|
Inverse square posted:Thanks but no, I have a tower I was going to use. Also I notice that all of those PCs either have a 970 when I'd prefer a 980, or they have a hardrive when I really only want one SSD in there. nobody ever got a good deal buying a prebuilt computer. Visit the PCPartspicking thread; tell them your needs, and they will help you put together a list of quality, reliable parts that will give you the best value for your money. then you can either put it together yourself (it isn't hard, and the OP of that thread has instructions) or visit a local PC store that will sell you the parts and also put it together for you for a small fee. Last time I checked, MSI had the quietest coolers. GPUs are fairly quiet when not under heavy load. Even then they aren't that noisy especially if you make smart decisions with your choice of case.
|
#
?
Feb 27, 2016 05:50
|
|
|
|
| # ? Jun 15, 2024 05:54 |
|
|
Just for the record building a PC is scary the first time you do it but it's super, super easy in reality. Watch a half hour of internet tutorials, spend an hour reading the motherboard manual (which will tell you where everything gets plugged in for your specific motherboard). Boom, you save a bunch over a prebuilt plus all the components are much more solid than you would ever get from an OEM. OEMs always cheap out on PSUs, sometimes their GPUs are weaker than the comparable retail parts, their coolers are often crap, and you'll pay a premium on stuff like memory and SSDs. The only real advantage is that you have someone to call when stuff breaks, but it's often pretty straightforward to debug especially when you have access to the internet's premiere  forum. Tell us a sob story and we'll probably help.However, be aware that thermal paste is an active religious warzone between the "Paste Spreading Federation" and the "Grain Of Rice Collective". We're mostly talking a degree or two at most and it mostly comes down to how skillfully you drop the chip and mount it, or how evenly you can spread. But everyone, however, can agree that MSI is wrong (what in the actual gently caress). Literally the only step you need to be super careful on is the CPU installation, because you can legitimately bend the pins very easily if you apply lateral force and it's all over if you do. But watch the videos, align the arrow, drop the chip straight in, and lock the socket down right away and you'll be fine. And nowadays the pins are on the mobo rather than the chip, which makes gently caress-ups cheaper since the motherboard is like $80 vs a $200+ CPU. Anything else you can probably work around - I've ripped off my share of PCIe locking tabs in my day and so on. Paul MaudDib fucked around with this message at 06:44 on Feb 27, 2016 |
|
#
?
Feb 27, 2016 06:26
|
|
|
Just don't take us seriously when we try to upsell you on something because it makes us feel better about spending ludicrous sums on tech by making it seem normal. Buy it anyway because it's awesome.
|
|
#
?
Feb 27, 2016 06:30
|
|
|
Paul MaudDib posted:But everyone, however, can agree that MSI is wrong[/url] (what in the actual gently caress). Oh my god! What is even happening there? Does he think he's icing a cake?
|
|
#
?
Feb 27, 2016 06:50
|
|
|
Peanut3141 posted:Oh my god! What is even happening there? Does he think he's icing a cake? On the principle that more is always better, MSI suggests you bathe your motherboard in toothpaste for improved thermal conductivity.
|
|
#
?
Feb 27, 2016 07:57
|
|
|
FaustianQ posted:On the principle that more is always better, MSI suggests you bathe your motherboard in toothpaste for improved thermal conductivity. Use as much as you want! Paste is what keeps it cool! For the love of god, don't do this. Although toothpaste apparently isn't the worst nor the best around. Mayonnaise is apparently an even better paste for the first two-three days, and then the smell becomes unbearable. 
Paul MaudDib fucked around with this message at 08:22 on Feb 27, 2016 |
|
#
?
Feb 27, 2016 08:06
|
|
|
Paul MaudDib posted:Just for the record building a PC is scary the first time you do it but it's super, super easy in reality. Watch a half hour of internet tutorials, spend an hour reading the motherboard manual (which will tell you where everything gets plugged in for your specific motherboard). Boom, you save a bunch over a prebuilt plus all the components are much more solid than you would ever get from an OEM. Yeah, I just recently broke a PCIe tab too. Would it still be okay using it? Would super gluing it back on work?
|
|
#
?
Feb 27, 2016 08:43
|
|
|
The little plastic tabs are just for retention at the back of the slot. It's not electrically needed, just a bit more insurance against movement. I've been running one of my cards. In a slot with no retention tab for years now.
|
|
#
?
