|
do they make a motherboard that is completely enclosed
|
#
?
May 3, 2020 03:33
|
|
|
|
| # ? Apr 24, 2024 04:05 |
|
|
taqueso posted:do they make a motherboard that is completely enclosed
|
|
#
?
May 3, 2020 03:39
|
|
|
taqueso posted:do they make a motherboard that is completely enclosed Yup!  
|
|
#
?
May 3, 2020 03:53
|
|
|
taqueso posted:do they make a motherboard that is completely enclosed NZXT makes a Z390: https://www.nzxt.com/categories/motherboards I'd assume they'll make a Z490 ~eventually~. They seem to be extremely lax on BIOS updates - the last one for the Z390 was a year ago. BIG HEADLINE fucked around with this message at 04:05 on May 3, 2020 |
|
#
?
May 3, 2020 04:03
|
|
|
I want a game console PC with no dumb PC standards ~800W TDP built in power supply 10L kthx.
|
|
#
?
May 3, 2020 21:15
|
|
|
eames posted:I assume the Comet Lake NDA was lifted earlier today. What comes after Comet Lake for the desktop? Tiger Lake?
|
|
#
?
May 4, 2020 01:26
|
|
|
Scientist Al Gore posted:What comes after Comet Lake for the desktop? Tiger Lake? Rocket Lake - it's still going to be on 14nm (however many pluses) and is expected to still be compatible with the same LGA1200 socket that Comet Lake will use. I think the other big thing about it is that Rocket Lake is supposed to be using the new Xe iGPUs (which Tiger Lake will also use, but Tiger Lake is going to be mobile-only) After Rocket Lake is supposed to be Alder Lake, which is going to be 10nm on desktop
|
|
#
?
May 4, 2020 01:38
|
|
|
gradenko_2000 posted:Rocket Lake - it's still going to be on 14nm (however many pluses) and is expected to still be compatible with the same LGA1200 socket that Comet Lake will use. Is Rocket Lake still essentially Sky Lake? If so, drat. If the rumors are true hopefully Alder Lake comes out later this year and I'll have yet another excuse to build a new pc 
|
|
#
?
May 4, 2020 01:43
|
|
|
These lakes all look the same
|
|
#
?
May 4, 2020 01:50
|
|
|
Scientist Al Gore posted:Is Rocket Lake still essentially Sky Lake? If so, drat. Yeah, Rocket Lake is still 14nm, except they're backporting all the I/O and iGPU stuff from Tiger Lake (which is 10nm-mobile)
|
|
#
?
May 4, 2020 01:50
|
|
|
Scientist Al Gore posted:Is Rocket Lake still essentially Sky Lake? If so, drat. Rocket Lake is reportedly a backport of Tiger Lake so there should be some IPC gains this time around. It’s 14nm but a new uarch and PCIe 4. Now, it's not all sunshine and roses, this is a core that was originally designed for 10++, with much higher density, so it will probably experience at least some clock regressions as things get farther apart and the timings loosen. I doubt it'll hit 5.0 let alone 5.3, they'll regress probably to 4.5-4.7 GHz territory. But still, 4.7 is a lot better than 5.3 in terms of thermals. This will move Intel back towards IPC competitiveness (probably about on par with Zen3 in productivity workloads) and away from just maxing the clocks. Paul MaudDib fucked around with this message at 01:58 on May 4, 2020 |
|
#
?
May 4, 2020 01:52
|
|
|
What are the enhancements with Tiger Lake that look interesting? I'm seeing DDR5, PCI Express 4.0 and support for newer versions of Display Port, HDMI, USB, etc. The only one that's useful appears to be PCI Express 4.0.
|
|
#
?
May 4, 2020 02:28
|
|
|
Scientist Al Gore posted:What are the enhancements with Tiger Lake that look interesting? I'm seeing DDR5, PCI Express 4.0 and support for newer versions of Display Port, HDMI, USB, etc. higher IPC in general, Ice Lake alone is already +18% IPC vs Skylake and Tiger Lake is a generation beyond that.
|
|
#
?
