I have wondered about folding@home as well, another distributed project begun in another age that seems not, from the outside, to have lived up to its early hype. In the meantime, it seems that Alphafold has surpassed conventional efforts at understanding protein folding, including folding@home's. https://moalquraishi.wordpress.com/2018/12/09/alphafold-casp...
The two approaches aren't directly comparable. Alphafold predicts the optimum structure for a sequence, while folding@home explores protein dynamics (how a protein's structure changes over time).
Anectdata: I set up an old computer with this in 2009, but a few months later I heard about this thing called cryptocurrency mining...
EDIT: The article notes that they are reaching a point of diminishing returns with the distributed nature of the SETI@HOME tasks, but I can't help wondering if things like crypto had an impact on participation with this and other @HOME-style academic projects[0].
I had the opposite problem. Knowing the importance of offsite backups I had the brilliant idea of backing up my wallet.dat on Dropbox without any additional encryption. My bitcoin were only worth a few bucks, who'd bother stealing them, right? Years later, I saw that BTC had gone up in price a few orders of magnitude so I tried to load my wallet.dat and cash out. Except when I opened my wallet file, the balance was zeroed out. Oops: https://www.troyhunt.com/the-dropbox-hack-is-real/
I looked at the wallet address, which looked liked it was used to slurp up BTC from a number of different sources. Didn't bother tracing it any further than that because I was so disgusted with myself for losing (potentially) thousands of dollars. Good news is that I signed up for Coinbase when they were giving out 0.1 BTC signup bonuses, so I still got to ride the wave with that. Made me feel better about writing off the BTC I mined the hard way on my Macbook Pro— better to think of it as an expensive way to learn about the importance of good passwords and 2 factor authentication :)
Nah, my guess is that whoever hacked Dropbox (or someone who bought the info from the person who hacked it) had a script to brute force passwords, then log into accounts and scan for wallet.dat files.
I'm not an expert in either this or cryptocurrency, but it always seemed like these two would be a good marriage of technologies rather than having crypto mining rigs waste resources on unproductive proof of work algorithms. Does anyone know why that wasn't feasible or didn't catch on?
Proof of work relies on computations that are hard to do but easy to check that someone has done. The entire network has to check someone's proposed result. There are very few computations that have sufficient asymmetry to the degree needed to make bitcoin work, and the types of things that SETI@home needs don't fit that bill.
Isn't Folding@Home trying to run an asymmetrical computation, though? The broad strokes as I understand them are that you have a sequence of amines, you arrange them into a three-dimensional protein (the hard part) by whatever method, and then you evaluate the energy bound within your structure (the easier part?). Lower energy is better.
The obvious problem is that there's no judgment of "correct" or "incorrect". Analogizing to a bitcoin proof-of-work, it would be as if a hash beginning with more zeroes always outranked a hash beginning with fewer zeroes, but there was no way to know whether your hash started with "enough" zeroes or not.
RE: folding@home, How much money will the medical cartel make off of the fruits of all that free labor? Do users even know what they're folding? How do users know that they're not building bioweapons?
For the most part this is fundamental research of the type that is typically funded from public grants, these are not pharma companies... As for what they are contributing too, it's all pretty well documented.
protein folding simulations don't produce a lot of money for pharma (I say that as somebody with a F@H publication to my name specifically doing research on GPCRs, which are a huge pharma target). users can see what the workunits are doing, and choose subprojects, presumably avoiding bioweapons if that's important to them.
How could the average technically savvy user, let's say willing to use Google a bit but not to obtain a biochemistry degree, tell which subprojects are possible bioweapons? Not a rhetorical question.
I'm going to guess that if you inspect each project it has a purported (positive) medical use and you'd have to make a judgement call on each project based on good knowledge of biology as to whether it had a bioweapons application.
I doubt, but cannot say for certain that F@H won't have any projects that are stated as bioweapons and I'm skeptical that somebody working on a bioweapon would try to cloak its intent, and come up with a plausible medical application so they could get free F@H cycles.
I remember the days of logging into national labs supercomputers and seeing the other jobs on the system, such as job_name "submarine_geometry_optimizer" and the user_name was "electric_boat" or something similar.
