|
Dramicus posted:It figures that the Intel thread would be in favor of reducing the number of threads and call it an improvement. Oy, page snipe. Kaby Lake-G chips were killed last night. That's such a peculiar set of products it's practically collectible: https://amp.tomshardware.com/news/intel-discontinue-kaby-lake-g-amd-graphics,40577.html NewFatMike fucked around with this message at 16:33 on Oct 8, 2019 |
# ? Oct 8, 2019 16:30 |
|
|
# ? May 30, 2024 14:57 |
|
Dramicus posted:It figures that the Intel thread would be in favor of reducing the number of threads and call it an improvement. oh man
|
# ? Oct 8, 2019 17:05 |
|
NewFatMike posted:
Friendship ended with AMD, now
|
# ? Oct 8, 2019 17:12 |
|
Saw an engineering sample of a 6950X on ebay - I was wondering, are ESs of unlocked chips also unlocked themselves? (just curious, 6950X isn't worth buying at this point even at ES pricing - "only" $500 is equal to the rumored pricing of the new comet lakes) Paul MaudDib fucked around with this message at 04:47 on Oct 9, 2019 |
# ? Oct 8, 2019 23:33 |
|
Yes, usually. There's not really a hard way to determine if one of them is locked or unlocked, but generally they are.
|
# ? Oct 9, 2019 01:03 |
|
NewFatMike posted:
Loading as a blank page for me for some reason, did they say if they're going to continue driver support?
|
# ? Oct 9, 2019 03:50 |
|
isndl posted:Loading as a blank page for me for some reason, did they say if they're going to continue driver support? It's the "amp" may it burn in hell: https://www.tomshardware.com/news/intel-discontinue-kaby-lake-g-amd-graphics,40577.html
|
# ? Oct 9, 2019 05:10 |
|
Volguus posted:It's the "amp" may it burn in hell: https://www.tomshardware.com/news/intel-discontinue-kaby-lake-g-amd-graphics,40577.html "Intel is refocusing its product portfolio. Our 10th Gen Intel Core processors with Iris Plus graphics are built on the new Gen11 graphics architecture that nearly doubled graphics performance. We have more in store from our graphics engine that will bring further enhancements to PCs in the future." That's one bold-rear end statement, but I'm just glad that we can finally put loving HD 620 in the goddamn ground.
|
# ? Oct 9, 2019 06:11 |
|
In other news Intel recently showed what’s essentially a NUC on a huge PCIe card, to be used for modular computers such as Razer‘s Project Christine. I suppose it could also be useful for upgrading AIO PCs though I suspect it’ll go the way of MXM modules... theoretically standardized but practically hardly ever worth upgrading. https://www.anandtech.com/show/14953/the-pc-on-a-gpu-intels-new-element-brings-project-christine-to-life Making the form factor worse than ATX for cooling seems to be questionable in this day and age but perhaps they designed it with watercooling in mind. Could be quite a neat idea with those leakproof quickconnectors. eames fucked around with this message at 09:25 on Oct 9, 2019 |
# ? Oct 9, 2019 09:17 |
|
Those designs look a lot closer to blade servers than consumer devices
|
# ? Oct 9, 2019 13:36 |
|
Dramicus posted:It figures that the Intel thread would be in favor of reducing the number of threads and call it an improvement.
|
# ? Oct 9, 2019 18:41 |
|
Apparently Intel is running into the same BIOS capacity issues as AMD earlier this year. The X299 updates for newer CPUS will drop support for discontinued Kaby Lake X SKUs.
|
# ? Oct 10, 2019 14:06 |
|
eames posted:Apparently Intel is running into the same BIOS capacity issues as AMD earlier this year. Heh. Well, at least nobody bought those.
|
# ? Oct 10, 2019 15:32 |
|
eames posted:Apparently Intel is running into the same BIOS capacity issues as AMD earlier this year. What ever happened to that Raid key on x299?
|
# ? Oct 10, 2019 15:54 |
|
eames posted:Apparently Intel is running into the same BIOS capacity issues as AMD earlier this year. I'm sure all ~200 people who bought KL-X chips will be furious.
|
# ? Oct 10, 2019 20:46 |
|
Too bad they discontinued the x299/itx I now want 18cores in mini itx
|
# ? Oct 11, 2019 01:40 |
|
Malcolm XML posted:Too bad they discontinued the x299/itx pretty sure AMD board partners will try that with threadripper.
|
# ? Oct 11, 2019 04:59 |
Malcolm XML posted:Too bad they discontinued the x299/itx
|
|
# ? Oct 11, 2019 11:31 |
|
wargames posted:pretty sure AMD board partners will try that with threadripper. Wouldn't it be neat to have a special board and case, where the CPU socket is on the back side of the board? Enormous amount of space around it for a giant, giant heatsink, and the components on the front of the board wouldn't need to be crammed in on little riser daughterboards and so on.. Maybe the space on the "front" could be used on my theoretical "mini ITX" sized board for a large chipset heatsink. I don't know, now I'm just talking nonsense.
