Good points. I've decided to give OpenGL a go, since I want to maintain portability to a variety of machines, including somewhat older/lower-end ones. For example, I may want to run small demos on RasPi type machines, rather than hauling big expensive machines to an exhibition.
With Vulkan out, is there any point in learning OpenGL right now? I'd like to accelerate my iterative art (see sig) with something like render to texture, and the higher-level tools don't seem too helpful for such tasks.
CPU fans always blow down onto the heat sink. Mounting them the other way allows a hot spot to form in the center as the fan tends to just pull air from around the heat sink instead of through it.
As opposed to the hot spot in the center of the blowing side? There's nothing blowing from the hub of the fan, and of course the highest speeds are on the outer edge. This doesn't yet account for idiotic heatsink designs that don't let air straight through, but instead divert it to the sides, thus leaving a high pressure center.
I agree with the general point though, it's almost always better to blow onto the heatsink. Try sucking out a candle, as our fluid mechanics professor used to challenge us.
The 1000x density and speed factors he's slobbering over are projected ceilings, which will take years to achieve in actual hardware.
Ah, just like the 10 GHz P4 that was just on the horizon, as the Networst architecture was designed for clock speeds rather than real-life performance and efficiency.
*sigh* you obviously missed my point: while "@" is pronounced "at", its meaning is roughly the same as that of "in". Why would you use "@" for a superficial wordplay when you have this wonderful chance to use its meaning to replace an entire element of the list ["requiescat", "in", "pace"], while keeping the other elements intact.
Is putting up with the headaches of windows really easier than learning something else?
Learning to use new operating systems doesn't seem too difficult when you consider things like tablets and "smart"phones. My parents' generation seems fine with those, but a computer running Linux seems totally alien to them -- like, how can you possibly write a document if it doesn't have Word? Of course, these are the same people who buy a new computer when Windows has crashed too hard. It's the appliance mentality.
I bought something at McDonald's and the total came to something and 35 cents. I gave the cashier the bills, a quarter and a dime. She insisted on giving me a nickel back as change. She insisted that a quarter and a dime added up to 40 cents. I finally took the nickel just to end the conversation.
This isn't necessarily a math problem per se, since you guys use arbitrary names instead of "10 cents" or "5 cents". The quarter is somewhat more tolerable, but even that might cause problems here. Remember those school problems like 1/4 + 1/10, wouldn't you rather do 0.25 + 0.10 when you're stressed out in work?
BTW, in Europe we don't even have a "quarter" denomination, the smaller coins are 50, 20, 10 and 5 cents for even easier math.
Looks like all the connectors are going away and everything is standardizing on USB Type-C connector. Good. Everything piggybacks over USB 3.1 including power. This looks like the best thing.
Ah, the "universal" serial bus. This will be great until USB 4 which is a completely new technology, but will have USB 2 and 3 pins bolted on to maintain backwards compatibility.
You do realize that there's no such thing as a pure digital display, right? LCD and Plasma are still all analog displays - they have digital display controllers driving an analog display surface.
This isn't quite true, IMHO -- you'll have to be more specific. The colour of an individual pixel still needs analog control, but with digital you can do exact pixel addressing. Contrast this to VGA where you're basically sweeping a spray can of a variable colour across the display. This is also why VGA monitors need adjustments to align this spray of colours with the display dimensions. No need for that with digital, where you always know which pixel you mean to light up.
This. I thought the whole idea of digital display connections was to make things bit-exact. Let's just go back to VGA and the fun of adjusting displays to the signal. Actually, let's go all the way back to analogue computers while we're at it.
I also found that to be true when I put together a system for my bedroom. And then I got a PicoPSU + transformer, only to find that the transformer had its own disagreeable little hum.
I went back to a traditional desktop and some fancy Noctua fans. They make noise I can hear but almost anything else I do is quiet enough for it to not be an issue.
There are full-size fanless PSUs that are much more robust than the PicoPSU+brick type combos. I've been very happy with my 460 W and 520 W models from Seasonic, no coil whine issues there.
I like to run my systems hot: cooling is more efficient the hotter you run. If it is specked for it, you might as well use it.
Specifications are not binary. Things will generally last longer if they don't run at the extreme of allowed specs all the time, due to effects such as electromigration.
In the second hand GPU market, a lot of people are always worried about overclocking history. What they should be asking about is temperatures. In overclocking, it's not the frequency that kills components, it's the heat. And it's easy to do heat damage while staying within allowed frequencies. Long-term heat damage is also an issue for other components around the main chip, such as VRMs on a GPU board.
I've been building silent, often fanless systems since 2003, so I'm always interested at these rare occasions when a commercial offering actually cares about noise. However, I'm somewhat worried about the peak temperatures of 80 C, and frankly it doesn't surprise me. Passive cooling is hard, and it's almost always better to aim for the low hum of large, slow fans. I'm running high end GPUs fully loaded all the time, and they stay around 50...60 C with aftermarket coolers (not water) in open cases, with 140 mm fans running at 7..10 V. The same goes for CPUs, though I'm not sure if they count as high end. Anyway, quiet and cool is easily done with aftermarket coolers that cost around 50 euros apiece. I live in a single-room apartment, so the lack of noise is pretty important.