Feb 27, 2016 09:40
|
|
|
Mutant Standard posted:Yeah, I just recently broke a PCIe tab too. Would it still be okay using it? Would super gluing it back on work? For the love of god do not pour any adhesive substances into your electrically conductive slot. The locking tab doesn't actually matter and there is literally no way that this repair technique will work out for you. You'll only manage to gently caress up your motherboard's contacts.
|
|
#
?
Feb 27, 2016 10:02
|
|
|
Double the bit about the CPU. It doesn't help that when you put down the lever for the first time, it feels like you're crushing something, but trust me, so long as you've got the chip centered in the socket, and it is very obvious if you have or have not, you gotta remember that you are basically putting enough force to compress 1151 or 2011 spring-loaded pins, depending on the socket. (Yes, Intel numbers their sockets after the number of pins in it.)
|
|
#
?
Feb 27, 2016 10:03
|
|
|
I remember completely loving the AGP slot on a socket A motherboard with 400FSB and DDR400 support and weeping because out of warranty. PCIe sockets seems to be built much stronger.SwissArmyDruid posted:Double the bit about the CPU. It doesn't help that when you put down the lever for the first time, it feels like you're crushing something, but trust me, so long as you've got the chip centered in the socket, and it is very obvious if you have or have not, you gotta remember that you are basically putting enough force to compress 1151 or 2011 spring-loaded pins, depending on the socket. (Yes, Intel numbers their sockets after the number of pins in it.) Having been with AMD and PGA for so long, my first experience with an LGA socket freaked me the gently caress out. I think I spent at least 15 minutes staring at the very simply retention mechanism in LGA775 wondering how I was loving this up and whether anything still worked. Of course, 939 to AM3+ had the nifty "you didn't change your paste in the last 6 months, enjoy removing the CPU and HSF at the same time".
|
|
#
?
Feb 27, 2016 10:28
|
|
|
Paul MaudDib posted:Just for the record building a PC is scary the first time you do it but it's super, super easy in reality. Except nothing prepares you for that first time you lower the retention arm down on your first LGA chip and think to yourself "I shouldn't be putting *this* much pressure on this loving CPU, should I?" Everything after that moment, save maybe getting the front panel connectors done with the absence of a 'guide block' is easy as hell, provided you don't have some DoA part right out of the box.
|
|
#
?
Feb 27, 2016 11:41
|
|
|
Nah I thought this is how you build a computer, https://www.youtube.com/watch?v=b85h_ujZ_vg
|
|
#
?
Feb 27, 2016 12:42
|
|
|
mpyro posted:Nah I thought this is how you build a computer, https://www.youtube.com/watch?v=b85h_ujZ_vg Thank you for this. I haven't laughed that hard in a while.
|
|
#
?
Feb 27, 2016 16:10
|
|
|
mpyro posted:Nah I thought this is how you build a computer, https://www.youtube.com/watch?v=b85h_ujZ_vg Haha the gently caress did I just watch?
|
|
#
?
Feb 27, 2016 17:16
|
|
|
mpyro posted:Nah I thought this is how you build a computer, https://www.youtube.com/watch?v=b85h_ujZ_vg I loving lost it at the banana. I mean stabbing poo poo with a screwdriver was painful but that's only because of bad memories (SOCKET A!!!  )
|
|
#
?
Feb 27, 2016 17:21
|
|
|
Considering some of the posts in the PCPartPicking thread I can only assume this is how people do their first build. With fewer eggs and breakfast cereals, obviously.
|
|
#
?
Feb 27, 2016 17:23
|
|
|
They're pairing Fury X2 with a blower cooler?  https://twitter.com/Roy_techhwood/status/703363511670349824
|
|
#
?
Feb 27, 2016 18:39
|
|
|
wat How does that even work, won't that create an imbalance in cooling? Also preemptive left field "Haha, jokes on you, we meant the Fury X Two, not the Fury X2!"
|
|
#
?
Feb 27, 2016 18:55
|
|
|
FaustianQ posted:How does that even work, won't that create an imbalance in cooling? Sure sounds like the thing would be noisy as hell but I have no clue what "an imbalance in cooling" is supposed to be
|
|
#
?
Feb 27, 2016 19:16
|
|
|
they should just integrate the heatsink into the case itself and then you just drop in a mitx mobo and cpu and then you're done
|
|
#
?