You could incorporate that in a quorum blockchain like Stellar (or Ripple I guess), but the whole point of those is that they remove the need for PoW so that's a bit useless. Maybe it's a cool way to build trust in such a network though..
Protein folding might work as a proof-of-work algorithm. One big problem is of course that it's useful work that could conceivably be monetized.
Proof-of-work works because a theoretical attacker would have to invest huge resources into an attack. For bitcoin it's trivial to say "for attack scenario X an attack has to spend at least Y", and as long as the expected reward is lower than Y the network is safe (that's for example one factor you use to determine how many confirmations you wait before accepting a payment). If the attacker could conceivably gain money outside the cryptocurrency for performing the proof-of-work, that makes the whole calculation less predictable, and likely exposes you to more situations where an attack is actually profitable.
What's needed is a large delta between the value of the work and the value of the currency. It's not a requirement that it be completely wasted computation, that's just how it happens to work right now.
Protein folding, where the value is speculative, cumulative, and difficult to cash out, is an excellent candidate. Especially since the ability to realize the value of a given fold is diminished by openly publishing it, which is a requirement of proof of work; the folding computation thus becomes a positive externality of the cryptocurrency, since everyone can use that knowledge as an economic stimulus.
An attack on proof-of-work wouldn't let you rewrite arbitrary transactions. At most an attacker could exclude certain recent transactions—most likely their own—reverting payments which were previously deemed "complete" and replacing them with conflicting transactions. All transactions still need to be signed with the proper keys, however, so the attacker can't just lay claim to others' coins.
Golem (www.golem.network) is building a general-purpose distributed computation platform where people who need compute power pay using cryptocurrency. And developers can build software to run on it (and get paid when it's used).
I vouched for your comment, as it's factually correct near as I can tell. I've been interested in the project, especially since Joanna Rutkowska joined them.
There were plenty, basically none of these savants understood economics so they didn't build a demand model, they only built a supply model. So people got paid in coin and just flooded the market until it was worth nothing.
Crypto mining rigs are proving far more productive than Seti or Folding@Home ever did. Economic pressures push the real energy heavy miners into places with renewable energy and also repurposing waste byproducts into electricity, so this is actually one of the most useful and cleanest sector on the planet, ironically the opposite of what people are saying. Protip: Headlines only tell you "how much" energy is being used and not the source, and this is very intentional.
I was in a similar boat to you, except I set up my new i7 rig (at the time) for SETI@HOME because I couldn't get mining to work properly, or it kept crashing!
Remnants of my attempts are at home on my PC somewhere, there are no wallets from my searches of my drives :-(.
I can remember my high school science teacher sometime between 2001-2002 telling us about this project and you'd see this running on his computer any time he wasn't using it.
I always wondered what came out of it and how cool it felt to feel like you could, as an individual, help support something massive by doing something so small.
>>I always wondered what came out of it and how cool it felt to feel like you could, as an individual, help support something massive by doing something so small.
You can still participate in Folding@Home, which in my opinion is a far more worthy pursuit - actual medical science came out of it and more is still being produced.
In case anyone reads the headline and thinks that the SETI research center itself is shutting down, or that the project failed in some way:
> "It's a lot of work for us to manage the distributed processing of data. We need to focus on completing the back-end analysis of the results we already have, and writing this up in a scientific journal paper," their news announcement stated.
Maybe it's PR speak, but it does make sense to me that there'd be a point of diminishing returns where enough data has been processed.
The singular in "journal paper" makes me sad. Surely there'd be more papers in this topic, on distributed computing for one, on the methodology of the analysis for another, and on any results for third.
I doubt cloud computing fits for their use case. They need constant power, and using "cloud" would be much more expensive than to have their own hardware (and of course even more expensive than using someone's else's computer).
The reason why SETI@home is not popular anymore is because how our hardware adapted. Back in the days the CPUs were constantly operating at the same speed. So if you left your computer on, it was using the same amount of power and generated the same amount of heat. SETI@home took advantage of it, by saying, let's make use of that wasted energy to do something useful.
