|
Arzachel posted:How the heck do you manage to be CPU bound in new games unless you're running some weird 1080p/2080ti combo or are turning down settings to hit 144hz?
|
#
¿
Jan 28, 2019 00:51
|
|
|
|
| # ¿ May 2, 2024 02:46 |
|
|
The one test someone did that claimed to significantly saturate and bottleneck on PCIe 3.0 x8 links took two RTX Titans doing a unusual compute task that generates a lot of PCIe traffic while not using the NVLink bridge that anyone actually doing that particular task would absolutely use. So basically nothing a consumer would do with a consumer GPU. Ashes of the Singularity also encounters a bandwidth limit on PCIe 3.0 x8, but only in SLI and only because Ashes of the Singularity is programmed in a way that disables the use of the NVLink bridge between the cards. Indiana_Krom fucked around with this message at 00:55 on Mar 23, 2019 |
|
#
¿
Mar 23, 2019 00:48
|
|
|
 This conversation is distressing because I just swapped my case and radiator fans with chromax 140mm because a couple of my Vardars were starting to squeak. Granted, the chromax fans I got are significantly quieter at all levels than the vardars while they perform about the same (within 1C) at load, so I guess I can live with it.
|
|
#
¿
Mar 30, 2019 19:20
|
|
|
Risky Bisquick posted:That much power and I/o puts it firmly up against high end spec pc’s. It’s going to be quite a powerful platform. Features yes, it is matching up against a high end PC quite well, but remember that a high end PC usually falls somewhere around 400-500w power consumption. This console should be good, but a high end PC of the same generation will still do all the same stuff three times faster for four times the energy cost.
|
|
#
¿
Apr 16, 2019 22:38
|
|
|
Wouldn't DDR5 require a different socket? Or at a bare minimum a different motherboard, because DDR5 is probably going to be keyed differently even if it has the same number of contacts?
|
|
#
¿
Apr 20, 2019 01:05
|
|
|
PC LOAD LETTER posted:Only way they'll pull off 4K @ 60fps is with scaling up from 1080p-1200p resolution. The bandwidth and GPU requirements are just too high for a native 4K @ 60fps console to be sellable even by late 2020 for typical console prices. I think by "late 2020s" he was meaning closer to 10 years from now, not next year. That is a good 2-3 generations down the line, I could believe 4k/60 native would be within reach of consoles by then.
|
|
#
¿
May 4, 2019 13:48
|
|
|
PC LOAD LETTER posted:I think you misread. Errr, sorry, thought you were replying to a different post that did say late 2020s... Yeah, its not happening at all in the next gen.
|
|
#
¿
May 4, 2019 17:39
|
|
|
Probably the main reason it isn't a baseline feature is cost; a non ECC DIMM will have some multiple of 8 chips on it while an ECC DIMM will have some multiple of 9 chips on it (for storing the parity data).
|
|
#
¿
May 6, 2019 22:13
|
|
|
Video games always run at the speed of the slowest component in the system. Most of the time its the GPU because it is an easy target, but sometimes games are just poorly optimized or poorly threaded and hit the CPU harder. Take the modern Far Cry games for instance, New Dawn bottlenecks on a i9-9900k at around 120 FPS average, which can be easily reached with a GTX 1080 Ti card at 1080p/ultra. Lower the resolution and the peaks reach higher, but the average barely moves from 120-125. Getting a bigger GPU like a 2080 Ti or turning down the graphical details is pointless in those games outside of 4k resolution (even 1440p is getting dangerously close to the CPU bottleneck on them). Actual CPU usage on a 9900k is around 30-40% while running the far cry benchmarks, but there are two threads that hit 100% usage and likely cause the bottleneck (one of them is probably the Denuvo DRM). The newest Tomb Raider game also bottlenecks pretty hard on anything less than an 8700k.
|
|
#
¿
May 11, 2019 12:52
|
|
|
Threadkiller Dog posted:I'm getting a 3900X for longevity mainly. Just in case games a few years from now suddenly start thirsting for more threads. I think something in the range of a 3900X or 9900K should still be adequate for games around 60 FPS in ~6 years. The 2500k/2600k were viable for so long because single threaded performance has only made relatively small improvements since they were released and there is not really any reason to expect that trend to change. If anything I suspect single threaded performance will increase less in the next 6 years than it has in the previous 6.
