Gigabyte: Z490 Motherboards Will Support Intel 11th Gen Rocket Lake CPUs

Gigabyte Z490 Aorus Xtreme (Image credit: Gigabyte)

Gigabyte has confirmed, via its Aorus livestream, that the recently launched Z490 motherboards come with support for Intel's future 11th-gen Rocket Lake-S processors.

Comet Lake-S just landed yesterday, so it might seem weird that there are already murmurs of its successor. Comet Lake-S marks the turning point for Intel since it's the first time that the Intel is introducing a 10-core processor on a mainstream platform, but that news isn't very exciting for most enthusiasts in light of AMD's impressive Ryzen 9 3950X that debuted with 16 cores and cutting-edge I/O, too. Intel is a step behind AMD in terms of adopting the PCIe 4.0 interface, a feature which the Red Team already supports on the Ryzen 3000-series (codename Matisse) chips. Ultimately, this is the reason why there's already plenty of interest in Rocket Lake-S.

Rocket Lake-S, which is rumored to arrive later this year to replace Comet Lake-S, will usher in a plethora of interesting improvements. Gigabyte's latest statement confirms that the Rocket Lake-S will indeed drop into the LGA1200 socket. Among the list of Rocket Lake-S improvements is alleged support for PCIe 4.0. Word on the street is that Intel was going to adopt PCIe 4.0 with Comet Lake-S, but things went sideways with the 400-series chipset, and the chipmaker eventually scrapped the feature.

That makes it necessary to clarify why many Z490 motherboards on the market are PCIe 4.0 ready, but there won't be full support for the feature. The 400-series chipset doesn't natively support PCIe 4.0, so motherboard vendors resorted to incorporating PCIe 4.0 timers, drivers, and re-drivers to enable the feature for future processors. At the end of the day, Rocket Lake-S will likely supply the PCIe 4.0 lanes. You can expect full PCIe 4.0 support on the direct PCIe lanes, but only some motherboards will enjoy the functionality on the first M.2 slot.

Intel has a policy of not commenting on future products, so we don't expect the chipmaker to make any statements about Rocket Lake-S or its compatibility with Z490 motherboards anytime soon. However, limited support on Z490 motherboards would give Intel the chance to launch new motherboards to bring complete support for PCIe 4.0, say Intel 500-series.

15 Comments Comment from the forums

InvalidError 01 May 2020 16:17

400-series Intel boards supporting 11th-gen CPUs shouldn't be surprising anyone when Intel has been doing two officially supported generations per platform for all but a few single-gen exceptions for well over 20 years.
Reply
Deicidium369 01 May 2020 18:21

InvalidError said:
400-series Intel boards supporting 11th-gen CPUs shouldn't be surprising anyone when Intel has been doing two officially supported generations per platform for all but a few single-gen exceptions for well over 20 years.
Exactly. 370 and 390 most recently. Z490 was for Comet Lake (before the boards were released that supported PCIe4) with the Z590 being for Rocket Lake S and officially support PCIe4. 470 likely to be released later in the less expensive tier.

Intel is 90%+ of the market - so when Intel releases PCIe4 - then it will be officially released in the main stream of desktops and servers and will no longer just be a niche implementation. Rocket Lake S will likely drop at the same time as Nvidia Ampere - which will most likely require PCIe4 - at least on it's high end - still backwards compatible with PCIe3. - but at reduced speed.
Reply
cryoburner 01 May 2020 21:12

Deicidium369 said:
Rocket Lake S will likely drop at the same time as Nvidia Ampere - which will most likely require PCIe4 - at least on it's high end - still backwards compatible with PCIe3. - but at reduced speed.
PCIe 3.0 x16 still provides significantly more bandwidth than today's high-end graphics cards require, so it will likely be years before PCIe 4.0 provides any tangible performance benefits for graphics cards, outside of some synthetic benchmarks. Today's highest-end cards are barely exceeding the bandwidth provided by PCIe 2.0, and that interface went into use over a decade ago. There's little doubt that the next graphics cards from AMD and Nvidia will "support" the PCIe 4.0 interface (AMD's cards from last year already do), but it isn't likely to be all that useful, let alone "required".

As for Rocket Lake, I'm doubting that well see it this year. It will probably be closer to a year from now before Intel's next generation of desktop processors come out. It wouldn't make much sense for them to release a new series of CPUs only half a year after Comet Lake comes out.
Reply
InvalidError 01 May 2020 22:14

cryoburner said:
PCIe 3.0 x16 still provides significantly more bandwidth than today's high-end graphics cards require, so it will likely be years before PCIe 4.0 provides any tangible performance benefits for graphics cards, outside of some synthetic benchmarks.
Don't forget the low-end: having 4.0x8 vs 3.0x8 makes a huge difference for the 4GB RX5500 with frame rates as much as doubling when the GPU runs out of local VRAM, pretty sure there would be some more room for additional scaling had it had a 4.0x16 interface instead of only x8 and with B550 motherboards just around the corner, there will be no need to break the bank to get 4.0x16 either. Ironic how 4.0x16 may deliver some of its greatest mainstream benefits to the entry-level.
Reply
cryoburner 01 May 2020 23:29

InvalidError said:
Don't forget the low-end: having 4.0x8 vs 3.0x8 makes a huge difference for the 4GB RX5500 with frame rates as much as doubling when the GPU runs out of local VRAM, pretty sure there would be some more room for additional scaling had it had a 4.0x16 interface instead of only x8 and with B550 motherboards just around the corner, there will be no need to break the bank to get 4.0x16 either. Ironic how 4.0x16 may deliver some of its greatest mainstream benefits to the entry-level.
I kind of think AMD might have done that just to make it look like PCIe 4.0 had some practical use for today's graphics cards, to help sell x570 boards, even if the people buying those boards are more likely to go with a higher-end card with an x16 interface and more VRAM anyway. Or maybe they just didn't care to make their sub-$200 parts all that competitive, since they have limited 7nm production capacity, and it probably makes more sense to put that toward higher-margin parts. They might have also wanted to give their partners a chance to offload existing RX 570/580/590 inventory.

