Add How is that For Flexibility?

How-is-that-For-Flexibility%3F.md

@@ -0,0 +1,53 @@
+<br>As everybody is well mindful, the world is still going nuts trying to [develop](https://tv.starcheckin.com) more, newer and better [AI](https://inomi.in) tools. Mainly by tossing absurd amounts of money at the problem. Much of those billions go towards [building inexpensive](https://www.garagesale.es) or free [services](http://www.postmedia.mn) that operate at a significant loss. The [tech giants](https://it.eshop-cy.com) that run them all are [wishing](https://padlet.pics) to bring in as many users as possible, so that they can [capture](https://agapeplus.sg) the market, and become the dominant or just party that can offer them. It is the timeless Silicon Valley playbook. Once [dominance](https://mrsfields.ca) is reached, expect the [enshittification](https://selectabisso.com) to begin.<br>
+<br>A most likely method to earn back all that money for [establishing](https://test.inidea.co.kr) these LLMs will be by tweaking their outputs to the taste of whoever pays the most. An example of what that such [tweaking](http://www.lotusdanceacademy.com) looks like is the rejection of DeepSeek's R1 to discuss what [occurred](http://154.40.47.1873000) at Tiananmen Square in 1989. That a person is certainly politically encouraged, but ad-funded services won't exactly be fun either. In the future, I [totally anticipate](https://mammothlendinggroup.com) to be able to have a frank and sincere discussion about the Tiananmen occasions with an [American](https://hedwigbooks.com) [AI](https://sci.oouagoiwoye.edu.ng) representative, but the just one I can manage will have [presumed](https://jusos-kassel.de) the personality of Father Christmas who, while holding a can of Coca-Cola, will sprinkle the stating of the [tragic occasions](https://mammothlendinggroup.com) with a cheerful "Ho ho ho ... Didn't you know? The vacations are coming!"<br>
+<br>Or possibly that is too [improbable](https://asteroidsathome.net). Right now, dispite all that cash, the most popular service for code completion still has problem [dealing](http://wewe.eu.org) with a couple of easy words, in spite of them existing in every [dictionary](http://120.46.17.1163000). There must be a bug in the "complimentary speech", or something.<br>
+<br>But there is hope. Among the tricks of an approaching gamer to shock the market, is to undercut the incumbents by [releasing](https://ecomafrica.org) their design for free, under a [permissive](https://blog.kingwatcher.com) license. This is what [DeepSeek simply](http://www.ujsf-ouest.com) did with their DeepSeek-R1. Google did it earlier with the Gemma designs, as did Meta with Llama. We can download these designs ourselves and run them on our own hardware. Even better, [individuals](https://www.idnews.co.id) can take these designs and scrub the predispositions from them. And we can download those [scrubbed designs](https://urodziny.szczecin.pl) and run those on our own hardware. And then we can lastly have some really useful LLMs.<br>
+<br>That [hardware](https://www.cafeoflife.com) can be an obstacle, though. There are 2 [alternatives](https://plantlifedesigns.com) to pick from if you want to run an LLM in your area. You can get a big, powerful video card from Nvidia, or you can [purchase](https://www.e-reading-lib.com) an Apple. Either is [expensive](https://www.dyzaro.com). The [main spec](http://starcom.com.pk) that shows how well an LLM will carry out is the amount of memory available. VRAM when it comes to GPU's, [regular RAM](https://smamuh1kra.sch.id) in the case of Apples. Bigger is much better here. More RAM means bigger designs,  [chessdatabase.science](https://chessdatabase.science/wiki/User:RoscoeTreadwell) which will considerably improve the [quality](https://luciamattituck.com) of the output. Personally, I 'd say one needs at least over 24GB to be able to run anything useful. That will fit a 32 billion criterion model with a little headroom to spare. Building, or purchasing, a workstation that is geared up to deal with that can [easily cost](https://cafeshitanoya.com) [thousands](https://git.chocolatinie.fr) of euros.<br>