|
# ? Oct 11, 2019 15:06 |
|
Looks like those earlier slide leaks might be right. Core i3 10100 spotted with 4/8 and 3.6 base: https://twitter.com/TUM_APISAK/status/1182640081623891969 If that means the rest of the slides are right, the stack is going to have HT across the board with 6/12 as i5s ($180 and up), 8/16s as i7s ($340 and up), and 10/20s as i9s ($409 and up). It also would mean a new socket.
|
# ? Oct 11, 2019 20:11 |
|
Cygni posted:Looks like those earlier slide leaks might be right. Core i3 10100 spotted with 4/8 and 3.6 base: if true that would be a welcome turn compared to the 9xxx series where only i9s got HT.
|
# ? Oct 11, 2019 20:40 |
|
Cygni posted:Looks like those earlier slide leaks might be right. Core i3 10100 spotted with 4/8 and 3.6 base: Really good lineup if the core/thread counts and pricing are true.
|
# ? Oct 11, 2019 20:43 |
|
movax posted:I think the demographics of this forum have mostly grown up past fanboy arguments of Intel vs AMD but there’ll always be a bit of crossover as was mentioned above. I sincerely think a combined x86 thread would be better and also sincerely want that thread to be immediately renamed to "x86: we literally cannot stop talking about Sandy Bridge" the instant someone starts yammering about how long they ran their 2500k or 2700k.
|
# ? Oct 11, 2019 22:07 |
|
I cleaned my desktop of dust the other day and noticed the CPU fan was running at full blast after. Checked CPU temps while under load. 100C. Searched, "Is 100C CPU bad?" And learned you need to reapply thermal paste after removing a cooler fan. That's my story for the day. Edit: Also Intel's stock coolers are a loving pain to reinstall. Those plastic pins are garbage.
|
# ? Oct 11, 2019 22:22 |
|
Methylethylaldehyde posted:Didn't we do the whole 'merge the threads' thing to the IT bitching threads like 2 years ago, and it was met with a resounding 'ehhhhhhh' and went back to 2 threads fairly quickly? Why is that, anyway? And aren't there three IT whining threads? More poo poo that pisses you off: My boss said I don't scream enough [SPAM] FW: RE: I entered your meeting details and ended up in a sex chat Working in IT 3.0: The Scrum Dumpster
|
# ? Oct 11, 2019 22:52 |
|
Cygni posted:If that means the rest of the slides are right, the stack is going to have HT across the board with 6/12 as i5s ($180 and up), 8/16s as i7s ($340 and up), and 10/20s as i9s ($409 and up). It also would mean a new socket. basically what I predicted a few months ago, 10C at $400-450, maybe a bit higher for super binned ones, rest of lineup settles into place beneath that. Should have guessed they'd still call a 10C an i9 since it's their flagship but I wasn't sure they'd want to be seen cutting the price of their "flagship" segment. They weren't gonna launch it at something stupid like $700 like people imagined though. it's really the only sensible way to do it, both from a perspective of where AMD has positioned their products and to unfuck the godawful mess that 9-series made of their lineup. 9-series took 3 segments worth of performance and awkwardly stretched it over 4 segments and ended up with a bunch of HT-disabled parts to avoid lower-tier parts stepping on higher parts, it was a complete mess. the one dumb thing is they still arbitrarily switched sockets yet again, I was still kinda hoping against hope they would come to their senses thermals are also going to be "challenging" for 10C chips, leaks didn't specify all-core boost but if they're trying to hold 5 GHz then yikes. I dunno if 5 GHz all-core will even be doable with a delid and LM at that point. Paul MaudDib fucked around with this message at 00:14 on Oct 12, 2019 |
# ? Oct 11, 2019 23:43 |
|
What's the thermal density (is that the right term? or maybe just watts/cm2?) on current chips compared to past generations?
|
# ? Oct 12, 2019 02:20 |
|
95 watts and it has been around there for like 10 years. There's also stuff like the 9980XE which gets up to 165 watts.
|
# ? Oct 12, 2019 02:26 |
|
Farmer Crack-rear end posted:What's the thermal density (is that the right term? or maybe just watts/cm2?) on current chips compared to past generations? Chips keep getting smaller/denser, and the desktop / server power envelopes haven't changed all that much, so higher. A 2500k was 95W at 216mm^2. A 8600k is still 95W, but at 149mm^2.