|
|
#
¿
Jun 8, 2019 22:18
|
|
|
Even in the US I do a lot of looking up stuff on Newegg/other retailers, and then make my actual purchase on Amazon. Though a lot of that is because Amazon's search algorithm is so horrible it makes me think the people that developed it must be actively malicious in wanting it to return garbage results. But outside of the horrible store front and search, once you actually purchase something on Amazon the customer support and service is better than pretty much anyone else on the web. I would certainly never buy a monitor off Newegg for instance, even if it saved me $50 or more, their return policy for stuff like displays is pretty horrible (like they conform to the cheapest possible standard they can get away with legally). Often it is just worth a little extra cash to have some assurance of real customer service if something goes wrong, especially in the US where regulators and consumer protections are inconsistent and unreliable at best.
|
|
#
¿
Jul 5, 2019 14:43
|
|
|
I think it is safe to say the 3900X is basically at parity with a 9900k in games if you take a look at the techpowerup review. The review pairs good memory settings/speeds on all platforms with a 2080 Ti and the difference between a 9900k and a 3900X is barely 7% even at 720p. And if you axe Far Cry 5 from the benchmark is probably a lot closer past 720p (that game is so horribly single thread IPC dependent it is ridiculous: it doesn't flatten out on a GPU limit till 4k). Any game that even slightly leans in to multi-threaded performance has the 3900X neck and neck with the 9900k.
|
|
#
¿
Jul 7, 2019 15:36
|
|
|
Sidesaddle Cavalry posted:What would a semi-custom perf process entail that would be different from the mobile process, anyways? We talking implementation of ASIC features in general or something more on the materials side of things, i.e. doping? Saw a video a while back and the crazy thing is that transistors are fundamentally analog devices, they don't switch on and off perfectly but instead they generate more of a voltage curve. A process can be optimized for different points on that voltage curve. A mobile process is going to try to make the transistors "off" state be as close to 0v as possible in order to reduce leakage which saves power/heat, but optimizing for that comes at the expense of switching time. More "off" transistors require more time to reach "off" so they switch slower. Where as faster transistors also leak more and consume more power overall in exchange for being able to very quickly cross back and forth from their threshold voltages. If you look at TSMCs 7nm as a process, I think it is for FinFET basically what 28nm was as the last practical high frequency planar process. The problem is the gates are getting to be too small even with FinFET and its causing power consumption/density to be too high for use in high frequency transistors. Intel didn't tighten up their geometry in 14++ to hit higher clock speeds, they actually made several aspects of their transistors bigger to make it work. This is likely why everyone is racing to develop GAAFET. 7nm is smaller and uses less power than larger nodes, but isn't any faster than those larger nodes. This is probably also a big component of what is wrong with Intel's 10nm which is in some ways a more aggressive target than TSMCs 7nm, they are all too small to control leakage at high frequencies.
|
|
#
¿
Jul 21, 2019 20:57
|
|
|
The only downside of using "too much" thermal paste is cleaning it up afterwards, as far as performance goes only using too little is a problem (when it fails to cover the entire surface), it is basically impossible to use too much. You could literally coat your entire socket in the stuff and it wouldn't hurt the thermals, but it would be a huge pain to clean up. So always just use enough that you can see a small amount get squeezed out of the edges uniformly and it will give you optimal cooling without being excessively difficult to clean up should you ever need to redo it.
|
|
#
¿
Aug 23, 2019 03:00
|
|
|
It wasn't always overblown, I recently dusted out my old 24" 144 Hz / 1080p TN and yeah it does look washed out because its best contrast ratio is about 750:1. But the AUO TN (24" 240 Hz / 1080p) I'm using now holds 960:1 and honestly looks fine even if I bring up the same image on the 1100:1 dell U2412M sitting next to it and compare them. Although it helps immensely that both monitors have been adjusted to the same 200 cd/m2 and 6500k white point using their respective OSD controls and a colorometer. Really any cheap screen is going to look like crap regardless of the panel type. A $175 monitor doesn't look like crap because its a TN, it looks like crap because its a $175 monitor.
|
|
#
¿
Sep 7, 2019 16:42
|
|
|
Victory Position posted:I asked for "give me a video card that stays cool forever" and was pointed to this thing: https://www.tomshardware.com/reviews/sapphire-rx-vega-64-nitro,5388.html Are you sure your friend didn't mishear or troll you? Because the card he recommended is an inefficient monstrosity that puts out a ton of heat.
|
|
#
¿
Sep 7, 2019 21:48
|
|
|
Its performance is going to be fine mind you, it just doesn't have a giant 3 fan x 2.5 slot cooler because the designers thought it would look pretty.