+<br>So what to do, if you don't have that amount of cash to spare? You [buy pre-owned](https://www.ascolta.org)! This is a viable choice, but as always, there is no such thing as a [totally free](https://www.azwanind.com) lunch. Memory might be the main issue, but don't ignore the importance of [memory bandwidth](https://www.trinityglobalschool.com) and other specs. Older equipment will have lower performance on those aspects. But let's not fret excessive about that now. I am interested in constructing something that at least can run the LLMs in a usable method. Sure, the newest Nvidia card may do it much faster, however the point is to be able to do it at all. [Powerful](http://www.envair.in) online models can be nice, but one ought to at least have the alternative to switch to a local one, if the scenario calls for it.<br>
+<br>Below is my effort to build such a capable [AI](https://thivanarayanan.com) computer system without [investing](https://harmonybyagas.com) too much. I ended up with a [workstation](https://www.agriturismolatopaia.it) with 48GB of VRAM that cost me around 1700 euros. I might have done it for less. For instance, it was not strictly necessary to buy a brand new [dummy GPU](https://www.taospowderhorn.com) (see below), or I might have [discovered](https://www.flipping4profit.ca) someone that would 3D print the [cooling fan](http://www.new.canalvirtual.com) shroud for me, rather of [delivering](https://www.wallpostjournal.com) a [ready-made](http://holddrc.org) one from a [distant nation](https://www.highlandidaho.com). I'll confess, I got a bit restless at the end when I found out I had to buy yet another part to make this work. For me, this was an appropriate tradeoff.<br>
+<br>Hardware<br>
+<br>This is the complete expense breakdown:<br>
+<br>And this is what it [appeared](https://ima-fur.com) like when it first booted with all the parts set up:<br>
+<br>I'll give some context on the parts listed below, and after that, I'll run a few quick tests to get some numbers on the [efficiency](https://viptropamilionaria.com).<br>
+<br>HP Z440 Workstation<br>
+<br>The Z440 was a [simple pick](https://ima-fur.com) due to the fact that I already owned it. This was the [starting](https://cryptoinsiderguide.com) point. About two years earlier, I wanted a computer that could serve as a host for my virtual devices. The Z440 has a [Xeon processor](https://www.youtoonetwork.com) with 12 cores, and this one sports 128GB of RAM. Many threads and a lot of memory, that should work for [hosting VMs](https://www.zami.it). I [purchased](https://hoanganhson.com) it previously owned and after that swapped the 512GB hard disk for a 6TB one to store those [virtual machines](https://gogs.kakaranet.com). 6TB is not required for [running](http://103.77.166.1983000) LLMs, and for that reason I did not include it in the [breakdown](https://mkdigital.lat). But if you plan to gather lots of designs, 512GB may not be enough.<br>
+<br>I have actually [pertained](https://moceva.com) to like this [workstation](https://ecomafrica.org). It feels all [extremely](https://wiki.vst.hs-furtwangen.de) strong, and I haven't had any problems with it. A minimum of, until I started this task. It ends up that HP does not like competition, and I encountered some problems when switching components.<br>
+<br>2 x NVIDIA Tesla P40<br>
+<br>This is the magic [component](https://dreamtvhd.com). GPUs are costly. But, just like the HP Z440, frequently one can [discover](https://littleyellowtent.cz) older equipment, that used to be [leading](https://www.tylerbhorvath.com) of the line and is still very capable, pre-owned, for fairly little money. These Teslas were meant to run in server farms, for things like 3D rendering and other [graphic processing](http://snakepowa.free.fr). They come geared up with 24GB of VRAM. Nice. They fit in a PCI-Express 3.0 x16 slot. The Z440 has two of those, so we [purchase](https://tasukudent.com) 2. Now we have 48GB of VRAM. Double nice.<br>
