|
slower but large cargo/passenger blimps using solar cells seems more plausible than trying to emulate high energy jet airliners with alternative fuel sources/mediums same thing with cargo ships (hybrid sail ships) and autos (trains) it's depressing that we're spending alot of energy (metaphorically and literally) on EVs while public transport systems are continuing to decay and VCs are funding dumb ideas like hyperloop
|
#
?
Dec 25, 2020 14:09
|
|
|
|
| # ? May 16, 2024 03:50 |
|
|
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.
|
|
#
?
Dec 25, 2020 15:07
|
|
|
Sir, this is the AMD thread
|
|
#
?
Dec 25, 2020 15:31
|
|
|
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
|
|
|
SwissArmyDruid posted:I'm given to believe that the real poo poo in 2H2 + O2 = 2H2O + heat is the LOX, as opposed to relying on ambient O2 This is why spacex sub-chills LOX until it's on the verge of becoming an oxygen slush: the density keeps going up.
|
|
#
?
Dec 25, 2020 19:02
|
|
|
PC LOAD LETTER posted:By all the rumors RKL gots a 250W PL2 power mode and Intel doesn't normally go doing that sort of thing for no reason. And there are plenty of "leaks" showing how it benches with a 125W or 250W power limit, Intel is not being subtle here with how they're presenting this thing or which way they want to push it. PL2 is short duration turbo and is not at all representative of what you will see in a steady state workload like gaming - that will fall back to PL1 which is usually equal to tdp assuming the chip is not overclocked either manually or with default settings by the mobo vendor. https://www.techpowerup.com/review/intel-core-i9-10900k/18.html Here's a review where they actually benched power consumption in games - note the 10900k is within 10w stock of the 3900x for total system power consumption. I would not expect the 11900k to be substantially different, just down 2 cores but with 18% better isoclock per core performance. I'm not saying the difference is irrelevant; I went ryzen myself because I couldn't fit a d15s and didn't want to go liquid. But it's gonna be nowhere near "180w to beat zen3" in games bad.
|
|
#
?
Dec 25, 2020 20:45
|
|
|
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.
|
|
#
?
Dec 25, 2020 21:13
|
|
|
No one who cares about gaming performance is going to care about short duration boost. They're going to be looking at sustained all core clocks. And if limited to 125W its not going to be that interesting, something like 4.2-4.3Ghz all core or so is what has been "leaked". That PL2 power spec is a pretty good indicator of what you can expect if you're going for sustained 5.3Ghz clocks like are being "leaked" everywhere and is what would be where things get more interesting. VorpalFish posted:assuming the chip is not overclocked either manually or with default settings by the mobo vendor. Oh yeah like neither of things are gonna happen. Just a minor caveat huh?
|
|
#
?
Dec 25, 2020 21:22
|
|
|
tbh i think both AMD and Intel have realized that gamers don't really care about power draw, they care about seeing the longer line on the review bar chart (at 480p resolution)
|
|
#
?
Dec 25, 2020 21:49
|
|
|
PC LOAD LETTER posted:No one who cares about gaming performance is going to care about short duration boost. They're going to be looking at sustained all core clocks. And if limited to 125W its not going to be that interesting, something like 4.2-4.3Ghz all core or so is what has been "leaked". I mean yes those things will happen. If you choose to crank it to 11 it will produce a lot of heat. None of that changes the fact that you *can* have a faster gaming chip at between 10-20w more power consumption, which is entirely possible to cool on a D15s or with 280mm of rad space. So no, to rebut your original point it's not a pyrrhic victory just because it's possible to cause insane power consumption by running the chip out of spec. Sustained 5.3 all core is absolutely an overclock.
|
|
#
?
Dec 25, 2020 21:51
|
|
|
VorpalFish posted:I mean yes those things will happen. If you choose to crank it to 11 it will produce a lot of heat. None of that changes the fact that you *can* have a faster gaming chip at between 10-20w more power consumption, which is entirely possible to cool on a D15s or with 280mm of rad space. So no, to rebut your original point it's not a pyrrhic victory just because it's possible to cause insane power consumption by running the chip out of spec. I’ve literally cooled a 10900K with the compact Noctua U9S, power consumption during gaming is so much lower than stress testing workloads. B-Mac fucked around with this message at 22:00 on Dec 25, 2020 |
|
#
?