|
|
#
¿
Sep 7, 2019 22:06
|
|
|
My gigabyte board has two 10k thermristor headers which I use with an XSPC g1/4 plug that fits into where the plugs for the alternate coolant line routes in my GPU block are. BIOS had some stupid limitations like the CPU fan header can only be set to monitor off the CPU highest core temperature, so I had to plug all the fans in to aux/case/pump headers, but otherwise it works. It takes a couple minutes for the fans to begin to spin up significantly when I dump a solid 375w of power into the loop, and around one minute to ramp back down once I go back to idle (because they don't even start speeding up till it passes 34C, and long idle is 27-28C). Before that I used to use speedfan with custom curves which behaved roughly the same once I got it dialed in even basing it off the actual CPU/GPU temps, but since it is no longer updated and wouldn't work on newer boards I had to find a different solution. Otherwise Corsair makes a fan controller that can do it, but an aquacomputer is the ultimate high end + brute force solution (completely independent programmable computerized multi-channel fan controller).
|
|
#
¿
Sep 14, 2019 23:54
|
|
|
Black Griffon posted:Can I talk about AMD motherboards here? Just installed my new Aorus, and there's this exhausting list of apps that I've never head about. Anyone know if any of these are useful for anything? I can't talk AMD motherboards, but I can talk gigabyte motherboards: Uninstall APP Center entirely and ignore that it ever existed, you do not need any of it unless you are dead set on controlling the RGB stuff outside of BIOS and even then you should not do it because RGB Fusion is an epic security hole. (The security on RGP Fusion is so hilariously exploitable that it casts doubt on all the other applications in app center, they are probably just as bad.)
|
|
#
¿
Oct 10, 2019 22:54
|
|
|
Gigabyte has not resolved that issue, but you can work around it by using the toggle switches on the board to soft disable it, then it only shows up if you flash a new version of BIOS or clear cmos and most of the time it is better to redo the settings manually anyway when you do that.
|
|
#
¿
Oct 26, 2019 19:20
|
|
|
My current PC has more cache than my first computer had total storage.
|
|
#
¿
Nov 11, 2019 23:45
|
|
|
AIOs also have the disadvantage of using all aluminum radiators, so even the massive 360mm ones usually barely equal or downright perform worse than a D15. Basically if you are going to use a water cooler these days it should either be a full custom loop, or you shouldn't use it at all. I have a water cooler myself, but it is a full custom loop with a massive 420x140x40mm radiator and all the fans and the pump speeds are controlled by the coolant temperature so it is very quiet and consistent and slightly outperforms even a D15 class cooler on both temperature and noise (for about 5 times the price). It is great, I really like it and wouldn't have my system any other way, but definitely if you aren't a nerd with a nearly unlimited budget then just get a heat pipe tower and be done with it.
|
|
#
¿
Apr 2, 2020 00:28
|
|
|
Paul MaudDib posted:Whether that matters to you is a different question, but there are still a fair number of CPU-intensive titles like Battlefield V, Metro Exodus, Far Cry 5, RDR2 where the 9900K makes a difference in minimum framerates, etc. Just gonna chime in here a bit about Far Cry 5; I did some pretty extensive benchmarking at different resolutions and settings to find the performance bottleneck and even with just a GTX 1080 at 1080p/ultra that particular game still hits the CPU limit on a 9900k. The Ubershit DRM is probably responsible for some of that, but basically you have to go past 1440p before you consistently hit the GPU bottleneck there. Literally turning down the details or using an aggressive resolution scale makes barely any difference in that game (the averages max out at about 10 FPS higher, but that is mostly in the peaks because the minimums basically don't budge).
|
|
#
¿
Apr 3, 2020 22:20
|
|
|
Paul MaudDib posted:Far Cry 5
|
|
#
¿
Apr 11, 2020 01:10
|
|
|
Some Goon posted:Yeah, the heat generated by the CPU is just a function of the work it's doing, better cooling doesn't make it get less hot. Technically it does put out less heat if you keep it cooler because leakage drops with temperature, just not by very much. I was actually pretty surprised when I delidded my 7700k and replaced the paste with liquid metal, it dropped the load power consumption by roughly 5 watts because it was 20C cooler. I was honestly expecting the difference to be too small to measure, but it was 120W to 115W (~4%).
|
|
#
¿
Jun 22, 2020 23:54
|
|
|
eggyolk posted:Just switched from Intel to AMD. I delidded a 3770k using the razor blade method, was like an hours worth of effort and terror, but I succeeded and replaced the cheap TIM with something better and got a ~5C reduction. When it came time to do my 7700k, I just bought a tool for it. $30 tool was worth it in time and terror savings and also how it doubled as a clamp to hold the heat spreader in the correct position/orientation when I glued it back on with rtv gasket maker after applying liquid metal which yielded a 20C reduction in load temps. The although only "overclocking" I did was setting its all core turbo to 4.5 GHz with an unlimited duration.