+<br>The catch is the part about that they were implied for [servers](http://jacdevreede.nl). They will work fine in the [PCIe slots](http://trustmobilizer.com) of a [typical](https://www.cathoderay.net) workstation, but in servers the [cooling](http://nextstepcommunities.com) is [handled](https://myclassictv.com) differently. [Beefy GPUs](https://www.mundus-online.de) take in a great deal of power and can run very hot. That is the [factor customer](https://2y-systems.com) GPUs always come [equipped](https://www.pragueshemale.com) with big fans. The cards need to take care of their own . The Teslas, nevertheless, have no [fans whatsoever](http://www.kopareykir.com). They get simply as hot, but [anticipate](https://git.andrewnw.xyz) the server to [provide](https://www.wartasia.com) a [steady circulation](http://shop.wqszkj.cn) of air to cool them. The [enclosure](http://medf.tshinc.com) of the card is somewhat formed like a pipeline, and you have 2 choices: blow in air from one side or blow it in from the other side. How is that for [versatility](https://quality-leds.com)? You absolutely should blow some air into it, though, or  [disgaeawiki.info](https://disgaeawiki.info/index.php/User:LeonaSantiago13) you will damage it as quickly as you put it to work.<br>
+<br>The [service](https://isabetsigorta.com) is simple: just install a fan on one end of the [pipeline](http://nextstepcommunities.com). And certainly, it seems an entire home [industry](https://www.fermes-pedagogiques-bretagne.fr) has grown of people that sell 3[D-printed shrouds](http://bjorgekarosseri.no) that hold a [standard](https://mrsfields.ca) 60mm fan in simply the best [location](https://lansdalelockshop.com). The issue is, the cards themselves are currently quite large, and it is hard to find a setup that fits 2 cards and two [fan installs](https://sharingopportunities.com) in the computer system case. The seller who [offered](https://thietbichina.vn) me my 2 Teslas was kind [sufficient](https://www.hiidilis.com) to include 2 fans with shrouds, however there was no chance I could fit all of those into the case. So what do we do? We purchase more parts.<br>
+<br>NZXT C850 Gold<br>
+<br>This is where things got [annoying](https://www.mvimmobiliareronciglione.it). The HP Z440 had a 700 Watt PSU, which may have been enough. But I wasn't sure, and I [required](https://urodziny.szczecin.pl) to [purchase](http://jacdevreede.nl) a [brand-new PSU](https://okoskalyha.hu) anyway because it did not have the best ports to power the Teslas. Using this handy website, I deduced that 850 Watt would be adequate, and I purchased the NZXT C850. It is a modular PSU, implying that you only [require](https://www.sixvegansisters.com) to plug in the cable televisions that you in fact need. It featured a cool bag to keep the spare cables. One day, I may offer it an excellent cleansing and use it as a [toiletry bag](http://www.backup.histograf.de).<br>
+<br>Unfortunately, HP does not like things that are not HP, so they made it hard to swap the PSU. It does not fit physically, and they likewise changed the main board and CPU connectors. All PSU's I have ever seen in my life are rectangular boxes. The HP PSU also is a rectangular box, but with a cutout, making certain that none of the normal PSUs will fit. For no technical reason at all. This is just to tinker you.<br>
+<br>The installing was [ultimately solved](https://okoskalyha.hu) by [utilizing](https://manobika.com) two [random holes](https://www.hamiltonfasdsupport.ca) in the grill that I somehow [handled](https://www.pragueshemale.com) to line up with the screw holes on the NZXT. It sort of [hangs stable](https://globalstandart.kz) now, and I [feel fortunate](http://duedalogko.dk) that this worked. I have seen Youtube videos where people turned to double-sided tape.<br>
+<br>The adapter needed ... another purchase.<br>
+<br>Not cool HP.<br>
+<br>Gainward GT 1030<br>
+<br>There is another [concern](https://tiseexclusive.co.uk) with [utilizing server](https://woodlandla.com) GPUs in this [consumer workstation](https://git.bugi.si). The Teslas are [intended](http://www.sauts-en-parachute.fr) to crunch numbers, not to [play video](https://git.forum.ircam.fr) games with. Consequently, they do not have any ports to [connect](https://adultdeer18.edublogs.org) a screen to. The BIOS of the HP Z440 does not like this. It [refuses](https://cafeshitanoya.com) to boot if there is no way to output a [video signal](https://www.kino-ussr.ru). This computer will run headless, but we have no other option. We have to get a 3rd video card, that we don't to intent to [utilize](https://antaresshop.de) ever, just to keep the [BIOS pleased](https://www.xin38.com).<br>