Dec 25, 2020 21:55
|
|
|
B-Mac posted:I’ve literally cooled a 10900K with the compact Noctua U9S, power consumption during gaming is so much lower than stress testing workloads. Oh for sure - my bias is heavily towards very quiet computers so in general when I throw out a cooler I think can cool something I'm usually way on the overkill side as far as what can technically work. I don't like an audible difference between idle and load.
|
|
#
?
Dec 25, 2020 21:59
|
|
|
B-Mac posted:I’ve literally cooled a 10900K with the compact Noctua U9S, power consumption during gaming is so much lower than stress testing workloads. stress testing is usually Prime95, which is gonna be even more lol on these chips because of AVX-512. I dunno if it still has the dual-512b units like Skylake-X or if it's just a single 512b unit this time around but regardless, at gaming clocks the prime95 power consumption is gonna be lol (I would assume it's the same as Ice Lake but I can't find a reference there either) but yeah, in practice, the stress tests are nothing close to a realistic load, gaming power consumption is way lower. and the flip side is that everyone looks at the power consumption but not the throughput. Prime95 is a pure AVX load with basically no other code, so going to 512b units means you will get twice as much work out of it, and that actually improves efficiency compared to doing the same amount of work on AVX2 units, because the AVX2 chip would run longer for the same task. Zen3 is still gonna be way more efficient in general, but there are probably a few specific tasks that AVX-512 will let Intel catch up on, like perhaps x265 encoding. Paul MaudDib fucked around with this message at 22:07 on Dec 25, 2020 |
|
#
?
Dec 25, 2020 22:03
|
|
|
VorpalFish posted:Oh for sure - my bias is heavily towards very quiet computers so in general when I throw out a cooler I think can cool something I'm usually way on the overkill side as far as what can technically work. I don't like an audible difference between idle and load. I’m the same way, I like my PC to be inaudible while under load so I generally run everything at stock and then drop voltage as much as possible. Ryzen is certainly more efficient than Intel currently but it doesn’t pull 200W while gaming like some people seem to think it does.
|
|
#
?
Dec 25, 2020 22:06
|
|
|
Cygni posted:tbh i think both AMD and Intel have realized that gamers don't really care about power draw, they care about seeing the longer line on the review bar chart (at 480p resolution) where you run into problem is thermal throttling, so you can draw a lot of energy but that has to go somewhere and there's only so much you can wick away before then it starts rolling back going oops can't go that fast captain and now your longer line on the review bar goes down because oh hey it can't actually take care of the heat and now suddenly that performance isn't good anymore. we can actually see this with RTX 3080s where undervolting boosts performance because drawing less voltage means less heat which means better performance. can also see this with Zen3s using Curve Optimizer to undervolt cores to boost performance. so i think people do care about power draw, in only as much as the power draw also isn't just shutting itself down by drawing too much and leaving performance on the table.
|
|
#
?
Dec 25, 2020 22:07
|
|
|
Xaris posted:well yeah. ultimately going from 50w to 85w-150w (or intels supposed 250w) doesn't really mean it'll draw that much all the time esp even when gaming it'll prob use mostly much smaller percentage of that. and chips and windows are pretty smart about really minimal idle power when it's not in use too so it's really kinda moot thing to worry about overall for the most part. More power draw means more heat which means either more displacement, more airflow, or both - that definitely matters since more airflow usually means more noise and more displacement usually means a larger system. How much that matters and at what thresholds is going to be a matter of preference that varies substantially from person to person.
|
|
#
?
Dec 25, 2020 22:14
|
|
|
VorpalFish posted:Sustained 5.3 all core is absolutely an overclock. But that is what people are interested in and what is being pitched in the leaks irregardless of stock settings. That is where the hype is at and where the clocks are high enough vs Zen3 to maybe offer enough performance in game to be interesting. If you limit it to a power envelope similar to Zen3 its not at all a bad chip but it becomes more of a coin flip situation rather than the "head and shoulders" winner that the 5.3Ghz clocks are being hyped as.
|
|
#
?