As computers became faster the heat and power consumption became a bigger issue, mechanisms were added to allow CPU scale its frequency down, and have "micro sleeps" during the time there's nothing to do. OSes similarly adapted. Instead of having an interrupt triggered at regular interval they implemented so called tickless mechanism, when those house keeping operations are fired when needed instead of regular intervals.
All those advancements helped a lot with portable computers (laptops, pads, phones etc) but also made it to the desktops and servers.
Because of those advancements it is no longer "free" to run SETI@home. If you run it, your computer will get warmer, it will turn extra fans too cool down and it will use more power, so people are less motivated to run it. That's what made me stop using SETI@home myself.
You're right, that might be the real reason. But did they report a decline in users or cpu hours?
What I thought of was more like Amazon, Google or Microsoft "donating" some idle instances, or something like that.
Well, I don't think they reported it, but back in 2000 a lot of my friends were running SETI@home, today I don't know a single person that does it. Frankly I thought it was shutdown much earlier.
As for Amazon, Google, Microsoft donating, the same thing applies, it is no longer free unused power, it actually cost money to do it. It's weird to think they would just let use their resources that way when they also work on optimizing their power usage to the point of having system that automatically shuts down idle computers.
If they want to donate to a cause, it's more efficient for them to give them money, or at very least provide service with some discounts. But IMO if they were doing it, they would not do it quietly, so I doubt public cloud providers contribute anything to SETI.
I had the brilliant experience of visiting the Allen Telescope Array and meeting Jill and her husband before the announcement of their $100MM donation a few years ago. I have to say that it is an incredibly lean operation ran by the smartest people I've ever met. The @HOME project was more about being able to help handle throughput they would've otherwise just had to dump. I think it makes sense to sit back and spend time reevaluating what's been collected such that the filters can be evolved into something more efficient. I don't think they've spent time seriously trying to apply AI, yet. I'd be eager to hear their opinions on the subject.
I worked in a debt collection center where all the call center people used Linux workstations running wy60 terminals and a custom check viewer I wrote in Java (had to reverse engineer our vendors that I did get working in Wine, but it crashed a lot).
Installations and updates were automated. After getting the company logo and xplanet in the background, I got permission to put SETI@Home and FOLDING@Home on the machines (I think I did half and half). After all the machines were barely doing anything other than terminals and rendering TIFFs. Our company name was listed as the team.
Looks like the stats are off line for SETI@Home teams .. wonder how much that cluster kept contributing after I left.
Honestly, I never cared much about SETI@home, always felt like an empty waste of resources. But folding@home or even just training LeelaChess by community really feels like a noble cause I am willing to spend some electricity for.
Ah, I used to fold. Sometimes I regret that I'm too young to hve enjoyed the early 'glory days' of personal computing, but thinking about it now, buying CustomPC magazine and finding inside it the team's folding@home league table will sound pretty weird one day, if it doesn't already.
Seems a bit misleading, they're not distributing more work units, not closing the research center. Nonetheless, this is sad news to me. My entire career path in the tech world was because I got a campus job working for SETI@home.
I remember being a younger pup and building fleets of machines (including overclocking, etc) for the sole purpose of running S@H, and all of the amazing software the spun up around it... SetiSpy, SetiDriver, SetiQ. I often wonder how machines of today would compare in work production to my old Duron 800, but alas.
it was cheaper for AMD to produce one chip and then selectively cut the multiplier lines to differentiate sku for different market segments, so if you closed the lines again you'd get a different, faster clock multiplier
Hold your CPU in your left hand and look very closely at the CPU, so as to clearly see the L1 bridges.
Use a business card to separate the bridges so that you do not connect the L1 bridges to each other. Work your way across the bridges from left to right (using a business card as a separation tool) and connect the bridges by rubbing the pencil back and forth over the bridges about twenty times until it is dark black, not the normal gold color. Make sure that all the bridges are reconnected, but not touching each other, and you are on your way.
I know it sounds incredible, but that is actually all there is to it. Your processor, if done correctly, is ready to be overclocked. It can now be set to run at different clock frequencies, eliminating the need to increase the FSB."
I had never heard of this trick. I love it though. Reminds me of using a hole punch to make a double sided floppy disk from a cheaper priced single sided floppy.[0]
Of course there was the AMD multi-core CPUs that had some of the cores disabled via software and sold at cheaper prices. Of course there were software hacks available to re-enable those cores.