May 4, 2020 02:33
|
|
|
Yeah, Rocket Lake is a whole new core architecture (intel's first new one on desktop since August 2015) on an old process. It is kind of a "tock" by intel's old naming. Still 14nm, still on Socket 1200. Care of Videocardz a while back:  It is after Rocket Lake (and after Zen 3 for AMD) that the whole market just turns to big question marks all at once. New sockets, new architectures, new nodes, new memory, new interconnects, new packaging, etc, for everyone all within a 1-2 year span. Gonna be wild, and the roller coaster is gonna start with mobile in a few months here.
|
|
#
?
May 4, 2020 04:34
|
|
|
What do they mean by CPU attached storage?
|
|
#
?
May 4, 2020 04:41
|
|
|
persistent memory?
|
|
#
?
May 4, 2020 04:42
|
|
|
E: I missed that it was in the new column, this is about something totally different. I don't know what that is aside from support for Optane on a consumer platform. Instead of the lanes to that storage device coming from the motherboard chipset, which is a switch and introduces latency and bandwidth restrictions, the lanes come directly from the CPU. Think of it like using a USB hub. If you have a bunch of thumb drives attached to it and read and write from all of them at once, you're going to get bandwidth limitations compared to plugging them all directly into their own ports on the PC itself.
|
|
#
?
May 4, 2020 04:45
|
|
|
Interesting, will that be supported by Non-Optane SSDs or just Intel?
|
|
#
?
May 4, 2020 04:51
|
|
|
Scientist Al Gore posted:Interesting, will that be supported by Non-Optane SSDs or just Intel? It is just an M.2 attached directly to PCIe lanes from the CPU. It can offer better performance in specific scenarios (that dont really matter to home users), and it will let Intel support one PCIe 4.0 SSD on their systems (which doesnt really matter to home users). AMD already does it on their side.
|
|
#
?
May 4, 2020 04:59
|
|
|
Will we also see a whole new slew of SSDs in the Fall? Assuming that when Rocket Lake comes out.
|
|
#
?
May 4, 2020 05:01
|
|
|
Scientist Al Gore posted:Will we also see a whole new slew of SSDs in the Fall? Assuming that when Rocket Lake comes out. Yeah, there's a newer Phison controller (there is one that is their previous gen with pcie 4.0 bolted on already) along with new drives from Intel, Samsung (980 Pro), Kioxia (Toshiba) and Western Digital all showing up in the near future with pcie 4 connections.
|
|
#
?
May 4, 2020 05:04
|
|
|
Scientist Al Gore posted:Will we also see a whole new slew of SSDs in the Fall? Assuming that when Rocket Lake comes out. I don't think youll see any Rocket Lake in Fall. Rumor is that Intel is accelerating Rocket Lake as much as they can, but even then, Winter at the earliest seems likely to me. Comet Lake isn't even out the door yet for desktop. Fall 2021, sure. 
|
|
#
?
May 4, 2020 05:20
|
|
|
Cygni posted:Yeah, Rocket Lake is a whole new core architecture (intel's first new one on desktop since August 2015) on an old process. It is kind of a "tock" by intel's old naming. Still 14nm, still on Socket 1200. Wait why are they removing the ~secure enclave~ in some of the models?
|
|
#
?
May 4, 2020 13:20
|
|
|
Hed posted:Wait why are they removing the ~secure enclave~ in some of the models? Each of those either leak information (including RSA keys) or let the SGX be controlled by an attacker. A few of them are remote attacks that just need privilege on the host. BlankSystemDaemon fucked around with this message at 15:30 on May 4, 2020 |
|
#
?
May 4, 2020 15:28
|
|
|
Thanks, I should have been more specific. My question is why do they say they’ve removed it from Rocket Lake-S? That implies to me they are going to do something different for the mobile chipset? Really I want to know if it’s going to be completely phased out as everything cycles. It certainly seems so.
|
|
#
?
May 4, 2020 16:18
|
|
|
https://www.anandtech.com/show/15763/first-atx12vo-consumer-motherboard-the-asrock-z490-phantom-gaming-4sr I'm surprised anyone is launching ATX12VO designs in the DIY space. I don't think its going to be a very big market initially. 
|
|
#
?
May 5, 2020 19:17
|
|
|
Cygni posted:https://www.anandtech.com/show/15763/first-atx12vo-consumer-motherboard-the-asrock-z490-phantom-gaming-4sr Why? Because the power supplies don't exist? Getting rid of the 24pin connector is a good thing. With parts obsoleteness rate in the PC industry I don't think anyone would mind not being able to use their old PSU.
|
|
#
?