I always wonder why proper cooling seems like an afterthought in components such as motherboards and cases, and why you always need aftermarket solutions if you don't want your machine to sound like a jet engine. For example, if the CPU were at the backside of the mobo, there would be no limit to the size of the heatsink. Yet the default is always a very crowded place in the middle of everything, where the "solution" is a small and whiny fan.
In all seriousness though... i guess there is such a thing as "times less" if taken literally rather than meaning fraction of...
Wow, imagine a world where words and phrases are taken literally. You could even say something like "one 10,000th of the power" to refer to the fraction, and actually be understood. But hey, language changes, let's just get over it and say jotain ihan vitun randomia, and let the reader/listener worry about the intended meaning.
I paid AMD/ATI money and they said "Ha, here's a shit ton of specs, write them yourself". Sorry. My job isn't to write display drivers, my job is to use the display drivers.
That's weird. I paid AMD and I got decent binary drivers. Not perfect, but decent for my purposes. I know they also provide some specs for open source devs, but so far I haven't found the open drivers good enough.
The savings come from smaller/faster hardware due to approximate carry propagation in the hashing cores. It has nothing to do with submitting false positives. False positives as easily weeded out by a regular CPU in less time than it takes to prepare a block and send it over the network.
So it's just how mining has been done since at least 2011, soon after we developed the first FPGA miners. Clock rates are gradually increased until a suitable error rate is reached, which depends on the individual chip and other factors.
I guess us Europeans have no need to whine, as we don't have such caps. Don't recall ever having one (I live in Finland). The terms of service generally include a right to throttle the connection in extreme situations, though it's hard to tell if this has ever happened to me. There are so many variables once the traffic is outside your ISP's network. We also understand the ISP may not have a total capacity equal to the sum of customer speeds. But this has little to do with monthly caps, unless you only think of data rates in monthly averages.
Slashdot Mirror
Good points. I've decided to give OpenGL a go, since I want to maintain portability to a variety of machines, including somewhat older/lower-end ones. For example, I may want to run small demos on RasPi type machines, rather than hauling big expensive machines to an exhibition.
With Vulkan out, is there any point in learning OpenGL right now? I'd like to accelerate my iterative art (see sig) with something like render to texture, and the higher-level tools don't seem too helpful for such tasks.
CPU fans always blow down onto the heat sink. Mounting them the other way allows a hot spot to form in the center as the fan tends to just pull air from around the heat sink instead of through it.
As opposed to the hot spot in the center of the blowing side? There's nothing blowing from the hub of the fan, and of course the highest speeds are on the outer edge. This doesn't yet account for idiotic heatsink designs that don't let air straight through, but instead divert it to the sides, thus leaving a high pressure center.
I agree with the general point though, it's almost always better to blow onto the heatsink. Try sucking out a candle, as our fluid mechanics professor used to challenge us.
The 1000x density and speed factors he's slobbering over are projected ceilings, which will take years to achieve in actual hardware.
Ah, just like the 10 GHz P4 that was just on the horizon, as the Networst architecture was designed for clock speeds rather than real-life performance and efficiency.
*sigh* you obviously missed my point: while "@" is pronounced "at", its meaning is roughly the same as that of "in". Why would you use "@" for a superficial wordplay when you have this wonderful chance to use its meaning to replace an entire element of the list ["requiescat", "in", "pace"], while keeping the other elements intact.
requiescat@pace
Fixed that for you.
Much bigger! Nuclear batteries. So they can fly even higher!
Higher, thus closer to the big yellow blob of a fusion power plant. I wonder if there's a technology to harness that...
Is putting up with the headaches of windows really easier than learning something else?
Learning to use new operating systems doesn't seem too difficult when you consider things like tablets and "smart"phones. My parents' generation seems fine with those, but a computer running Linux seems totally alien to them -- like, how can you possibly write a document if it doesn't have Word? Of course, these are the same people who buy a new computer when Windows has crashed too hard. It's the appliance mentality.
It's time to bury Windows in a deep, dark hole, remember it for all the good stuff it brought to computing
Ah! I remember the moment in the late 1990s when I buried Windows in a deep dark hole, and all the good stuff that moment brought to my computing.
AFAIK, current solutions for external GPUs are simply different electromechanical formats for PCI Express, so this shouldn't be an issue.
I bought something at McDonald's and the total came to something and 35 cents. I gave the cashier the bills, a quarter and a dime. She insisted on giving me a nickel back as change. She insisted that a quarter and a dime added up to 40 cents. I finally took the nickel just to end the conversation.
This isn't necessarily a math problem per se, since you guys use arbitrary names instead of "10 cents" or "5 cents". The quarter is somewhat more tolerable, but even that might cause problems here. Remember those school problems like 1/4 + 1/10, wouldn't you rather do 0.25 + 0.10 when you're stressed out in work?