Oh, yes, the Phenom II 550 Black Edition. The early ones like the one I still have stashed away were an unsung bargain. My motherboard has a BIOS option to unlock extra cores, and so I got four cores for the price of two; at the stock clock speed all cores were rock-solid stable. I never did find any unbuffered 4GB ECC RAM sticks, so I had to settle for 4x1GB.
With 16 GB and a SATA3 controller or PCIE-to-NVMe adapter it would still be a respectable performer today, in spite of the DDR2.
The Duron had a set of circuit traces exposed on the ceramic chip carrier that were cut by laser at the factory to set the maximum allowed clock multiplier (and thus its final speed). You could reconnect all the cut traces with pencil lines and use a motherboard that let you set the multiplier directly to overclock freely.
Yes, I owned two of those - alas they ended up being the period of bad capacitors but for the money - wow. Was amazing and for many, one of the only dual cpu systems they ever owned as a home user - even to this day unless you buy a cheap second hand workstations/server platform of ebay or the like.
Indeed they were. I had that exact setup. Overclocked the hell out of them. Unlike the 550 MHz someone else suggested, I'm 97.6% sure I only had them clocked at 450 MHz.
I built a bunch of them and still have mine in the attic. But alas your probably like me and by the time I actually purchased one for myself the overclocking abilities were much less than the original ones which went from 300->500Mhz fairly easily. IIRC its because Intel was selling 400 and 450Mhz ones and binning all the good ones into the upper tier bins.
Dual celery story:
Me being around 14, not allowed to build my own PC, but I got to spec it, I ordered dual CPU's.
The company put in a dual slot 1 motherboard, and then two celerons in slot1 adapters.. that didn't support SMP.
After complaining for a while, they traded the two celerons and adapters for a single P3 (I think I was being had). It took a year or two before I was able to afford the second P3.
I still hold a grudge towards that particular computer shop, or I would, hadn't they gone out of biz.
I did run a lot of SETI@home on that P3, it earned me my first two certificates :D
I had a BP6 and remember 450 too (I may still have it tucked away somewhere). I think there was a 333 that could be pushed to 550 or something but not in SMP and I think it was slot 1 and not socket 370.
+1 here who had the Abit BP6 dual celery setup. Had a pair of 400s running at close to 500 all the time. Both were lapped down to copper running FEP32 global win coolers. The wonders of Windows 2000 gaming machine with the GeForce 256 which was quickly followed by a GeForce 2 GTS upgrade as soon as it game out. Running quake on one CPU and no stutter MP3s on priority on the other. Good times.
yeah it was something like, you ran a lower clock speed but a higher multiplier? I've been out of the hardware game... well, ever since then. But that's my recollection. 450 vs 464 was the first place you were likely to see stability issues depending on your specific setup, but most were stable to at least 464.
I still have 2 of these in my old hardware box. I used to run them at 4.5x100MHz on Abit BH6 single CPU mobo. Great way for me to bump performance as a poor CompSci student. And yeah, I ran SETI@home on them as well.
Hahah, I can relate so much with the 'poorerer' part, I'll raise you another one: my first '686' was a Cyrix 6x86 because those were the cheapest you could get. Eventually got my hands on a Pentium II and had to scavenge parts to replace my setup.
Good times. Nowadays I just buy a laptop with everything soldered on it and complain a lot if I have to reboot it every month.
Waay back in the day.. 286 12mhz with 1mb ram and a 10mb RLL drive. Setting the jumpers on the drive controller block was a nightmare that I still have. I think there were 16 or 32 pins and all the jumpers had to be just right or no dice (no drive). Further along I had a 40mhz Cyrix 4x86 I think, with 4mb ram and a Trident VGA 256 color card. I remember the ram being old school block chips and not SIMM modules. It's been a long time, so I could be mistaking the memory type between mobos. Was a great 3.11 machine for the time.
Seems to have been a fairly common trick with the Durons. I sadly never experienced the rush of trying it mostly because Athlons and PIIIs were all I'd ever played with when the 1ghz chips started coming out, but this was probably what made the Durons so attractive to student hackers. https://arstechnica.com/civis/viewtopic.php?f=8&t=840954
Tangent: I love that I can pull up forum threads from two decades ago.