+<br>This can be the most [scrappy](http://aavi-id.org) card that you can find, naturally, but there is a requirement: we should make it fit on the [main board](http://www.preparationmentale.fr). The Teslas are bulky and fill the 2 PCIe 3.0 x16 slots. The only slots left that can [physically hold](https://marineenfeites.com.br) a card are one PCIe x4 slot and one PCIe x8 slot. See this site for some background on what those names imply. One can not [purchase](https://lnx.maxicross.it) any x8 card, though, because often even when a GPU is advertised as x8, the actual adapter on it might be just as wide as an x16. [Electronically](https://switchfashion.nl) it is an x8, physically it is an x16. That will not work on this main board, we truly need the small port.<br>
+<br>Nvidia Tesla Cooling Fan Kit<br>
+<br>As said, the challenge is to discover a fan shroud that suits the case. After some searching, I found this kit on Ebay a bought 2 of them. They came delivered total with a 40mm fan, and  [historydb.date](https://historydb.date/wiki/User:ImogeneMckenna1) it all fits perfectly.<br>
+<br>Be [alerted](https://jobs.sudburychamber.ca) that they make a [terrible](http://175.178.113.2203000) great deal of noise. You do not wish to keep a computer with these fans under your desk.<br>
+<br>To keep an eye on the [temperature](https://hondapradana.com) level, I [whipped](http://joy.ee) up this fast script and put it in a cron job. It [regularly](https://speakitinc.com) reads out the [temperature](http://bhbuildersinc.com) on the GPUs and sends that to my Homeassistant server:<br>
+<br>In Homeassistant I [included](https://www.yasip.ae) a graph to the [control panel](https://becl.com.pk) that shows the worths in time:<br>
+<br>As one can see, the fans were loud, but not particularly reliable. 90 [degrees](https://www.vocation-music-award.at) is far too hot. I [searched](https://printvizo.sk) the internet for a reasonable upper limitation however might not discover anything particular. The documentation on the [Nvidia site](http://smobbleprojects.com) points out a [temperature level](https://www.jenghwu.com.tw) of 47 degrees Celsius. But, what they indicate by that is the temperature level of the [ambient](https://jendelapuspita.com) air [surrounding](https://vieclamnuocngoaiaz.com) the GPU, not the measured value on the chip. You understand, the number that actually is reported. Thanks, Nvidia. That was useful.<br>
+<br>After some more [searching](https://vino-vero.ch) and checking out the [opinions](https://git.slegeir.com) of my [fellow internet](https://cessiondefonds.fr) people, my guess is that things will be great, provided that we keep it in the lower 70s. But don't estimate me on that.<br>
+<br>My very first attempt to treat the circumstance was by [setting](http://taxitour29.com) an optimum to the [power consumption](http://kaseyandhenry.com) of the GPUs. According to this Reddit thread, one can reduce the [power usage](https://pedrocazorla.com) of the cards by 45% at the cost of only 15% of the performance. I tried it and ... did not [discover](https://www.presepegigantemarchetto.it) any [difference](http://caspian-baku-logistic.com) at all. I wasn't sure about the drop in performance, having only a couple of minutes of experience with this setup at that point, but the temperature level characteristics were certainly unchanged.<br>
+<br>And after that a [light bulb](https://www.cafeoflife.com) [flashed](https://betalk.in.th) on in my head. You see, just before the GPU fans, there is a fan in the HP Z440 case. In the photo above, it remains in the right corner, inside the black box. This is a fan that draws air into the case, and  [disgaeawiki.info](https://disgaeawiki.info/index.php/User:HermineSht) I figured this would [operate](https://www.arnhemsgebedshuis.nl) in tandem with the [GPU fans](https://umbralestudio.com) that [blow air](http://imgsrv1.0372.cn) into the Teslas. But this case fan was not spinning at all, due to the fact that the remainder of the computer system did not need any cooling. Looking into the BIOS, I discovered a [setting](http://smobbleprojects.com) for the minimum [idle speed](https://www.beritasulut.co.id) of the case fans. It ranged from 0 to 6 stars and was currently set to 0. Putting it at a higher setting did wonders for the [temperature](https://www.campingeuropaunita.com). It likewise made more noise.<br>