Had they given the RX 5500 cards a PCIe 3.0 x16 interface, this would have not been an issue. As it is, it's kind of hard to recommend the RX 5500, when the 1650 SUPER, 1660 and AMD's older cards exist. And anyone upgrading an older PCIe 2.0 system probably shouldn't even consider one of these cards.
Reply
spongiemaster 01 May 2020 23:56

cryoburner said:
It wouldn't make much sense for them to release a new series of CPUs only half a year after Comet Lake comes out.

That's the glass is half full perspective. The glass is half empty perspective says, whatever architecture Rocket Lake is using is getting released over 5 years after the last architecture update, Skylake. Whenever it is ready, we need to ship it.

Remember Broadwell? Most people probably don't. If you blinked, you missed it. The only 2 desktop variants Intel released were on June 2nd, 2015. Skylake was released only 2 1/2 months later on August 15th, 2015. So there is precedent for Intel releasing CPU's and then going, "Just kidding. Here are the real CPU's we're releasing." If Rocket Lake is ready, I can't see Intel holding it back until next year, especially if the Ryzen 4000 series put a beating on Comet Lake.
Reply
InvalidError 02 May 2020 01:32

cryoburner said:
As it is, it's kind of hard to recommend the RX 5500, when the 1650 SUPER, 1660 and AMD's older cards exist.
The 4GB RX5500 performs on par and sometimes even better than the 1650S when it isn't bottlenecked by 3.0x8. There are many games where 3.0x8 costs it 10-20% in performance even when VRAM isn't an issue. AMD kind of screwed up and created an SKU without a market by launching it without budget-oriented boards that support 4.0 to make it work as intended.
Reply
larkspur 02 May 2020 15:18

cryoburner said:
And anyone upgrading an older PCIe 2.0 system probably shouldn't even consider one of these cards.
I can confirm that. As soon as it was revealed it was a x8 card I completely lost interest. Ended up with a RX 580 8gb on a wicked deal. The funny thing is, even if I had a PCIe 3 system I still wouldn't have bought it :)
Reply
cryoburner 03 May 2020 04:56

spongiemaster said:
Remember Broadwell? Most people probably don't. If you blinked, you missed it. The only 2 desktop variants Intel released were on June 2nd, 2015. Skylake was released only 2 1/2 months later on August 15th, 2015. So there is precedent for Intel releasing CPU's and then going, "Just kidding. Here are the real CPU's we're releasing." If Rocket Lake is ready, I can't see Intel holding it back until next year, especially if the Ryzen 4000 series put a beating on Comet Lake.
I actually have a laptop with a Broadwell CPU, though they offered a full lineup in the mobile space, and those processors were "current-gen" for longer.

However, as far as the desktop CPUs are concerned, there were a lot of differences compared to what we are seeing with the launch of Comet Lake. As you said, Broadwell only offered two mainstream models on the desktop, one i7 and one i5, whereas Comet Lake is offering a full lineup containing 32 different processor variations, including i9s, i7s, i5s, i3s, Pentiums and Celerons. That doesn't exactly sound like something they plan to replace in a matter of months.

Also, Broadwell desktop CPUs were designed to plug into existing motherboards, so they were still arguably a viable option even after Skylake launched. They brought much improved integrated graphics to the platform, along with a bit better efficiency on a new process node, but not much else. Comet Lake, on the other hand, needs new motherboards, and is increasing thread counts at any given price point. Even if efficiency hasn't really improved, adding SMT puts their core and thread counts largely on par with AMD's current offerings, and their higher clock rates should allow most Comet Lake processors to match or take the lead from AMD in terms of multithreaded performance. That seems like a product that should be able to stand on its own, and Intel has generally taken their time in responding to new AMD product launches.

An Intel roadmap leaked last year indicated that Comet Lake would be launching in Q2 of 2020, and Rocket Lake in Q2 2021, and given the accurate timing of Comet Lake's launch, I haven't seen anything suggesting that roadmap has been changed substantially. People expecting Rocket Lake this year are probably going to be disappointed. And with what little we know about Rocket Lake so far, its very possible that it might not even be much of an improvement over Comet Lake.
Reply
InvalidError 03 May 2020 07:28

cryoburner said:
Also, Broadwell desktop CPUs were designed to plug into existing motherboards, so they were still arguably a viable option even after Skylake launched.
No, they weren't viable options: the IPC increase was not enough to offset the clock handicap so Broadwell ended up performing better in some cases, just as much worse in as many others and about the same overall. Broadwell was primarily intended for all-day laptops and embedded applications where low power and a massively improved IGP were key selling points. AFAIK, Intel initially had no plan to make socketed variants, that only changed due to a combination of OEM demand and people starting to panic about Intel possibly going BGA-only (soldered to motherboard) for future chips when no desktop chips were forthcoming.

As for the hype about Rocket Lake possibly launching this year, on one hand you have the 3DMark results from an unknown Intel engineering sample and on the other hand, you also have Zen 3. Intel appears to have working silicon and a fairly strong incentive to at least announce something if it can get it up to speed in time to spoil AMD's launch. Also, if Intel is on track to start making chips on 7nm by Q4 next year, launching Rocket Lake next year would put it within about the same number of months from 7nm parts anyway. Intel needs to catch up on mainstream fab process and re-establish its lead on architecture, it will need to step up its launch schedule of delayed and back-ported designs either way.
Reply

Show more comments