|
That kind of thing makes a small difference with Intel CPUs due to how the CPU loading mechanism is designed, but it's completely unnecessary with AMD CPUs.
|
#
?
Nov 12, 2023 20:31
|
|
|
|
| # ? May 29, 2024 16:48 |
|
|
justdan posted:Hey, quick question: Intel's LGA1700 has an issue with the stock retention mechanism that made these "contact frames" come into the market, but they won't do anything for AM5. The only good part of that thing is that it might keep some of the thermal paste from going into those annoying crevaces, but thats absolutely not worth the cost unless you are swapping CPUs every day (and then you are probably using a graphite pad). Don't buy. ah poo poo, efb on the next page, shoulda figured!!
|
|
#
?
Nov 12, 2023 20:50
|
|
|
also noctua has a thermal paste bundle that includes a plastic bracket to deal with paste spill
|
|
#
?
Nov 12, 2023 20:53
|
|
|
Noctua has a similar thing but the only claim they make about it is that it stops thermal paste from dripping down into the underside of the IHS Because I guess some people think that actually is a problem?
|
|
#
?
Nov 12, 2023 20:59
|
|
|
Thanks for all the replies! Very helpful.
|
|
#
?
Nov 12, 2023 21:11
|
|
|
Klyith posted:Because I guess some people think that actually is a problem? The custom retention brackets for Intel have gotten a fair amount of media attention, with review videos on youtube getting millions of views across the lot of them. On the other hand, custom AMD retention brackets are pretty useless, and it's much harder to make a story out of "this company released a $15 product that doesn't help," so most people don't seem to be aware of that part. They only hear about how much it helps with Intel and assume it must with AMD too, and thus a market of suckers is born. The AM5 socket retains CPUs in a similar manner to LGA1700, but perhaps because the PCB and IHS are more square, or maybe because of how drat thick the IHS is, there's not much flex at all, if any.
|
|
#
?
Nov 12, 2023 21:32
|
|
|
Klyith posted:Noctua has a similar thing but the only claim they make about it is that it stops thermal paste from dripping down into the underside of the IHS If one is using a thermal paste that is conductive, then I could see wanting a bit of protection to keep it from getting under the IHS.
|
|
#
?
Nov 12, 2023 22:02
|
|
|
Subjunctive posted:I dunno OP, sounds like bullshit to me. I just used whatever came with my cooler IIRC. It seems like it would make thermal paste application way cleaner, and cleanup easier, since it won't gloop down under the ihs. Looks kind of neat, but it's far from necessary on am5 as far as I'm aware. Edit: I guess it may help with cooling, as it's contacting the ihs in a significant way and has some surface area to it HalloKitty fucked around with this message at 22:17 on Nov 12, 2023 |
|
#
?
Nov 12, 2023 22:12
|
|
|
I've been pretty gratuitous with paste application the last CPU changes and repastes, and never has something dripped anywhere. At best, there's some excess paste caked between the edge of the IHS and the retention bracket, which you only notice when removing the cooler and is cleaned up easily.
|
|
#
?
Nov 12, 2023 22:18
|
|
|
Normal thermal paste isn't conductive, so unless you are both quite excessive in applying it and incredibly anal about invisible messes, I can't imagine caring.
|
|
#
?
Nov 12, 2023 22:19
|
|
|
Dr. Video Games 0031 posted:The custom retention brackets for Intel have gotten a fair amount of media attention, with review videos on youtube getting millions of views across the lot of them. On the other hand, custom AMD retention brackets are pretty useless, and it's much harder to make a story out of "this company released a $15 product that doesn't help," so most people don't seem to be aware of that part. They only hear about how much it helps with Intel and assume it must with AMD too, and thus a market of suckers is born. I was talking more about the thermal paste getting into all the nooks and crannies of the AM5 IHS. People care about that for some reason. I don't get it, despite myself having been known to carefully clean off a CPU with alcohol and a q-tip. Koskun posted:If one is using a thermal paste that is conductive, then I could see wanting a bit of protection to keep it from getting under the IHS. True, but you need something with way *more* protection than these gimmicks HalloKitty posted:Edit: I guess it may help with cooling, as it's contacting the ihs in a significant way and has some surface area to it I doubt the fitting is tight enough to really do any metal-to-metal conduction. But it might help people torque down the cooler harder, which does help.
|
|
#
?
Nov 12, 2023 22:28
|
|
|
Are those Kryosheet thermal contact pads that I occasionally see advertised on GN/HU any good if you're pretty much just gaming and not doing anything CPU-intensive? Sometimes feels like my biggest impediment to casually tinkering around with builds is the prospect of having thermal paste get everywhere (and it does, it always does).
|
|
#
?
