|
Combat Pretzel posted:So question, I've read that this liquid metal paste stuff has supposedly higher thermal conductivity than solder (80W/mK for liquid metal vs. 50W/mK for solder, whatever moon units those are), why the hell isn't Intel using it? Because that poo poo eats through aluminum over time, and it requires more precision than a two-second 'gloop' by an Indonesian factory worker on a line.
|
#
?
May 30, 2017 22:04
|
|
|
|
| # ? Apr 28, 2024 17:24 |
|
|
Combat Pretzel posted:80W/mK for liquid metal vs. 50W/mK for solder, whatever moon units those are wikipedia posted:In SI units, thermal conductivity is measured in watts per meter-kelvin (W/(m·K)). gently caress am I ever glad I got into network engineering and not physics
|
|
#
?
May 30, 2017 22:11
|
|
|
Kazinsal posted:gently caress am I ever glad I got into network engineering and not physics The 1st year physics course I took in university seemed actively hostile towards students
|
|
#
?
May 30, 2017 22:16
|
|
|
mewse posted:The 1st year physics course I took in university seemed actively hostile towards students A lot of colleges treat courses like this as "weeder" classes. Even when it's not official, a lot of teachers are assholes and resent those lazy kids in those other majors and make it as bland and unengaging as possible. In my university it was the Digital Logic course, which was required for virtually any engineering field.
|
|
#
?
May 30, 2017 22:23
|
|
|
Paul MaudDib posted:A lot of colleges treat courses like this as "weeder" classes. Even when it's not official, a lot of teachers are assholes and resent those lazy kids in those other majors and make it as bland and unengaging as possible. It wasn't just bland and unengaging, we had to do lab reports and they nitpicked them for asinine reasons and every small criticism they had marked it down a mark out of 10, so a full 10% drop on the assignment. From talking to other people, they were told to mark them like that, probably as a weeder course like you said. My entire university experience seemed passively hostile towards students but that's another story
|
|
#
?
May 30, 2017 22:28
|
|
|
But there surely are non-corrosive thermal compounds that come close to the performance of solder?
|
|
#
?
May 30, 2017 22:41
|
|
|
There's the conductive part that's a worry, too. I mean, I'm sure they'd figure it out if they cared enough. But you stack enough issues on top of each other and figuring it out becomes non-trivial, and in the meantime they don't really need to because they're still within the thermal envelope they're targeting and if it also just so happens to curtail overclocking a little and therefore allow better market segmentation that might convince people to move up the product ladder, well that's just all the more reason to keep slathering the chips with lovely TIM.
|
|
#
?
May 30, 2017 22:54
|
|
|
BIG HEADLINE posted:Because that poo poo eats through aluminum over time, and it requires more precision than a two-second 'gloop' by an Indonesian factory worker on a line. 'scuse me, a Malaysian, Chinese, or Vietnamese factory worker 
|
|
#
?
May 30, 2017 22:57
|
|
|
BIG HEADLINE posted:Because that poo poo eats through aluminum over time, and it requires more precision than a two-second 'gloop' by an Indonesian factory worker on a line. There is no aluminum under the IHS. However that stuff is highly conductive and can flow out of place.
|
|
#
?
May 31, 2017 00:23
|
|
Don Lapre posted:There is no aluminum under the IHS. However that stuff is highly conductive and can flow out of place. Yeah, this, the IHS is copper on the inside. Also the solder they use on higher end chips is not cheap, both in materials cost and most importantly the complex way it has to be applied because there are like 14 layers of material to form a proper bond between the chip and the IHS, it's an expensive process. Anyway, smaller chips have problems with solder cracking/voids forming in the solder over time due to heating/cooling cycles and these issues will only affect more chips as thermal density gets higher and chips get smaller. Back with Ivy Bridge they had to switch to TIM, it's unsurprising that with how much smaller everything has gotten since then that they are starting to need to use TIM on the higher core count chips.
|
|
|
#
?