|
# ? Oct 12, 2019 02:43 |
|
Farmer Crack-rear end posted:What's the thermal density (is that the right term? or maybe just watts/cm2?) on current chips compared to past generations? Excluding the iGPU:
(the other half being the die itself is a bit thicker now to resist mechanical working and cracking, because smaller dies don't have the mechanical strength to resist the expansion/contraction of the heatspreader layer as well... this is Intel's concession to die longevity.) Paul MaudDib fucked around with this message at 03:50 on Oct 12, 2019 |
# ? Oct 12, 2019 03:37 |
|
Paul MaudDib posted:the one dumb thing is they still arbitrarily switched sockets yet again, I was still kinda hoping against hope they would come to their senses I have a Z370 board and obviously this bothers me too. Can’t help but wonder if they did this with pressure from the motherboard manufacturers. I would’ve considered the 10C as a drop in replacement but now I’m just going to ride it out to see what the future brings for both AMD and Intel. Strong single thread performance (when overclocked) and low memory latencies are two things that the *lake SKUs have going for them but there’s also the high power consumption and ongoing security issues.
|
# ? Oct 12, 2019 06:31 |
|
Paul MaudDib posted:Excluding the iGPU: Yeah, getting cooling close enough to the transistors to be meaningful is going to be a significant issue in upcoming process nodes. The stresses from thermal cycles between the die and substrate/heat spreader/etc is only going to get worse as transistors shrink. The breakthrough that will allow smaller nodes to clock as aggressively as Intel's 14+++ is probably going to be a packaging/substrate technology instead of having much to do with the process node itself.
|
# ? Oct 12, 2019 12:57 |
|
... meanwhile Intel is already stacking DRAM on top of the Die of their 10nm CPUs (Lakefield, Surface Neo), AMD suggests that this tech is going to scale to half a kilowatt or more (HPC) and nobody knows how they're going to cool any of that. maybe we'll see those fabled in-silicon microchannel heatpipe coolers in the next few years, with copper around the Die to act as a heatspreader.
|
# ? Oct 12, 2019 13:16 |
|
Perhaps switching the packaging, heat spreaders or even the semiconductor itself to diamond which has the highest thermal conductivity of any bulk material. Or some weird carbon nanotube heat spreader technology (although diamond is also mostly carbon and probably cheaper than trying to scale up manufacturing on some weird bulk CNT material). But I think in the near term it is mostly going to come down to some type of mechanical solution, existing coolers are more than sufficient to dissipate the total wattage if you can just get the heat in to the cold plate quickly enough.
|
# ? Oct 12, 2019 13:37 |
|
The real trick is going to be getting that mechanical solution to work in a Surface / sleek-looking professional laptop.
|
# ? Oct 12, 2019 13:48 |
|
a quick search for microchannel cooling shows that they can handle >1kW/cm2 but the customers of this tech aren't civilian so I suspect it's way too expensive for consumer products at this point. I guess tiny integrated peltier elements could also work. There were so many rumors about these technologies 5-10 years ago but nothing made it to market (yet).
|
# ? Oct 12, 2019 13:54 |
|
eames posted:... meanwhile Intel is already stacking DRAM on top of the Die of their 10nm CPUs (Lakefield, Surface Neo), AMD suggests that this tech is going to scale to half a kilowatt or more (HPC) and nobody knows how they're going to cool any of that. it is completely bonkers how everyone has gone all-in on 3D stacking without any idea how to cool it, everyone’s just sure it’s the next big thing. DRAM on die (or die on dram) makes a certain amount of sense but nobody has working solutions to cool big powerful dies in the middle of the stack
|
# ? Oct 12, 2019 15:59 |
|
Paul MaudDib posted:it is completely bonkers how everyone has gone all-in on 3D stacking without any idea how to cool it, everyone’s just sure it’s the next big thing. The solution could be as simple as "dont run this at 5ghz". IPC should go up since everything is on the same substrate, so the question is will it go up enough to make 2ghz chips match performance of their high clocked counterparts
|
# ? Oct 12, 2019 16:05 |
|
Bringing DRAM closer to the die is for bringing data closer to the CPU so it can wait less and work more, which means it will also consume more power. Stacking everything together will definitely make cooling and power consumption worse in pretty much every way, because it sticks more power consuming and thermal dissipating components in less space than ever before and helps them all work harder besides.
|
# ? Oct 12, 2019 16:17 |
|
|
# ? May 30, 2024 14:57 |
|
Indiana_Krom posted:Bringing DRAM closer to the die is for bringing data closer to the CPU so it can wait less and work more, which means it will also consume more power. Stacking everything together will definitely make cooling and power consumption worse in pretty much every way, because it sticks more power consuming and thermal dissipating components in less space than ever before and helps them all work harder besides. I hope you don't think that they plan to make these chips 95w TDP. 7nm doesn't scale well with high power usage. They will try to target the optimal spot of the efficiency-curve. More likely we'll get similar performance of today's chips but at 15w or 7w so it can be passively cooled and tossed in a tablet or surface pro
|
# ? Oct 12, 2019 16:35 |