Dec 25, 2020 23:02
|
|
|
PC LOAD LETTER posted:But that is what people are interested in and what is being pitched in the leaks irregardless of stock settings. That is where the hype is at and where the clocks are high enough vs Zen3 to maybe offer enough performance in game to be interesting. People are interested in either the best performance they can get for their use case that they can cool or the best value for their use case. Personally I don't give a poo poo if my clocks are 2ghz if they're somehow giving more consistent frame times. If the 18% IPC from ice lake / tiger lake tracks, rocket lake should be able to beat zen3 for gaming at stock clocks in a stock thermal envelope. At that point the question becomes what they price at. You can buy the 10700k at $320 but I'm guessing they're going to price higher since they're trying to reposition it as top of stack even at though it's 8c16t. Probably still a solid buy for high end gamers at $400, but I think they'll probably go higher.
|
|
#
?
Dec 25, 2020 23:44
|
|
|
VorpalFish posted:People are interested in either the best performance they can get for their use case that they can cool or the best value for their use case. Personally I don't give a poo poo if my clocks are 2ghz if they're somehow giving more consistent frame times. The lowest rocket lake sku is going to be some type of i5, right? So no successor to the underdog of the year, the i3 10100?
|
|
#
?
Dec 25, 2020 23:56
|
|
|
VorpalFish posted:You can buy the 10700k at $320 but I'm guessing they're going to price higher since they're trying to reposition it as top of stack even at though it's 8c16t. Probably still a solid buy for high end gamers at $400, but I think they'll probably go higher. At $400 it'd be $50 cheaper than the 8c/16t 5800X, so yeah, if it can beat Zen3 at gaming, still have 8c for productivity work, and equal or undercut the 5800X, it would be a pretty strong product. 150W isn't all that hard to cool, after all.
|
|
#
?
Dec 26, 2020 00:29
|
|
|
Twerk from Home posted:The lowest rocket lake sku is going to be some type of i5, right? So no successor to the underdog of the year, the i3 10100? I haven't seen any rumors. Hopefully they stop with the memory clock fuckery on the lower end skus. DrDork posted:At $400 it'd be $50 cheaper than the 8c/16t 5800X, so yeah, if it can beat Zen3 at gaming, still have 8c for productivity work, and equal or undercut the 5800X, it would be a pretty strong product. Yeah. The 5600x being as good as it is sets a ceiling on what they can charge imo if they have to market on gaming even though it's "only" 6c12t, current environment where you can't buy computer parts at all notwithstanding. VorpalFish fucked around with this message at 00:53 on Dec 26, 2020 |
|
#
?
Dec 26, 2020 00:50
|
|
|
Any thoughts on moving up from a 2400G to a Ryzen 3600 on a B350 board, versus getting an i5-10400F and a new H410/B460 board? The price difference means the Intel CPU + the board is going to come out to about the same total as just the AMD CPU by itself. My use-case is just purely gaming. I feel like the 3600 is still better as far as not losing out on memory tuning, but I think I just need someone to tell me outright.
|
|
#
?
Dec 26, 2020 02:40
|
|
|
gradenko_2000 posted:Any thoughts on moving up from a 2400G to a Ryzen 3600 on a B350 board, versus getting an i5-10400F and a new H410/B460 board? The price difference means the Intel CPU + the board is going to come out to about the same total as just the AMD CPU by itself. Yes, the Intel chip will underperform on a locked motherboard, though would be fine on a 470 iirc. What are you about to pay for a 3600, given that they've disappeared, and moreover what fps are you trying to hit?
|
|
#
?
Dec 26, 2020 02:42
|
|
|
3600s are back on Newegg for $199 right now fwiw https://www.newegg.com/amd-ryzen-5-...-569-_-12252020
|
|
#
?
Dec 26, 2020 02:50
|
|
|
Some Goon posted:Yes, the Intel chip will underperform on a locked motherboard, though would be fine on a 470 iirc. What are you about to pay for a 3600, given that they've disappeared, and moreover what fps are you trying to hit? the 3600 is priced at 230 USD (just doing a straight conversion from my currency) and the i5-10400F is priced at 173 USD. To be clear, I'm in Asia so prices are always elevated from MSRP. I'm trying to get to 75 FPS at 1440p. I recently upgraded to a 3060Ti so I know it's my CPU that's the slowest component in my build right now.
|
|
#
?