|
|
#
¿
Aug 15, 2020 15:19
|
|
|
HalloKitty posted:The thing that still impresses me the most happens to be the humble microSD card. Goddamn there's so much packed in to a tiny space. You can get 1TB ones now! My current CPU has more cache than my first computer had in hard drive capacity.
|
|
#
¿
Sep 16, 2020 00:14
|
|
|
CaptainSarcastic posted:I have a Gigabyte Aorus Elite x570 but I refuse to install any of Gigabyte's software. It doesn't have a good reputation, and most of the functions their apps perform can be performed better by other applications or from the BIOS. Without a good reason I wouldn't run any of Gigabyte's software, especially not anything that requires their "app suite." The software has also been known to expose massive security holes. But this is hardly unique to Gigabyte, honestly nobody should run motherboard vendor utilities.
|
|
#
¿
Nov 15, 2020 04:23
|
|
|
Arzachel posted:Speedfan hasn't been updated in years, sadly. Argus is very solid but payware, Fan Control popped up recently and might be a freeware alternative but I haven't tried it. Thanks for pointing this one out, just checked it on my gigabyte z390 aorus master and it works for all the PWM fans (unfortunately it doesn't work for the DC fans). I have all my fan curves baked into BIOS anyway based off a thermristor in the coolant of my water loop, but I like having an option to temporarily override the curves and have been missing speedfan for the last couple years. Now if only this had a "delta" curve that could adjust fan speeds based off the difference between two temperature sensors (like say for instance the difference between the coolant temperature and the ambient temperature).
|
|
#
¿
Nov 15, 2020 14:10
|
|
|
Chaitai posted:I have 2 pairs of Corsair Vengeance LPX 16GB (2x8GB) DDR4 DRAM 3200MHz. Doesn't hurt to try, just make a backup if you are really concerned and then try plugging the drive, windows has gotten pretty resilient these days.
|
|
#
¿
Nov 19, 2020 16:58
|
|
|
bus hustler posted:Something I've noticed to this day is that motherboard manufacturers still basically say "if it is not broken there is no need to update the BIOS," when everyone else is obsessed with up to date BIOS including the security side of the house. Windows (and presumably other modern operating systems) can and do load microcode security fixes into the CPU during the boot process, so even if BIOS isn't doing it directly the updated microcode is still applied to the system at runtime. So basically check the change log of your BIOS and if it doesn't address an issue or compatibility you are actually experiencing then just skip it, there is no particular risk of missing security fixes.
|
|
#
¿
Dec 6, 2020 15:50
|
|
|
A nearly 20C delta between ambient and the coolant on 720mm worth of radiator is not that great even for 500w of load but if the fans are <500 RPM it is probably not too unusual. I've got 375w worth of load on a 420mm radiator (3x140) and a ~15C delta is about the limit where anything short of stopping the fans entirely dumps all the heat before it can get any warmer though, but bigger fans. You should be able to feel heat just pouring out of the radiators with that kind of delta. That being said, the heat spreader on modern CPUs is basically an insulator. The tiny size of the die combined with the high power consumption means they all run hot and there isn't much that can be done about it short of popping the spreader off and doing direct die cooling. CPUs regularly approaching 90C even on exceptionally good cooling is not uncommon or even a sign of a problem as long as they aren't throttling right away by a huge amount. As long as they sustain the performance/power output, your cooling is working normally and you shouldn't worry about it.
|
|
#
¿
Dec 15, 2020 23:19
|
|
|
K8.0 posted:Batteries will keep improving, but they by definition can't compete with fuels in terms of energy density. Gasoline, diesel, and kerosene are extremely close to ideal ideal chemical stores of energy. In cars it may be possible to outperform them in some senses because an electric drivetrain can be so much smaller, lighter, and more efficient than a combustion one, which frees up a lot of space and weight so batteries can lag behind on mass and size while keeping the total system competitive. For planes, this is never going to be the case. We'd need a fundamentally different sort of energy storage for combustion to ever be displaced there. Anyone telling you batteries are still going to improve their energy density sufficiently to compete with fuels is either totally ignorant of physics or straight up lying to you. The main improvements we'll hopefully see going forward are charge rate and durability. The theoretical energy storage capacity in Wh/kilogram for lithium type batteries is an astronomical number much higher than the energy density of common fossil fuels. The problem is getting a cell that could approach that theoretical capacity without immediately exploding or only having the endurance of a single cycle. It isn't that fossil fuels have some ideal energy density, it is that they are near the ideal state of being stable and easy to transport while also being volatile enough to extract meaningful work out of burning them.