Feb 27, 2016 19:42
|
|
|
Paul MaudDib posted:However, be aware that thermal paste is an active religious warzone between the "Paste Spreading Federation" and the "Grain Of Rice Collective". We're mostly talking a degree or two at most and it mostly comes down to how skillfully you drop the chip and mount it, or how evenly you can spread. But everyone, however, can agree that MSI is wrong (what in the act Actually the X method is the best application method. https://www.pugetsystems.com/labs/articles/Thermal-Paste-Application-Techniques-170/ 
|
|
#
?
Feb 27, 2016 19:46
|
|
|
Col.Kiwi posted:What is this supposed to mean exactly? The blower would pull in cool air, move it past gpu heat sink one, then the warmer air would move across the second gpu's heatsink before exiting the case. The second GPU would always thermally throttle before the first, preventing you from using both to max potential. If that's how it's actually designed.
|
|
#
?
Feb 27, 2016 19:48
|
|
|
Col.Kiwi posted:What is this supposed to mean exactly? The die closest to the fan being much cooler, resulting in potentially weird throttling issues as the die further from the fan ends up clocking lower? Look at the designs for most dual GPU coolers (590, 690, 6990, 7990, 295X2, etc) and placement of fan compared to stock blowers meant for single gpu boards. The Fury X2 appears to be equipped with the latter as a cooling solution, which is odd.
|
|
#
?
Feb 27, 2016 19:51
|
|
SlayVus posted:Actually the X method is the best application method. What is the margin of error on those measurements?
|
|
|
#
?
Feb 27, 2016 20:01
|
|
|
AVeryLargeRadish posted:What is the margin of error on those measurements? shh not posting error bars invalidates 99.9% of all "benchmarks" ever done
|
|
#
?
Feb 27, 2016 20:05
|
|
|
Zero VGS posted:The blower would pull in cool air, move it past gpu heat sink one, then the warmer air would move across the second gpu's heatsink before exiting the case. The second GPU would always thermally throttle before the first, preventing you from using both to max potential. If that's how it's actually designed. I've never owned or installed a dual GPU card, I'd only seen photos and handled aftermarket coolers and I was only familiar with designs that use a single larger heatsink to cool both GPUs. Which I believe would lead to roughly equal gpu temps no matter the fan(s). Designs like this: https://www.arctic.ac/worldwide_en/accelero-twin-turbo-690.html So I assumed all dual GPU cards used a large single heatsink design. Googling it shows I'm very wrong. However all the cards I can find on google image search with two seperate heatsinks for the two GPUs are open air cooler designs with multiple fans. I think you're worrying about nothing. I'd imagine they were smart enough not to design a cooler with two seperate heatsinks and a single blower fan at the end, because as you explained that is just an obviously stupid design. (edit: if unclear, I'm saying since we know it's a blower I think we are probably safe to assume they used a single large heatsink) I'm not an engineer so it's always possible I'm missing something here. Col.Kiwi fucked around with this message at 20:20 on Feb 27, 2016 |
|
#
?
Feb 27, 2016 20:12
|
|
|
SlayVus posted:Actually the X method is the best application method. The only thing to take from this graph is that they didn't do enough tests and as far as in practice, do whatever is easiest. Let me try to contribute actually. I've laid down a lot of epoxy adhesive with a similar viscosity to many thermal compounds. Our spec calls for complete coverage of the faying surfaces. The plates are then torqued to 20 inch pounds with some screws. Our spec calls for complete coverage on one of the surfaces before they are mated. This is achieved by laying down an arbitrary pattern, then using a spreader that has small 1/20" (i do not remember the actual spec) fins on it to smooth it. Once the parts are torqued down, the inspection is ensuring that there is a bead of squeeze-out around the entire diameter of the surface of no more than 1/8". This is all necessary because obviously you can't pull the thing apart and look at it after the fact. From this I would suggest that a good process would be to use the dumb old credit card technique followed by a rice grain in the center and then checking for a small bead of squeeze out. Also, it is a good idea to try one application, clamp the chip, then remove it and check the coverage. Clean it all off again and adjust if necessary. Oh yeah, and if you are worried about the squeeze out screwing stuff up, make sure you are using a non-conductive paste/grease. LRADIKAL fucked around with this message at 20:38 on Feb 27, 2016 |
|
#
?
Feb 27, 2016 20:29
|
|
|
THE DOG HOUSE posted:they should just integrate the heatsink into the case itself and then you just drop in a mitx mobo and cpu and then you're done Zalman had some fun ones in that regard.
|
|
#
?