May 31, 2017 00:45
|
|
|
I'm certainly not going to delid anything. I'd like to keep using an air cooler (still not sure whether I want to risk annoying pump noises), and probably a fairly big one, so getting it to apply an even load on the die once you tilt the mainboard vertical will probably be tricky. That said, if say the 8C is rated to turbo to 4.3GHz, it'll hopefully do so in an acceptable temperature range.
|
|
#
?
May 31, 2017 00:59
|
|
|
AVeryLargeRadish posted:Yeah, this, the IHS is copper on the inside. Yeah, except if you're paying $799+ for a CPU, there should drat well be a budget for that expensive soldering process. I understand leaving it out on the G-series and even the i3/i5/i7 lines, but these i9s are effectively Xeons.
|
|
#
?
May 31, 2017 02:16
|
|
|
Are xeons soldied?
|
|
#
?
May 31, 2017 02:18
|
|
|
Scarecow posted:Are xeons soldied? I'm pretty sure 2011 and 2011v3 are soldered, so the i9s *not* being soldered is disturbing.
|
|
#
?
May 31, 2017 02:27
|
|
BIG HEADLINE posted:Yeah, except if you're paying $799+ for a CPU, there should drat well be a budget for that expensive soldering process. I understand leaving it out on the G-series and even the i3/i5/i7 lines, but these i9s are effectively Xeons. Huh, I thought the i9s were getting soldered.
|
|
|
#
?
May 31, 2017 02:38
|
|
|
AVeryLargeRadish posted:Huh, I thought the i9s were getting soldered. Someone delidded one already and it sure didn't look soldered. https://www.youtube.com/watch?v=cCv7hF7kepU @1:32 he confirms (turn on captions) at least HIS Skylake-X wasn't soldered, just pasted. In the same video you see just how much wasted space there is on the Kaby-X package. It's complete horseshit that they didn't put eDRAM on it - it could've fit 256-512MB of it. BIG HEADLINE fucked around with this message at 05:40 on May 31, 2017 |
|
#
?
May 31, 2017 05:34
|
|
|
BIG HEADLINE posted:Someone delidded one already and it sure didn't look soldered. 
|
|
#
?
May 31, 2017 10:39
|
|
|
Yeah, that's why I'm curious whether it's officially official, or if this is just with engineering samples, which I presume his chip is. I'd figure the engineering samples would use the final manufacturing process, so it'd be kinda official, but I don't really know how Intel operates, so...
|
|
#
?
May 31, 2017 11:26
|
|
|
I've got a 4770k atm and it really pissed me off how much of a difference deliding that temp wise it got me (15c) that i just know that if I got a 7900x i'd be reaching for the delidding gear all over again just because of how poo poo their paste is, but I dont nor should i be doing that on a 999USD cpu That alone has me wanting to go to threadripper
|
|
#
?
May 31, 2017 13:20
|
|
|
First reports claim the TIM is still terrible. https://www.kitguru.net/components/cpu/dominic-moass/intels-new-skylake-xkaby-lake-x-chips-still-use-low-quality-tim
|
|
#
?
May 31, 2017 13:57
|
|
|
Combat Pretzel posted:I'm certainly not going to delid anything. I'd like to keep using an air cooler (still not sure whether I want to risk annoying pump noises), and probably a fairly big one, so getting it to apply an even load on the die once you tilt the mainboard vertical will probably be tricky. That said, if say the 8C is rated to turbo to 4.3GHz, it'll hopefully do so in an acceptable temperature range. Delid, use liquid metal on the core, set the CPU in the socket, put the heat spreader back on the core and use the socket retaining mechanism to hold the heat spreader on the CPU, then put your heatsink back on as normal. I did that on my 6700k and got my temps down by 15-20c. No extra overclocking though.
|
|
#
?
May 31, 2017 14:04
|
|
|
I really want that Core i7-7820X. Release date is June, how soon will we get reviews on these processors?
|
|
#
?
May 31, 2017 19:31
|
|
|
Anywhere from a week before release date to the morning of the same day, if I had to guess at when the official NDA will lift and you'll see reviews from sites that actually get official samples.
Eletriarnation fucked around with this message at 21:09 on May 31, 2017 |
|
#
?