Nov 12, 2023 23:29
|
|
|
Cross-Section posted:Are those Kryosheet thermal contact pads that I occasionally see advertised on GN/HU any good if you're pretty much just gaming and not doing anything CPU-intensive? Yeah, they're fine. The performance is slightly worse than fresh thermal paste, and the up-front cost is higher. But they last forever. So if you are a dedicated non-tinkerer who wants to put the PC together and then ignore it for 5 years (other than the occasional dust cleaning) they're pretty ideal. IIRC GN's biggest caveat when they reviewed it was that the sheets are pretty fragile, which can make their supposed re-usability tricky. Cross-Section posted:Sometimes feels like my biggest impediment to casually tinkering around with builds is the prospect of having thermal paste get everywhere (and it does, it always does). Stop worrying about precise application. Draw a big X from corner to corner on the IHS, cap the tube, and then put the heatsink on. Any video you have seen on youtube about thermal paste application, showing demonstrations with glass sheets -- "don't do it this way, look at the trapped air bubbles!" -- are completely bogus misinformation.
|
|
#
?
Nov 12, 2023 23:51
|
|
|
Cross-Section posted:Are those Kryosheet thermal contact pads that I occasionally see advertised on GN/HU any good if you're pretty much just gaming and not doing anything CPU-intensive? I havent tried the Kryosheet ones, but I use the IC graphite thermal pads extensively for my dumb CPU collection. Probably have 10 of em. They are absolutely good enough for regular use, providing you arent trying to eek out every degree. They are a few degrees hotter generally speaking, but it makes no difference in 99% of regular use applications from my testing. The one thing I will say is that the pads are electrically conductive and they do move around when mounting, and they can also get cut around the edges and shed tiny filaments of conductive graphite around, so you do have to be more careful than you generally would be when mounting with paste. https://www.youtube.com/watch?v=2owXLBV2e7g
|
|
#
?
Nov 12, 2023 23:55
|
|
|
Klyith posted:Stop worrying about precise application. Draw a big X from corner to corner on the IHS, cap the tube, and then put the heatsink on. I literally do just this and there's still a pretty high probability of paste ending up all over my fingers, or on the pins, or even in the slot. I blame my shaky big butterfingers. Not to mention all the rubbing alcohol and paper towels I end up going through just switching CPUs or coolers around. Cygni posted:The one thing I will say is that the pads are electrically conductive and they do move around when mounting, and they can also get cut around the edges and shed tiny filaments of conductive graphite around, so you do have to be more careful than you generally would be when mounting with paste. This, however, sounds terrifying because at the very least, the paste (well, the paste I use) is non-conductive so the only thing that really gets affected when I klutz up with the paste is the resale value lol
|
|
#
?
Nov 13, 2023 00:06
|
|
|
Cross-Section posted:Not to mention all the rubbing alcohol and paper towels I end up going through just switching CPUs or coolers around. Grey fingertips are a sign of diligent work.
|
|
#
?
Nov 13, 2023 02:44
|
|
|
Subjunctive posted:Grey fingertips are a sign of diligent work. or aluminum oxide
|
|
#
?
Nov 13, 2023 03:05
|
|
|
I do a small but thick X, put the heat sink on and move it around in squares to spread the paste before clamping down. Someone’s probably gonna tell me that’s bad, but my temps are fine.
|
|
#
?
Nov 13, 2023 11:28
|
|
|
Who cares as long as it works. I do five solid "dots" on AM4 and have no issues with temps.
|
|
#
?
Nov 13, 2023 11:35
|
|
|
|
|
#
?
Nov 13, 2023 13:23
|
|
|
Cross-Section posted:I literally do just this and there's still a pretty high probability of paste ending up all over my fingers, or on the pins, or even in the slot. I blame my shaky big butterfingers. PTM7950 I think is non conductive and won't get everywhere during application. Not reusable though and kind of a pain to line up/remove the plastic film. I've started using it instead of paste cause it's supposed to last longer though.
|
|
#
?
Nov 13, 2023 14:46
|
|
|
I, for one, am totally down with this miraculous resurgence of Oculink. https://liliputing.com/minisforums-mini-pc-with-a-70w-ryzen-7-7840hs-chip-and-63-gbps-oculink-port-now-available-for-479-and-up/ Kinda makes me regret buying the Minisforum B550.
|
|
#
?
Nov 13, 2023 17:13
|
|
|
SwissArmyDruid posted:I, for one, am totally down with this miraculous resurgence of Oculink. feeding 70 Watts to a 7840HS should show us how the upcoming desktop APUs will perform
|
|
#
?
Nov 13, 2023 17:26
|
|
|
FuturePastNow posted:feeding 70 Watts to a 7840HS should show us how the upcoming desktop APUs will perform i am not expecting too much out of it, realistically speaking, the limiting factor for a mini PC without dedicated GPU these days is what the igpu can handle, and iirc, that's still a 780M in the 8000 series. Which, I mean, I guess it's fine, but at the same time, I was kind of hoping for igpu improvements to keep pace with their dgpu silicon, or what's the point. SwissArmyDruid fucked around with this message at 18:44 on Nov 13, 2023 |
|
#
?
Nov 13, 2023 18:41
|
|
|
FuturePastNow posted:feeding 70 Watts to a 7840HS should show us how the upcoming desktop APUs will perform I suspect handhelds and laptops with soldered LPDDR5X will be faster at gaming. Being limited to DDR5-5600 SODIMMs in a mini-pc will place a big limit on how much performance you can expect. Desktop APUs will be a little better, but not much. Dr. Video Games 0031 fucked around with this message at 01:09 on Nov 14, 2023 |
|
#
?