Feb 27, 2016 21:16
|
|
|
Jago posted:Let me try to contribute actually. I've laid down a lot of epoxy adhesive with a similar viscosity to many thermal compounds. I guess the main difference is that with thermal compound, you want the bare minimum to fill the micro-scale voids between the contacting metal surfaces and no more. Any extra just increases the overall thermal resistance.
|
|
#
?
Feb 27, 2016 22:35
|
|
|
Jago posted:The only thing to take from this graph is that they didn't do enough tests and as far as in practice, do whatever is easiest. The problem with computers though is that your method would eventually damage something. Twenty inch pounds is way too much for the CPU and socket. Just a few months ago Intel CPUs were getting bent because too much force was used from normal mounting brackets. With a CPU you just can't torque out the air bubbles that you would get with a spread method. Which is why you go with something like a pea or line method.
|
|
#
?
Feb 27, 2016 22:48
|
|
|
Wait you guys don't solder it
|
|
#
?
Feb 27, 2016 23:54
|
|
|
I used to be on a 60hz 2560x1600 primary and 60hz 1920x1080 I swapped out the 1920x1080 with a 144hz gsync in the same resolution When I run a twitch stream or watch a full screen video file on the large monitor I have issues with windows applications. They will close like normal but the image of the window stays behind. Or the application will open but the application wont display even though I can see it on my task bar. Im sure it has something to do with mismatched hz. When I set the the gsync monitor to 60hz it doesnt happen. Is there a way I can force 144hz only in games and 60hz on desktop? Fauxtool fucked around with this message at 01:39 on Feb 28, 2016 |
|
#
?
Feb 28, 2016 01:37
|
|
|
SlayVus posted:The problem with computers though is that your method would eventually damage something. Twenty inch pounds is way too much for the CPU and socket. Just a few months ago Intel CPUs were getting bent because too much force was used from normal mounting brackets. Wow, no poo poo, use the normal retention hardware as indicated. Point I missed saying, was that even at that amount of compression, I have seen first hand that there can remain gaps where no goo spread.
|
|
#
?
Feb 28, 2016 02:20
|
|
|
Jago posted:Wow, no poo poo, use the normal retention hardware as indicated. The thing about CPUs is that you're really only looking for proper thermal conductivity right on the die itself. Since the dies are so small and are in the center, the rest of the heatplate is really only there to distribute load. The dot method with properly spreading thermal compound is all you need for applications in regards to CPUs: https://www.youtube.com/watch?v=EyXLu1Ms-q4
|
|
#
?
Feb 28, 2016 02:42
|
|
|
I am looking at buying a new video card this week. I'm not huge on video games and my 560ti has kept up nicely. I want to get a video card that is Vulkan capable for SPIR-V. I'm looking at around the $325 mark (AUD) which is either a 960 or a 380 from what I can see. I don't understand the market very much, what would be the better of the two for GPGPU (not using CUDA).
|
|
#
?
Feb 28, 2016 02:51
|
|
|
Rukus posted:The thing about CPUs is that you're really only looking for proper thermal conductivity right on the die itself. Since the dies are so small and are in the center, the rest of the heatplate is really only there to distribute load.
|
|
#
?
Feb 28, 2016 16:23
|
|
|
FaustianQ posted:I remember completely loving the AGP slot on a socket A motherboard with 400FSB and DDR400 support and weeping because out of warranty. PCIe sockets seems to be built much stronger. That's why you were supposed to twist it off. =P
|
|
#
?
Feb 28, 2016 16:47
|
|
|
So I guess this goes here, but apparently there is a rumor floating about that Bristol Ridge will come at the high end with 16CU parts. Considering there is no sign of HBM for Bristol APUs, I'm a bit dismissive of it (to say nothing of total die size). But say AMD is doing it, how well would 16CU would perform dual channel or even quad channel DDR4 from 2400mhz to 4000mhz? Would this be roughly Iris Pro 6200 performance, or would it choke and only 12 CU is needed to match an Iris Pro 6200?
|
|
#
?
Feb 28, 2016 17:45
|
|
|
|
| # ? Jun 15, 2024 05:54 |
|
|
My method is the "Alright, a nice drop here oops it stringed out well not of it hit the motherboard, aggggh the heat sink is unaligned aaaah gently caress it just moved off the cpu gently caress it gently caress it the temps will be fine."
|
|
#
?
Feb 28, 2016 18:08
|
|