May 5, 2020 19:41
|
|
|
Shaocaholica posted:Why? Because the power supplies don't exist? Getting rid of the 24pin connector is a good thing. In principle you should be able to make an ATX12VO aftermarket/replacement cable for anything that has modular cables. You don't get the benefits of the smaller/less complex PSU, but it doesn't hold anything up either. (or just a straight-up ATX-to-ATX12VO adapter cable)
|
|
#
?
May 5, 2020 19:43
|
|
|
I will love it if that 24 pin connector goes the way of the dodo! Also the psu makers need a standard for their modular cables, because that stuff is all over the place often with disastrous results.
|
|
#
?
May 5, 2020 19:46
|
|
|
Any standard or trend to minimize DIY PC cabling is good no matter how minor.
|
|
#
?
May 5, 2020 20:02
|
|
|
priznat posted:I will love it if that 24 pin connector goes the way of the dodo! Yep, I got a refurbished EVGA PSU, and they gave me a SATA power cable meant for another one of their PSU lines. It plugged in fine, but when I used it, the computer wouldn't even boot (probably a ground fault). Luckily it didn't fry anything, but it's still incredibly stupid that manufacturers can't standardize it within their own lines because they're just buying stuff from third parties and slapping their names on them.
|
|
#
?
May 5, 2020 20:03
|
|
|
Inept posted:Yep, I got a refurbished EVGA PSU, and they gave me a SATA power cable meant for another one of their PSU lines. It plugged in fine, but when I used it, the computer wouldn't even boot (probably a ground fault). Luckily it didn't fry anything, but it's still incredibly stupid that manufacturers can't standardize it within their own lines because they're just buying stuff from third parties and slapping their names on them. Yeah the pins are actually swapped in a dangerous way when going between corsair and evga power supply cables, which led to every one of the cables in the lab getting labelled for the appropriate psu Also the fact you can plug in a pcie 6 pin cable into a cpu power 8 pin socket is also dumb as hell. Fix the keying because if it is doable somebody will do it!!
|
|
#
?
May 5, 2020 20:15
|
|
|
priznat posted:Also the fact you can plug in a pcie 6 pin cable into a cpu power 8 pin socket is also dumb as hell. Fix the keying because if it is doable somebody Fixed.
|
|
#
?
May 5, 2020 20:26
|
|
|
Oh absolutely, it’s been done in my very lab with exciting results.
|
|
#
?
May 5, 2020 20:28
|
|
|
Here's a super simple convention. You know how the PCIe connectors have square ones and the one with 2 beveled corners used for keying. Just make one of those 12V and the other GND. No way to fuckup the pinouts between different makers, etc. Doesn't matter how many pins you use.
|
|
#
?
May 5, 2020 20:33
|
|
|
Ain't none of you ever seen what happens when you plug a pci card in backwards
|
|
#
?
May 5, 2020 20:38
|
|
|
Shaocaholica posted:Why? Because the power supplies don't exist? Getting rid of the 24pin connector is a good thing. Because the supplies don't exist and the standard was explicitly made for OEMs that have to meet bulk efficiency requirements in California, not DIY. It also raises the price/complexity of motherboards, which generally have a shorter expected use period than power supplies. I'm personally all for the standard in a world where all the minor rails do is drive SATA and some old integrated ICs that should be replaced anyway (im looking at you ITE), but again, I'm just surprised anyone is bothering to make DIY boards for it already. But when you look at the board and specs, its pretty clear who the intended market is.
|
|
#
?
May 5, 2020 21:12
|
|
|
The power supply should just plug directly into a distribution board/wire harness built in to the case behind the motherboard tray, and the motherboard/drives/etc should all plug in to that, the only cables it should need are for fan hubs, pumps, GPUs and stuff like that. Like the power connector on the motherboard would be on the back side of the board and would just slot right in to a matching plug on the case.
|
|
#
?
May 5, 2020 22:13
|
|
|
If we going in the all +12V direction, why not just have PCI-E 8-pin replace the bulky 24-pin on the mobo side while standardizing all modular cables on the PSU end FFS.
|
|
#
?
May 6, 2020 04:05
|
|
|
|
| # ? Apr 24, 2024 04:05 |
|
|
why not full 48v like servers. highest efficiency, smallest cables
|
|
#
?
May 6, 2020 04:16
|
|