Ah! Athlons.
My overclocking journey picked up when I built a DIY watercooler out of a bent 5mm cooper sheet (at which thickness you stop calling it "sheet"?) polished to a mirror finish.
At some point, during the peak of easter european winter I kept the heat exchanger outside in the -20*C temperatures!
I remember doing this too, it unlocked the multiplier which they locked at the factory by slicing through some metal blocks with a laser. My Duron 600 would run at 800 with a normal cooler & fan, then I added a really loud Delta fan and a copper heatsink and it was stable at a magical 1Ghz. Eventually I got cocky enough to solder a resistor between a certain pin on the motherboard and earth which effectively increased all the voltage settings in the BIOS so I could push enough voltage through it to run at 1.2Ghz - it ran like this for years as my only PC. Good times :)
I had a K6-233 which was always on the verge of overheating (or maybe I was obsessing a bit too much about monitoring it). I envied my friend's K6-200, which was more stable it seemed.
My K6-233 got so hot it demagnetized its CPU fan. I noticed after the floating point unit fried the first time I tried to play some post-quake FPS.
After that, it would boot, but the antivirus software I was running said it had a virus. I guess that was the only msdos program in the boot sequence that performed floating point.
I had a K6-II which I put a big heatsink and large fan on (had to ghetto mod the mounting), sawed a 120mm hole in the case just above the cpu and put another fan there. It overclocked handsomely and stable after that. I don't remember the exact frequencies but I think it was 266MHz stock and clocked close to 400MHz with some modifications.
I was right along with you when I was helping my parents pick out their next PC purchase, a p3 450, and from then on just consumed every single cpu article and review I could find, so much knowledge that I picked up about the computer industry in general has helped me later in my career.
I worked for a large insurance company, my first job, when SETI@Home was launched. It was a typical large cubicle farm. I remember sneaking around to any unused WindowsNT workstation around me and installing it. Good times, good times.
When I first got my i7 6-core with quad GTX-690s (2 GPUs per card), I naively let Seti@Home run rampant when the computer was idle. Room they were in got up to around 90F when I let it run 100%.
I suspect in the very early days of cooperative computing, one could genuinely argue that there were "idle cycles" that were going to waste if a computer was powered on and not computing something. But this became more and more inaccurate throughout the 1990s as new power management technologies became a part of computer designs, and is completely wrong now. Phenomena like Bitcoin mining and feeling your laptop get warmer depending on what it's doing show that easily, as does your room-heating experience.
I'm still curious about the extent to which the claim of idle compute power going to waste was true when various projects were set up. For example, the press release about the Cooperative Computing Awards (I've forgotten whether this was 1998 or 1999) says
> However, the computer industry produces millions of new computers each year, which sit idle much of the time, running screen savers or waiting for the user to do something.
> [the CPU-idle instruction] was not specifically designed to reduce power consumption until the release of the Intel DX4 processor in 1994.
I participated in DESChall and distributed.net's RC5 challenge and I remember being told, and telling others, that we were only using resources that would otherwise have gone to waste. Were we already greatly mistaken about that in 1997?
I would think of compute resources "going to waste" in three separate groups:
- The baseline electricity the computer is using. This will get used and produce no S@H work either way (in an era where our computers didn't go to sleep)
- The added electricity to do S@H work. This certainly was not idle, but also was a relatively small amount relative to an entire household's use.
- The hardware itself.
I would say that S@H were making use of hardware that was "going to waste" (this is CAPEX the project didn't need to spend), was having volunteers donate the added electricity, and could be said to be making use of the baseline electricity inasmuch as it would be used either way, and if S@H were to set up their own compute cluster, they would also incur similar overhead.
I don't know what system efficiencies were like in the late 90's, but I would totally believe someone if they said "idle" was 40% the power consumption of "max usage". Using that number as an example, "donated electricity" gets (1/0.6=) 1.6x as much compute-per-watt as if it was donated to run an equivalent, dedicated, machine. Maybe you can convince yourself that you were using "idle cycles" and donating on top of that?