Dec 26, 2020 03:48
|
|
|
gradenko_2000 posted:the 3600 is priced at 230 USD (just doing a straight conversion from my currency) and the i5-10400F is priced at 173 USD. To be clear, I'm in Asia so prices are always elevated from MSRP. I don't know the 10400F's numbers with the slower memory off the top of my head but I'd be shocked it couldn't do 75fps in most titles. It's usually in the 100+ range where differences emerge. At the same price mise go with the ryzen since you've already got a fine board and it should have more headroom.
|
|
#
?
Dec 26, 2020 04:57
|
|
|
Indiana_Krom posted: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. Yeah, fossil fuels are fairly middle-of-the-road when it comes to energy density; as you said, it's just that they hit a decent balance of stability/transportability, but internal combustion is a fairly inefficient process - even under ideal conditions, you're not going to see it surpass 70%. This is part of why hydrogen is being pursued long-term: if you can utilize renewables and (ideally) nuclear power in combination to assist with the production of hydrogen, then you'll end up with a fuel that is substantially superior to any fossil fuel in terms of energy density. Anyway, to make this energy related, this is not something GloFo cares about because its owners prefer
|
|
#
?
Dec 26, 2020 05:34
|
|
|
Cygni posted:tbh i think both AMD and Intel have realized that gamers don't really care about power draw, they care about seeing the longer line on the review bar chart (at 480p resolution) History might not repeat itself, but it sure does loving rhyme.
|
|
#
?
Dec 26, 2020 07:37
|
|
|
im not a cpu genius but it also seems that "shoving more power at it" seems to be one of the only tools left in the belt anyway that might not need specific tens of billions of R&D we also spent a few years determining desktop computing was dead then the consoles became PCs (when lots of idiots thought it would be the other way around) and a lot of people starting computing full time at home again. also also the gains between these generations are basically useless for anyone BUT gamers and high end pros, who do only care about Number Up.
|
|
#
?
Dec 26, 2020 15:09
|
|
|
A 10400 might get a few more frames in games than a 3600 but in general computer usage I sincerely doubt I'd be able to tell the difference between them in a blind test. Most people will get a good experience with any computer above a certain level of performance.
|
|
#
?
Dec 26, 2020 17:32
|
|
|
bus hustler posted:the gains between these generations are basically useless for anyone BUT gamers and high end pros This has actually been true for a good while. Earlier this year I built a machine for someone who had been struggling with an ancient, third-hand, truly lovely laptop. After asking them about their use case (standard office productivity and the occasional Zoom call) I built it around a Ryzen 2200G -- but gave it 16G of DDR4-3000 RAM and a Crucial 500G NVMe SSD. No dGPU. They were thrilled, and I expect it's gonna be a few years before that machine starts to feel slow to them. Put your money where it's important for your use case. As an example on the flip side, my compute nodes use high core-count CPUs and fast RAM, but paired with the lowest-capacity, cheapest, no-name m.2 SSDs I can get my hands on. Those machines will never be I/O bound, and will only ever hold transient data. The drives cost about $30, and when one fails (which doesn't happen all that often because, again, they're not being stressed or doing tons of writes) I can reinstall the machine and have it back in service in around 20 minutes, including the hardware swap.
|
|
#
?
Dec 26, 2020 17:40
|
|
|
bus hustler posted:im not a cpu genius but it also seems that "shoving more power at it" seems to be one of the only tools left in the belt anyway that might not need specific tens of billions of R&DD You can only shove so much power at it before you can't cool it anymore especially with the transistor density and power density on modern cpus, cooling a 250w CPU is really hard. Push it too far and you're looking at chillers or ln2 for cooling which is.. questionable for regular use or maybe that concept being worked on with liquid cooling channels through the die itself.
|
|
#
?