BTW, in Europe we don't even have a "quarter" denomination, the smaller coins are 50, 20, 10 and 5 cents for even easier math.
Looks like all the connectors are going away and everything is standardizing on USB Type-C connector. Good. Everything piggybacks over USB 3.1 including power. This looks like the best thing.
Ah, the "universal" serial bus. This will be great until USB 4 which is a completely new technology, but will have USB 2 and 3 pins bolted on to maintain backwards compatibility.
You do realize that there's no such thing as a pure digital display, right? LCD and Plasma are still all analog displays - they have digital display controllers driving an analog display surface.
This isn't quite true, IMHO -- you'll have to be more specific. The colour of an individual pixel still needs analog control, but with digital you can do exact pixel addressing. Contrast this to VGA where you're basically sweeping a spray can of a variable colour across the display. This is also why VGA monitors need adjustments to align this spray of colours with the display dimensions. No need for that with digital, where you always know which pixel you mean to light up.
http://iki.fi/teknohog/rants/v...
This. I thought the whole idea of digital display connections was to make things bit-exact. Let's just go back to VGA and the fun of adjusting displays to the signal. Actually, let's go all the way back to analogue computers while we're at it.
I also found that to be true when I put together a system for my bedroom. And then I got a PicoPSU + transformer, only to find that the transformer had its own disagreeable little hum.
I went back to a traditional desktop and some fancy Noctua fans. They make noise I can hear but almost anything else I do is quiet enough for it to not be an issue.
There are full-size fanless PSUs that are much more robust than the PicoPSU+brick type combos. I've been very happy with my 460 W and 520 W models from Seasonic, no coil whine issues there.
I like to run my systems hot: cooling is more efficient the hotter you run. If it is specked for it, you might as well use it.
Specifications are not binary. Things will generally last longer if they don't run at the extreme of allowed specs all the time, due to effects such as electromigration.
In the second hand GPU market, a lot of people are always worried about overclocking history. What they should be asking about is temperatures. In overclocking, it's not the frequency that kills components, it's the heat. And it's easy to do heat damage while staying within allowed frequencies. Long-term heat damage is also an issue for other components around the main chip, such as VRMs on a GPU board.
I've been building silent, often fanless systems since 2003, so I'm always interested at these rare occasions when a commercial offering actually cares about noise. However, I'm somewhat worried about the peak temperatures of 80 C, and frankly it doesn't surprise me. Passive cooling is hard, and it's almost always better to aim for the low hum of large, slow fans. I'm running high end GPUs fully loaded all the time, and they stay around 50...60 C with aftermarket coolers (not water) in open cases, with 140 mm fans running at 7..10 V. The same goes for CPUs, though I'm not sure if they count as high end. Anyway, quiet and cool is easily done with aftermarket coolers that cost around 50 euros apiece. I live in a single-room apartment, so the lack of noise is pretty important.
I always wonder why proper cooling seems like an afterthought in components such as motherboards and cases, and why you always need aftermarket solutions if you don't want your machine to sound like a jet engine. For example, if the CPU were at the backside of the mobo, there would be no limit to the size of the heatsink. Yet the default is always a very crowded place in the middle of everything, where the "solution" is a small and whiny fan.
They're in your car, and it's moving at exactly 88 mph. Not even Google can beat physics.
In all seriousness though... i guess there is such a thing as "times less" if taken literally rather than meaning fraction of...
Wow, imagine a world where words and phrases are taken literally. You could even say something like "one 10,000th of the power" to refer to the fraction, and actually be understood. But hey, language changes, let's just get over it and say jotain ihan vitun randomia, and let the reader/listener worry about the intended meaning.
In the future, we might even use silicon-based components in data transmission. But for now, it's pure Si-Fi.
Also: Now that we have used "bandwidth" to describe bit rate, what are we going to call the old scientific/technical concept of "bandwidth"?
This. How do you get corrupted data from bad networking hardware into a virtual machine, without it passing through a real NIC first?
I paid AMD/ATI money and they said "Ha, here's a shit ton of specs, write them yourself". Sorry. My job isn't to write display drivers, my job is to use the display drivers.
That's weird. I paid AMD and I got decent binary drivers. Not perfect, but decent for my purposes. I know they also provide some specs for open source devs, but so far I haven't found the open drivers good enough.
The savings come from smaller/faster hardware due to approximate carry propagation in the hashing cores. It has nothing to do with submitting false positives. False positives as easily weeded out by a regular CPU in less time than it takes to prepare a block and send it over the network.
So it's just how mining has been done since at least 2011, soon after we developed the first FPGA miners. Clock rates are gradually increased until a suitable error rate is reached, which depends on the individual chip and other factors.
I guess us Europeans have no need to whine, as we don't have such caps. Don't recall ever having one (I live in Finland). The terms of service generally include a right to throttle the connection in extreme situations, though it's hard to tell if this has ever happened to me. There are so many variables once the traffic is outside your ISP's network. We also understand the ISP may not have a total capacity equal to the sum of customer speeds. But this has little to do with monthly caps, unless you only think of data rates in monthly averages.