That's really part of what I was asking about. The claim from Wikipedia about Intel's HLT instruction shows that this started to change a lot by 1994, but I don't know if there are other similar considerations.
Before 2002 (AMD's Cool'n'Quiet) and 2005 (Intel's SpeedStep) CPUs didn't do any frequency scaling for power savings. When the CPU wasn't asleep it was running at it's full voltage and clock speed. While the HLT instruction kept let them 'idle' they only used marginally less power than when active.
So running distributed.net or Seti@Home was using unused processing capacity for a marginal increase in power draw. If you ran them overnight with your monitor (CRT typically) your computer was drawing less power than if you were sitting up all night on IRC or playing Quake. This changed once frequency scaling was commonplace on desktops.
I'm surprised it lasted so long, the idle computer donation made way more sense back when processors didn't really change power consumption or downclock.
I ran it before I went to school in the morning and before I went to bed at night for a few days. Then the novelty wore off. Also, it interfered with my downloads. I also remember I watched Contact on tv around that time and it got me hyped about doing my part to discover extraterrestial life.
Did SETI@home accomplish anything useful with its 21 years of computing? Outside of the general "it's always good to do science, even with negative results" which I agree with.
I ran a large cluster of university machines surreptitiously for this project many years ago.
So, is this all they did in 21 years?
Funny people don't criticize it for wasting electricity when people complain about how Bitcoin wastes so much which seems like a far more practical utility than SETI.
Ah, good memories of being temporarily banned in the EE labs at UT Austin for running set@home on a bunch of the unix boxes. Must have been around 2000.
I got chewed out at my workplace around 1999 or so. We had a couple hundred SUN workstations on the network and anyone could run commands remotely on any workstation. So I set up a cron job to spawn SETI@home jobs on every workstation every night at 9pm, shutting them down at 6am the next morning.
I was way up there in the rankings for a while, until the sysadmins started investigating the strange traffic patterns...
Prime95 helped me weed out a flaky power supply. The system would randomly freeze. The cursor would just stop moving, along with everything else. Otherwise the machine was working perfectly.
I figured it that Prime95 triggered the freeze within 30 seconds of staring the torture test. Then it was just a matter of swapping out the ram and then the power supply. The new power supply let the computer go on running the test.
IMO there isn't much of one, at least not enough for me to feel motivated to run Prime95 (the GIMPS client), but others may find the justifications here more convincing:
Man alive. I first ran some SETI@home stuff on my gaming rig in 2001. After a year or two I figured it was wasted energy. Surely they had more efficient rigs helping out than my lowly Unreal 2004 machine, right?
355 comments
[ 3.6 ms ] story [ 382 ms ] threadEDIT: The article notes that they are reaching a point of diminishing returns with the distributed nature of the SETI@HOME tasks, but I can't help wondering if things like crypto had an impact on participation with this and other @HOME-style academic projects[0].
[0]: https://boinc.berkeley.edu/projects.php
The obvious problem is that there's no judgment of "correct" or "incorrect". Analogizing to a bitcoin proof-of-work, it would be as if a hash beginning with more zeroes always outranked a hash beginning with fewer zeroes, but there was no way to know whether your hash started with "enough" zeroes or not.
I'm going to guess that if you inspect each project it has a purported (positive) medical use and you'd have to make a judgement call on each project based on good knowledge of biology as to whether it had a bioweapons application.
I doubt, but cannot say for certain that F@H won't have any projects that are stated as bioweapons and I'm skeptical that somebody working on a bioweapon would try to cloak its intent, and come up with a plausible medical application so they could get free F@H cycles.
I remember the days of logging into national labs supercomputers and seeing the other jobs on the system, such as job_name "submarine_geometry_optimizer" and the user_name was "electric_boat" or something similar.
Proof-of-work works because a theoretical attacker would have to invest huge resources into an attack. For bitcoin it's trivial to say "for attack scenario X an attack has to spend at least Y", and as long as the expected reward is lower than Y the network is safe (that's for example one factor you use to determine how many confirmations you wait before accepting a payment). If the attacker could conceivably gain money outside the cryptocurrency for performing the proof-of-work, that makes the whole calculation less predictable, and likely exposes you to more situations where an attack is actually profitable.