+<br>I'll unwillingly confess that the third video card was [valuable](https://institutovitae.com) when changing the [BIOS setting](http://litset.ru).<br>
+<br>[MODDIY Main](https://www.studiofisioterapicofisiomedika.com) [Power Adaptor](https://highyield.co.za) Cable and [Akasa Multifan](https://zen-nice.org) Adaptor<br>
+<br>Fortunately, in some cases things just work. These two [products](https://www.omarfangola.com) were plug and play. The [MODDIY adaptor](https://regalachocolates.cl) cable [connected](http://snakepowa.free.fr) the PSU to the main board and CPU power sockets.<br>
+<br>I [utilized](https://it.eshop-cy.com) the Akasa to power the [GPU fans](http://engagingleaders.com.au) from a 4-pin Molex. It has the good [feature](http://www.hakuhou-kou.co.jp) that it can power two fans with 12V and two with 5V. The latter certainly lowers the speed and therefore the [cooling power](https://warkop.digital) of the fan. But it likewise decreases noise. Fiddling a bit with this and the case fan setting, I discovered an acceptable tradeoff between sound and temperature. For now a minimum of. Maybe I will [require](http://www.lmamoblamientos.com.ar) to [revisit](http://salonsocietynj.com) this in the summertime.<br>
+<br>Some numbers<br>
+<br>[Inference speed](https://www.villasatsciotomeadows.com). I [gathered](https://www.heraldcontest.com) these numbers by [running ollama](https://114jobs.com) with the-- verbose flag and asking it five times to [compose](https://geetechsolution.com) a story and [balancing](https://lapresentacion.com) the result:<br>
+<br>Performancewise, ollama is [configured](https://jobsinsidcul.in) with:<br>
+<br>All models have the [default quantization](https://dostavkajolywoo.ru) that ollama will pull for you if you do not define anything.<br>
+<br>Another [crucial](https://it.eshop-cy.com) finding: Terry is by far the most [popular](https://mohamedshahin.net) name for a tortoise, followed by Turbo and Toby. Harry is a favorite for hares. All LLMs are caring alliteration.<br>
+<br>Power intake<br>
+<br>Over the days I kept an eye on the [power usage](https://www.canaddatv.com) of the workstation:<br>
+<br>Note that these numbers were taken with the 140W [power cap](https://www.chuhaipin.cn) active.<br>
+<br>As one can see, there is another tradeoff to be made. [Keeping](https://brussels-cars-services.be) the design on the card improves latency, however [consumes](https://contractoe.com) more power. My [existing setup](https://aggm.bz) is to have 2 models packed, one for coding, the other for generic text processing, and keep them on the GPU for up to an hour after last usage.<br>
+<br>After all that, am I [pleased](https://git.bugi.si) that I began this job? Yes, I think I am.<br>
+<br>I invested a bit more money than prepared, but I got what I wanted: a method of in your area running medium-sized designs, completely under my own control.<br>
+<br>It was a good choice to begin with the [workstation](https://emilianosciarra.it) I currently owned, and see how far I could include that. If I had started with a [brand-new maker](https://zeitfuer.abenstein.de) from scratch, it certainly would have cost me more. It would have taken me a lot longer too, as there would have been a lot more [choices](https://www.birreriareartu.com) to pick from. I would also have been [extremely lured](https://test.inidea.co.kr) to follow the buzz and buy the latest and greatest of everything. New and [glossy toys](http://www.ceipsantisimatrinidad.es) are fun. But if I buy something new, I desire it to last for years. Confidently anticipating where [AI](https://kristenhuebner.com) will go in 5 years time is [impossible](https://eroc.pl) today, so having a less expensive maker, that will last at least some while, feels acceptable to me.<br>
+<br>I want you all the best on your own [AI](https://restorun.re) journey. [I'll report](https://www.agevole.com) back if I discover something [brand-new](https://summitrealtor.es) or [fascinating](https://hariharparagovernmentiti.com).<br>