Nov 13, 2023 23:21
|
|
|
Is CAMM going to be a thing for this gen, or only the one after this?
|
|
#
?
Nov 14, 2023 00:25
|
|
|
Tuna-Fish posted:Is CAMM going to be a thing for this gen, or only the one after this? I don't think JEDEC have even finalized the spec yet
|
|
#
?
Nov 14, 2023 00:38
|
|
|
Tuna-Fish posted:Is CAMM going to be a thing for this gen, or only the one after this? This gen of memory, ddr5? Yes. This gen of CPUs / platfroms? Dunno, seems to be a 2024 thing at the earliest now. They were talking about it happening this year, but then nobody was buying PCs. It's definitely not a thing you launch in a cold market -- the biggest deal for CAMM is high-end  laptops where shaving 3mm is big news.Arzachel posted:I don't think JEDEC have even finalized the spec yet Pretty sure they did? They definitely had a vote to approve it. And it's being extended: https://www.jedec.org/news/pressreleases/jedec-expands-camm-standardization-include-two-key-memory-technologies
|
|
#
?
Nov 14, 2023 01:03
|
|
|
I read that but I still don’t know what a CAMM is.
|
|
#
?
Nov 14, 2023 01:26
|
|
|
Subjunctive posted:I read that but I still don’t know what a CAMM is. new memory module format developed by dell and accepted as a jedec standard for anyone to use. it uses compression mounting to keep a low profile, and is designed in such a way to ensure shorter trace lengths. it strikes a compromise between the performance and low-profile nature of soldered memory with the modularity of sodimms   
|
|
#
?
Nov 14, 2023 02:03
|
|
|
Wild—laptop only at this point I assume, unless ASRock has a special baby that supports it.
|
|
#
?
Nov 14, 2023 02:49
|
|
|
Dr. Video Games 0031 posted:new memory module format developed by dell and accepted as a jedec standard for anyone to use. it uses compression mounting to keep a low profile, and is designed in such a way to ensure shorter trace lengths. it strikes a compromise between the performance and low-profile nature of soldered memory with the modularity of sodimms Only thing I'd add is that the link is strictly point-to-point, and modules are 128 bits. So a normal laptop would have exactly one of those, and to upgrade memory you'd have to remove it and replace it with a larger one.
|
|
#
?
Nov 14, 2023 03:02
|
|
|
Subjunctive posted:Wild—laptop only at this point I assume, unless ASRock has a special baby that supports it. Basically, it exists to allow socketable LDDR.
|
|
#
?
Nov 14, 2023 03:49
|
|
|
Subjunctive posted:Wild—laptop only at this point I assume, unless ASRock has a special baby that supports it. hobbesmaster posted:Basically, it exists to allow socketable LDDR. Somewhere, an ASRock engineer's ears started to burn a little bit...
|
|
#
?
Nov 14, 2023 08:04
|
|
|
Assuming they could lift their head out of the pile of cocaine they're swimming in face-first long enough to notice.
|
|
#
?
Nov 14, 2023 08:43
|
|
|
This was released from Micron a few days ago: LPCAMM2 seems to be a pretty big part of their future plans. DDR5 in general seems like its going to have a long life, no replacement on the horizon by 2029ish. Also GDDR7 seems to sync with rumors for next gen GPUs in early 2025.
|
|
#
?
Nov 14, 2023 09:26
|
|
|
Cygni posted:
LPCAMM is actually, amusingly, a competing standard to CAMM. It's not just LPDDR on a CAMM module, which is in the spec, but a separate LPDDR-only compression attached memory module that isn't a JEDEC standard and isn't compatible with CAMM. LPCAMM is basically just a socket for LPDDR chips, the compression mount is directly under the dies:   Compare that to CAMM where the dies are on another part of the PCB: 
|
|
#
?
Nov 14, 2023 09:44
|
|
|
BurritoJustice posted:LPCAMM is actually, amusingly, a competing standard to CAMM. It's not just LPDDR on a CAMM module, which is in the spec, but a separate LPDDR-only compression attached memory module that isn't a JEDEC standard and isn't compatible with CAMM.
|
|
#
?
Nov 14, 2023 10:10
|
|
|
Looking at the two I don't see what advantage the surface area of camm had over lpcamm
|
|
#
?
Nov 14, 2023 12:02
|
|
|
|
| # ? May 29, 2024 16:48 |
|
|
ijyt posted:Looking at the two I don't see what advantage the surface area of camm had over lpcamm LPCAMM can only fit large amounts of memory if there are many dies in each memory package. CAMM can do larger amounts of memory with more traditional packaging. The advantage (other than compactness) for LPCAMM is that trace lengths are even shorter than CAMM, and it should be able to reach very close to the power use/maximum speeds of soldered memory.
|
|
#
?
Nov 14, 2023 12:17
|
|