Protein folding, where the value is speculative, cumulative, and difficult to cash out, is an excellent candidate. Especially since the ability to realize the value of a given fold is diminished by openly publishing it, which is a requirement of proof of work; the folding computation thus becomes a positive externality of the cryptocurrency, since everyone can use that knowledge as an economic stimulus.
I'd like to see someone try it.
Crypto mining rigs are proving far more productive than Seti or Folding@Home ever did. Economic pressures push the real energy heavy miners into places with renewable energy and also repurposing waste byproducts into electricity, so this is actually one of the most useful and cleanest sector on the planet, ironically the opposite of what people are saying. Protip: Headlines only tell you "how much" energy is being used and not the source, and this is very intentional.
Remnants of my attempts are at home on my PC somewhere, there are no wallets from my searches of my drives :-(.
I always wondered what came out of it and how cool it felt to feel like you could, as an individual, help support something massive by doing something so small.
You can still participate in Folding@Home, which in my opinion is a far more worthy pursuit - actual medical science came out of it and more is still being produced.
> "It's a lot of work for us to manage the distributed processing of data. We need to focus on completing the back-end analysis of the results we already have, and writing this up in a scientific journal paper," their news announcement stated.
Maybe it's PR speak, but it does make sense to me that there'd be a point of diminishing returns where enough data has been processed.
https://link.springer.com/article/10.1007/s10723-019-09497-9
Well... We haven't found anyone we can talk to yet so, in a sense, we are still failing.
Eventually we will.
The reason why SETI@home is not popular anymore is because how our hardware adapted. Back in the days the CPUs were constantly operating at the same speed. So if you left your computer on, it was using the same amount of power and generated the same amount of heat. SETI@home took advantage of it, by saying, let's make use of that wasted energy to do something useful.
As computers became faster the heat and power consumption became a bigger issue, mechanisms were added to allow CPU scale its frequency down, and have "micro sleeps" during the time there's nothing to do. OSes similarly adapted. Instead of having an interrupt triggered at regular interval they implemented so called tickless mechanism, when those house keeping operations are fired when needed instead of regular intervals.
All those advancements helped a lot with portable computers (laptops, pads, phones etc) but also made it to the desktops and servers.
Because of those advancements it is no longer "free" to run SETI@home. If you run it, your computer will get warmer, it will turn extra fans too cool down and it will use more power, so people are less motivated to run it. That's what made me stop using SETI@home myself.
As for Amazon, Google, Microsoft donating, the same thing applies, it is no longer free unused power, it actually cost money to do it. It's weird to think they would just let use their resources that way when they also work on optimizing their power usage to the point of having system that automatically shuts down idle computers.
If they want to donate to a cause, it's more efficient for them to give them money, or at very least provide service with some discounts. But IMO if they were doing it, they would not do it quietly, so I doubt public cloud providers contribute anything to SETI.
Installations and updates were automated. After getting the company logo and xplanet in the background, I got permission to put SETI@Home and FOLDING@Home on the machines (I think I did half and half). After all the machines were barely doing anything other than terminals and rendering TIFFs. Our company name was listed as the team.
Looks like the stats are off line for SETI@Home teams .. wonder how much that cluster kept contributing after I left.
... no
Now there's a name I haven't heard in a while. Duron, the poorerer man's Celeron.
"Unlocking the Duron and Athlon Using [a] Pencil
Hold your CPU in your left hand and look very closely at the CPU, so as to clearly see the L1 bridges.
Use a business card to separate the bridges so that you do not connect the L1 bridges to each other. Work your way across the bridges from left to right (using a business card as a separation tool) and connect the bridges by rubbing the pencil back and forth over the bridges about twenty times until it is dark black, not the normal gold color. Make sure that all the bridges are reconnected, but not touching each other, and you are on your way.
I know it sounds incredible, but that is actually all there is to it. Your processor, if done correctly, is ready to be overclocked. It can now be set to run at different clock frequencies, eliminating the need to increase the FSB."
http://computer-communication.blogspot.com/2007/06/unlocking...