Dec 26, 2020 17:59
|
|
|
FuturePastNow posted:A 10400 might get a few more frames in games than a 3600 but in general computer usage I sincerely doubt I'd be able to tell the difference between them in a blind test. Most people will get a good experience with any computer above a certain level of performance. Most people will honest to goodness report a computer at work feels faster if you take it away, maybe do basic maintenance on it, and put it back with a new KB/Mouse/Monitor with no other changes.
|
|
#
?
Dec 26, 2020 21:31
|
|
|
bus hustler posted:we also spent a few years determining desktop computing was dead then the consoles became PCs (when lots of idiots thought it would be the other way around) and a lot of people starting computing full time at home again. also also the gains between these generations are basically useless for anyone BUT gamers and high end pros, who do only care about Number Up. I take some small satisfaction from always feeling like the push to touch-screen interfaces and pretending that tablets were the future of computing was misguided. Work from home (and remote learning) has demonstrated the need for actual computers in the home. It's not like they were going away from workplaces anyway, but plague year helped quash the idea they were unnecessary at home. It's also unfortunately demonstrated the class inequity involved - lower income people are forced to rely on phones, both as a device and as a means of connecting the Internet, which just don't meet the needs for remote work or school. A decent modern laptop (which are ironically cheaper than higher end phones) and decent broadband should be considered a bare minimum for people going forward.
|
|
#
?
Dec 26, 2020 23:45
|
|
|
I just wish used computers were more of a thing like used cars, because on one hand you are absolutely right, on the other hand a perfectly serviceable computer for most needs is actually affordable for more people than ever before. One thing that got massively exposed IMO was how underpowered and stagnant laptop and even basic desktop graphics have become when a lot of perfectly fine, downright fast laptops still under warranty poo poo the bed with Zoom.
|
|
#
?
Dec 26, 2020 23:50
|
|
|
bus hustler posted:I just wish used computers were more of a thing like used cars, because on one hand you are absolutely right, on the other hand a perfectly serviceable computer for most needs is actually affordable for more people than ever before. Yeah, used computers should be more of a thing - I used to work for a Microsoft Authorized Refurbisher, and it was possible to get a decent machine through us. I know of similar organizations in other areas, but they are definitely not as well-known as they should be or utilized as much as they could be. I haven't run into the underpowered for Zoom thing but can believe it - I just did a fresh Linux install on my old netbook and I'm not sure it could handle it. My couple year old Chromebook handles Zoom fine, but it's got more serious hardware than the 2011 vintage netbook.
|
|
#
?
Dec 27, 2020 00:03
|
|
|
My Pixel Slate hates Zoom - it's very upsetting. An m3 8100 isn't setting any speed records, but it'd be awfully nice not to have everything chug as soon as it's running.
|
|
#
?
Dec 27, 2020 00:55
|
|
|
The problem with consumer laptops is that they are engineered to fall apart after like 3 years. The used market for Thinkpad or Latitude laptops is actually a thing though. You can totally replace keyboards or fans or whatever, they’re engineered to be repairable. The caveat is that you have to be willing to take older hardware with all that entails. Worse idle power, worse battery life, older processor with fewer cores (usually topping out at 4), etc.
|
|
#
?
Dec 27, 2020 01:00
|
|
|
Paul MaudDib posted:The problem with consumer laptops is that they are engineered to fall apart after like 3 years. Yeah, but used desktops are totally viable, especially since the capacitor plague ended. And used monitors, keyboards, mice, and what have you are plentiful. For basic use an old 60hz TN monitor is going to be fine for most people, and there are lots of them out there.
|
|
#
?
Dec 27, 2020 01:42
|
|
|
|
| # ? May 16, 2024 03:50 |
|
|
"Laptops, desktop sales see 'renaissance;' shortages won't ease until 2022" https://www.reuters.com/article/us-tech-hardware-yearend-idUSKBN28Y12M Keep on DIY-ing, nerds. Unless you wanted a Mac; Apple seems to have supply. Edit: Denver Micro Center had 5600Xs in stock all day yesterday, and most of today. I think that's the first time i've seen a 5000-series part be in stock for more than 24h. mdxi fucked around with this message at 01:54 on Dec 27, 2020 |
|
#
?
Dec 27, 2020 01:43
|
|