0 support (something it's going to take Intel another year to get working correctly)

Ah yeah you are probably right, I forgot about that bottleneck which is absolutely ridiculous.
 
Intel CPU's have been stuck at 16 PCIe lanes for all mainstream SKU's. The platform has had far more than that for some time. You get 40 PCIe lanes by having 24 lanes off the PCH. Yes, there is the DMI 3.0 bottleneck which people whine about, but its rarely an actual problem. Saturating this bus is something you typically only do in NVMe RAID testing.

AMD really isn't much better in all actuality. AMD gives you 24 PCIe lanes off the CPU, but these are split for specific functions the user has no real control over. 16 lanes are allocated for graphics. 4x lanes are reserved for NVMe storage, although the option is there to split these into SATA ports. Another 4x lanes are reserved for the PCH. Intel doesn't count the four that are used for the PCH off the CPU total. Basically, it's marketing bullshit.

I've tested the same devices on both AMD and Intel's platforms. On paper, AMD would seem to have an advantage with dedicated lanes for NVMe storage going right to the CPU. However, in the real world or even in the world of synthetic testing, there isn't any appreciable difference in performance between these two competing designs where storage is concerned.
 
Last edited:
Back
Top