Of course there was the AMD multi-core CPUs that had some of the cores disabled via software and sold at cheaper prices. Of course there were software hacks available to re-enable those cores.
[0]https://en.wikipedia.org/wiki/Double-sided_disk
With 16 GB and a SATA3 controller or PCIE-to-NVMe adapter it would still be a respectable performer today, in spite of the DDR2.
Good times. Nowadays I just buy a laptop with everything soldered on it and complain a lot if I have to reboot it every month.
I also maxed out the RAM to 32mb and upgrade the SRAM in my Trident video card so I could do 256 color in Windows 95.
I also had a 40mb Erwin Tape Drive for backing up my 100mb HDD. It wasn't fast but it was faster than reinstalling Tie Fighter from Floppies.
https://en.m.wikipedia.org/wiki/Intel_80487SX
Tangent: I love that I can pull up forum threads from two decades ago.
After that, it would boot, but the antivirus software I was running said it had a virus. I guess that was the only msdos program in the boot sequence that performed floating point.
Celeron 950MHz, 512MB RAM, 20GB HDD, 12x CD-ROM (not DVD :), floppy, add-on IDE Ethernet card. Win XP actually.
We have an unused Win8 desktop that we'll be refurbishing and gifting to her soon.
Great era to build your own stuff.
Perhaps "they" (the committee) were afraid that computers were getting powerful enough that we might just find something. Better pull the plug.
Sad to see they're shutting down.
I'm still curious about the extent to which the claim of idle compute power going to waste was true when various projects were set up. For example, the press release about the Cooperative Computing Awards (I've forgotten whether this was 1998 or 1999) says
> However, the computer industry produces millions of new computers each year, which sit idle much of the time, running screen savers or waiting for the user to do something.
https://www.eff.org/press/releases/eff-offers-cooperative-co...
The main thing I can think of that would have supported the concept of idle cycles going to waste was
https://en.wikipedia.org/wiki/HLT_(x86_instruction)
> [the CPU-idle instruction] was not specifically designed to reduce power consumption until the release of the Intel DX4 processor in 1994.
I participated in DESChall and distributed.net's RC5 challenge and I remember being told, and telling others, that we were only using resources that would otherwise have gone to waste. Were we already greatly mistaken about that in 1997?
- The baseline electricity the computer is using. This will get used and produce no S@H work either way (in an era where our computers didn't go to sleep)
- The added electricity to do S@H work. This certainly was not idle, but also was a relatively small amount relative to an entire household's use.
- The hardware itself.
I would say that S@H were making use of hardware that was "going to waste" (this is CAPEX the project didn't need to spend), was having volunteers donate the added electricity, and could be said to be making use of the baseline electricity inasmuch as it would be used either way, and if S@H were to set up their own compute cluster, they would also incur similar overhead.
I don't know what system efficiencies were like in the late 90's, but I would totally believe someone if they said "idle" was 40% the power consumption of "max usage". Using that number as an example, "donated electricity" gets (1/0.6=) 1.6x as much compute-per-watt as if it was donated to run an equivalent, dedicated, machine. Maybe you can convince yourself that you were using "idle cycles" and donating on top of that?
So running distributed.net or Seti@Home was using unused processing capacity for a marginal increase in power draw. If you ran them overnight with your monitor (CRT typically) your computer was drawing less power than if you were sitting up all night on IRC or playing Quake. This changed once frequency scaling was commonplace on desktops.
I ran a large cluster of university machines surreptitiously for this project many years ago.
Surprised you didn’t just use TACC credits for UT students.
I was way up there in the rankings for a while, until the sysadmins started investigating the strange traffic patterns...
https://www.mersenne.org/
I figured it that Prime95 triggered the freeze within 30 seconds of staring the torture test. Then it was just a matter of swapping out the ram and then the power supply. The new power supply let the computer go on running the test.
https://primes.utm.edu/notes/faq/why.html
I hope you're kidding ...
We will probably never find or communicate with ET, so the effort is a waste of time and resources. Space is just too big.
But interesting numbers are already known to be valuable in group theory, encryption, and other areas. Golomb numbers are used today, for example:
https://en.wikipedia.org/wiki/Golomb_ruler
And my standard disclaimer for technological illiterates: "It's not all about you."