|
|
#
¿
Dec 25, 2020 16:38
|
|
|
My CPU/GPU temps relative to the coolant temp have remained the same for the entire 3+ years since I installed the coolers on them with thermal grizzly kryonaut. If your temps have not changed it is not necessary to repaste.
|
|
#
¿
Jan 15, 2021 23:07
|
|
|
Lord Stimperor posted:Hey y'all need your help. Did you remove the plastic cover protecting the bottom of the AIO before installing it? Is the pump actually working? Those idle temps seem high enough to suggest the cooler isn't working at all. Prime95 has several modes: (large, blend, small), do they all crash, or does one mode crash more than the others? If there is a difference between modes it could help narrow it down (crashing frequently on small FFT but not as often on large would indicate CPU/cache issues, the opposite would indicate memory issues).
|
|
#
¿
Jun 13, 2021 12:24
|
|
|
For the little toy pumps in AIOs, yeah run them at 100% all the time. For a custom loop with a massive D5 pushing it though, go ahead and set it down to like 30-40% and keep it there forever.
|
|
#
¿
Jun 14, 2021 22:18
|
|
|
Lord Stimperor posted:By the way how do you all get this tremendous amount of hardware debugging experience? Are you all hardcore overclockers who live and breathe this stuff, does that knowledge overlap with an electrical engineering degree, or what is it? The advice I'm getting in this thread is far deeper and more consistent than what I would have received in random hardware forums (where it usually amounts to 'your 5000 watt power supply is probably underpowered'). That makes me feel really pampered and cared for. Probably can't speak for everyone, but nerds gotta nerd out. If the rest of them are anything like me, I was building PCs from junk parts in the basement when most kids would be playing with the easy lego kits so lots of us literally grew up doing this stuff. Just ask how many of us remember the days of jumpers and dip switches and other oddball things like using a conductive pencil to mark a line between two points on a CPU package to unlock it for overclocking. Hardware troubleshooting is a skill some people just develop from experience like smashing stuff together from a pile of old junk PC parts when we were kids and seeing if we could get it to work (and play Doom).
|
|
#
¿
Jun 19, 2021 13:17
|
|
|
CaptainSarcastic posted:The harshest lesson I learned in troubleshooting was to not skip steps. There's nothing like spending hours trying to figure out a software problem that was actually a hardware problem, or vice versa. Or doing a bunch of troubleshooting on a Mac showing virus-like network traffic only to discover that Limewire was choking out their slow-rear end connection. I recall an incident a couple years ago where my brother spent the better part of a weekend trying to get this USB-C docking station working with his laptop and it kept on refusing to work in one way or another (you could get parts of it to work together, but never all of them at once). Finally he called me since I am more in to hardware troubleshooting and the first thing I said was: "Lets get the simple things out of the way first, did you try a different type-C cable?". It was the cable. Once a local pastor brought his computer and printer which refused to print to my house, and I plugged the USB cable all the way into the printer for him. A company in town had a computer that wouldn't boot, hadn't worked in a week and they were at their wits end, I examined the case closely and used a ball-point pen to get the reset switch unstuck. Always check the basic solutions that are so simple it makes the people asking for help feel stupid, you would be amazed how many times it pays off, saves hours of troubleshooting and teaches people a valuable lesson that not every problem has to have a complicated cause.
|
|
#
¿
Jun 19, 2021 23:12
|
|
|
Lord Stimperor posted:... the Mainboard wasn't the problem. God loving damnit. Installed the new one, took me about 30 attempts to burn an iso to USB that would actually boot up and bam, blue screen. If the RAM passes memtest, and the motherboard is good, I'd suspect the CPU before anything else.
|
|
#
¿
Jun 23, 2021 22:57
|
|
|
|
| # ¿ May 2, 2024 02:46 |
|
|
I have two monitors, one is 1440p/240 Hz and the other is 1200p 60 Hz, right now my GTX 1080 is idling at 13w (139 MHz). Nvidia cards failing to idle only happens when you have three or more displays connected, as long as it is just two displays nvidia cards will still idle down normally without issue and don't seem to care about different refresh rates.
|
|
#
¿
Jul 2, 2021 00:55
|
|