Alongside its 9th Gen Core Mobile launch, Intel today is also significantly expanding its range of 9th Gen Core Desktop processors, adding 25 new SKUs. The additions include refinements to the existing i3, i5, i7, and i9 lineups, as well as the introduction of new 9th Gen SKUs to the lower-end Celeron and Pentium Gold classes.
As we saw with the complete 8th Gen Core lineup, these new parts cover nearly the complete range of price, performance, and feature categories, with varying support for hyperthreading, integrated graphics, cache levels, and memory frequency. However, as indicated late last month, all parts support Intel's Optane memory.
No word on exact retail availability of these parts but, like today's mobile launch, OEMs will be announcing systems shortly.
As usual it’s that “Up To 40
As usual it’s that “Up To 40 PCIe” lanes that can be questioned and what will actualy be available will not be that much or if actually 40 PCIe lanes will cost much more.
And that Check Mark for Integrated graphics when there should be the exact Graphics Version Number/Naming provided for each specific SKU.
Yes let’s only bring out that Top End Gen 11 Graphics for mostly benchmarking and marketing reasons while the actual graphics made use of will have way less EUs available on average.
I do really want to see L1, L2, L3 cache sizes listed in total and on a per core basis and most of these slides are marketing obsfucation by omission of some necessary details like the aforementioned cache metric breakdowns and actual Graphics used instead of some very nondescript Check Mark.
Intel will get design wins as usual what with Laptop OEM’s dependent on Intel’s engineering assistance and marketing assistance that adds up to a large incentivised reason for Laptop/Other OEM’s bottom lines.
But with so much more competition from AMD across a few generations of Zen/Zen+ and now Zen2 based offerings that are arriving shortly Intel is forced to offer more cores once again. And AMD’s Zen2 SKUs on the AM4 platform going as high as 16 cores 32 threads will really force Intel to counter with even more than just updates that offer so little for the extra cost.
i would just wait. wikichip
i would just wait. wikichip already has cometlake listed as 8 core 16 threads for i7 and 8 cores 8 threads for i5.
so they will just relabel the current i7 as i5 and the current i9 as i7. there is a bios update for a new stepping of these chips. i would be willing to bet that a big part of that new stepping is so they have a line of dies with no igpu (not just turned off like some current chips but actually not on there in the first place).
this will allow intel to cut some fat. think about it. a 9900k with no igpu would be smaller than an 8700k. an 8700k would be smaller than a 7700k. and the new 10 core cometlake will be smaller than the current 8 core i9.
they may also have changed the way turbo boost works to be more flexible like amds turbo. apparently the new chips can boost above their rated turbos if the cooling is there and the power limit is there.
since you can change the power limit on most non oem z series boards does that mean you can buy a locked part, cool it with a huge cooler like a noctua dh15 and have it near the performance of an overclocked version? maybe not that far but would be nice if it would come close.
and get the prices down on these. the chip shortage is really caused by going to 28 and 56 cores for xeon and the demand for new 6 and 8 core parts on the desktop. nobody would upgrade for years because it was all quads. but now they are buying like hotcats. the xeon production is nearly caught up. with the new 8 core chips being smaller than the old 6 core chips (the ones with no igpu that is) the price for the 8 core 16 thread parts could be nearing 330-360 within a couple months of launch).
oh and the linux patches that said “cometlake comes off of coffeelake” meant that these chips will work in current 300 series boards. which is great. wikichip also has them listed as working with current boards
Intel is really segementing
Intel is really segementing more and more above their usual too many segements all while Intel is in no hurry to end any of their the parts shortage issues on consumer SKUs. Intel would rather have their margins remaing higher even if that means loosing some market share to AMD. Intel is a very high overhead business operation that can not afford its gross margins slipping too far below the mid 50% range because Intel has massive fixed costs to cover for operations like chip fabs and fat middle management layers and their pay/perks packages that drink down the Revenues like sailors and beer at fleet week.
Look that those 6 core and 6 thread i5 offerings and Hyperthreading(SMT) disabled and the same on more of the i7 and i9 segements that also are more HT(SMT) disabled. And the Intel MB paltforms are the same way bing segemented by PCIe lanes and some reqiuiring RAID Keys to be purchased at extra costs.
Intel’s HT(SMT) disabled offeings continue to grow relative to their HT(SMT) enabled offerings all throughout Intel’s Consumer market segements with only one T sries part getting a full 8 core 16 thread option and costing much higher also relative to the non HT(SMT) enabled SKUs.
Most certianly Intel’s marketing will be pushing that 8 core parity with AMD’s Ryzen 7, 8 core/16 thread, options but most of Intel’s 8 core options will be single threaded with HT(SMT) disabled and that’s probably the result of more die harvesting on Intel’s part to make use of more cosnumer dies with defective HT(SMT) hardware. This is more of a rebranding by Intel with some minor tweaks and the costs are still not competative with AMD and Intel is not worried at all about that becase of Intel’s addiction to higher margins.
Intel’s binning operations for consumer parts are as complex as they have ever been what with Intel trying to get as many usable SKUs and price/feature segements created in order to field sufficient numbers of total consumet offerings during this shortage period that appears to be going to last longer than was stated eariler.
AMD’s got its 3rd generation Ryzen/Zen2 desktop parts coming on 7nm and offering possibliy 16 cores/32 threads max for mainstream market usage and 2 older generations of 8 core/16 thread Ryzen desktop offerings that will be continuing to come down lower in ASP. And all of the Ryzen generations of 8 core parts have that SMT enabled and are unlocked consumer offerings with AMD offering more value and features/dollar as well as an overall price/performance metric that Intel can still not match.
AMD’s Ryzen Raven Ridge APUs are in only their second generation and Zen+/12nm based but even with the first generation Raven Ridge APUs the Graphics offered was better overall on average with most of the AMD parts coming with the higher end Vega nCU options and much lower costs than Intel charges for its better graphics options.
AMD is also increasing its average memory clock offerings and the Ryzen 7 2800H APU(35 – 54 W) supports DDR4-3200 dual-channel memory.
Wikichip is not listing the max memory clocks suppoted in the Raven Ridge tables and Wikipedia is not currently listing the max memory clocks support on the Raven Ridge 2nd generation/3000 series parts currently but if that Ryzen 7 2800H APU parts is any indicator then AMD will be ahead there also. Anandtech list the H series APU parts at DDR4-3200 Dual Channel.
That 2800H if some OEM laptop OEM grows a pair will definitely make for some interesting high performance workhorse laptop SKU compared to Intel’s current offerings and the 2800H sporting Vega 11 graphics. But the OEM laptop market is funny in that it’s become so dependent on one CPU supplier’s incentives over so many years and that holds true currently as well.
Is Intel’s hyperthreading
Is Intel’s hyperthreading still security threat?