May 31, 2017 21:05
|
|
|
Rabid Snake posted:I really want that Core i7-7820X. Release date is June, how soon will we get reviews on these processors?
|
|
#
?
May 31, 2017 21:50
|
|
|
Combat Pretzel posted:There's a slight temptation to go with the 7900X for the solder and PCIe lanes, but I don't think I'll be able to justify that price gap to myself. What solder on the 7900X? https://www.youtube.com/watch?v=I1Bv8Mxnnlc
|
|
#
?
Jun 1, 2017 01:16
|
|
|
OK, then it's the 7820X.
|
|
#
?
Jun 1, 2017 01:19
|
|
|
Combat Pretzel posted:OK, then it's the 7820X. That wont be solidered ether
|
|
#
?
Jun 1, 2017 01:21
|
|
|
.
sincx fucked around with this message at 05:55 on Mar 23, 2021 |
|
#
?
Jun 1, 2017 01:22
|
|
|
Delid this
|
|
#
?
Jun 1, 2017 01:30
|
|
|
If I were to move to X299 with an i9, this would probably be the board I would go with.
|
|
#
?
Jun 1, 2017 04:59
|
|
|
SlayVus posted:If I were to move to X299 with an i9, this would probably be the board I would go with. It reminds me in a strange way of those tacky fiber optic "plants"/table pieces that people had in the late 80s/early 90s.
|
|
#
?
Jun 1, 2017 05:44
|
|
|
I haven't bought a mobo in a couple years. What's the deal with all the shrouding on new boards?
|
|
#
?
Jun 1, 2017 16:51
|
|
|
WAR DOGS OF SOCHI posted:I haven't bought a mobo in a couple years. What's the deal with all the shrouding on new boards?
|
|
#
?
Jun 1, 2017 17:05
|
|
|
Hey if you're willing to forgo a warranty, silicon lottery dot com is where you can buy delidded, liquid metaled, binned CPUs for a small price bump. Delidding is pretty super easy now anyway, they have these kits you can buy that basically crack your CPU open like some skull-screw torture device from one of the SAW movies. You don't even have to LM it, but if you stuck some Kryonaut TIM on there you'd probaby get a decent temp drop over the lovely paste stamp thing they use.
|
|
#
?
Jun 1, 2017 17:13
|
|
|
WAR DOGS OF SOCHI posted:I haven't bought a mobo in a couple years. What's the deal with all the shrouding on new boards? There's really no reason for it other than it makes it look a little "cleaner". None of the enterprise stuff (supermicro, asrock rack etc) have that. I don't mind it but I wouldn't go out of my way to get it either.
|
|
#
?
Jun 1, 2017 17:41
|
|
|
What are the implications of Intel integrating Thunderbolt 3 in their CPUs for desktop users? Aside from more ports on motherboards/on cases due to no royalties.
|
|
#
?
Jun 1, 2017 22:37
|
|
|
ufarn posted:What are the implications of Intel integrating Thunderbolt 3 in their CPUs for desktop users? Aside from more ports on motherboards/on cases due to no royalties. It's actually a bigger deal than you're presenting it as, Intel is actually opening the standard up, not just integrating it in their CPUs. So you could see stuff like AMD CPUs with built-in support too. Now that there's an actual target audience besides Apple users, presumably there will be more/better offerings of things like Thunderbolt monitors, drives, and external GPU enclosures. Up until now it's been a chicken-and-egg problem, consumers don't care because there's no devices, and device makers don't care because nobody besides Apple actually has Thunderbolt ports.
|
|
#
?
Jun 1, 2017 23:24
|
|
|
Right, so now it can be as ubiquitous as USB instead of being as ubiquitous as FireWire
|
|
#
?
Jun 1, 2017 23:29
|
|
|
Paul MaudDib posted:
Nonsense, there was that one Sony laptop that shipped for like a year with it too! And some custom Dells. 
|
|
#
?
Jun 1, 2017 23:31
|
|
|
|
| # ? Apr 28, 2024 17:24 |
|
|
What advantages would TB have over USB3.1? Unless there's way less CPU overhead, I don't see any quick adoption.
|
|
#
?
Jun 1, 2017 23:47
|
|