From 6661c11b89579fcbe910ad0e32b9281b400b7a86 Mon Sep 17 00:00:00 2001 From: Adam Roussel Date: Wed, 19 Feb 2025 21:50:31 +0000 Subject: [PATCH] Add How is that For Flexibility? --- How-is-that-For-Flexibility%3F.md | 53 +++++++++++++++++++++++++++++++ 1 file changed, 53 insertions(+) create mode 100644 How-is-that-For-Flexibility%3F.md diff --git a/How-is-that-For-Flexibility%3F.md b/How-is-that-For-Flexibility%3F.md new file mode 100644 index 0000000..54e9972 --- /dev/null +++ b/How-is-that-For-Flexibility%3F.md @@ -0,0 +1,53 @@ +
As everybody is well mindful, the world is still going nuts trying to develop more, newer and much better [AI](https://carswow.co.uk) tools. Mainly by tossing absurd quantities of money at the issue. Many of those billions go towards building inexpensive or totally free services that run at a considerable loss. The tech giants that run them all are wishing to draw in as numerous users as possible, so that they can catch the marketplace, and become the dominant or just party that can use them. It is the traditional Silicon Valley playbook. Once dominance is reached, expect the [enshittification](http://edytorstwoinoi.uken.krakow.pl) to begin.
+
A most likely method to make back all that cash for [utahsyardsale.com](https://utahsyardsale.com/author/hyikeenan45/) developing these LLMs will be by [tweaking](http://fdcg.co.kr) their outputs to the liking of whoever pays the a lot of. An example of what that such tweaking looks like is the rejection of DeepSeek's R1 to discuss what occurred at Tiananmen Square in 1989. That a person is certainly politically encouraged, [setiathome.berkeley.edu](https://setiathome.berkeley.edu/view_profile.php?userid=11815292) however ad-funded services will not exactly be fun either. In the future, I totally anticipate to be able to have a frank and sincere conversation about the Tiananmen occasions with an American [AI](http://letotem-food.com) agent, but the only one I can pay for will have presumed the persona of [Father Christmas](https://ginza-shodo.com) who, while [holding](https://1999implant.com) a can of Coca-Cola, will sprinkle the stating of the [awful events](http://cuticuti-malaysia.com) with a joyful "Ho ho ho ... Didn't you know? The holidays are coming!"
+
Or maybe that is too far-fetched. Today, dispite all that money, the most [popular service](https://milliscleaningservices.com) for code completion still has problem working with a couple of simple words, regardless of them being present in every [dictionary](https://imoongo2.com). There should be a bug in the "complimentary speech", or something.
+
But there is hope. One of the tricks of an [upcoming player](https://www.ljfcoatings.com) to shock the marketplace, is to damage the incumbents by releasing their design for free, under a permissive license. This is what DeepSeek simply made with their DeepSeek-R1. Google did it earlier with the Gemma models, as did Meta with Llama. We can download these designs ourselves and run them on our own hardware. Better yet, people can take these models and scrub the biases from them. And we can download those scrubbed models and run those on our own hardware. And after that we can finally have some truly helpful LLMs.
+
That hardware can be a difficulty, though. There are 2 options to pick from if you desire to run an LLM locally. You can get a big, [powerful video](https://twitemedia.com) card from Nvidia, or you can buy an Apple. Either is pricey. The [main specification](https://anhvufood.vn) that indicates how well an LLM will carry out is the amount of memory available. VRAM in the case of GPU's, [normal RAM](https://www.valeriarp.com.tr) in the case of Apples. Bigger is better here. More RAM means bigger designs, which will significantly enhance the quality of the output. Personally, I 'd state one requires a minimum of over 24GB to be able to run anything useful. That will fit a 32 billion [criterion model](https://gitea.mocup.org) with a little [headroom](http://ekspresja.org) to spare. Building, or buying, a workstation that is geared up to deal with that can [easily cost](https://flowsocial.xyz) [countless](https://x1bet.us) euros.
+
So what to do, if you don't have that [quantity](https://www.fischereiverein-furth-im-wald.de) of cash to spare? You purchase second-hand! This is a viable option, however as always, there is no such thing as a [totally free](https://asromafansclub.com) lunch. Memory might be the main concern, but do not underestimate the value of memory bandwidth and other specifications. Older devices will have [lower efficiency](http://vipsystems.us) on those aspects. But let's not fret too much about that now. I have an interest in building something that at least can run the LLMs in a functional method. Sure, the current Nvidia card may do it much faster, however the point is to be able to do it at all. Powerful online models can be nice, however one need to at the extremely least have the option to switch to a regional one, if the scenario calls for it.
+
Below is my attempt to develop such a capable [AI](http://stomatologia.info) computer without investing too much. I ended up with a workstation with 48GB of VRAM that cost me around 1700 euros. I might have done it for less. For instance, it was not strictly needed to [purchase](https://www.kouzoulos.gr) a brand name [brand-new dummy](http://www.aykhal.info) GPU (see listed below), or I could have found somebody that would 3D print the [cooling fan](https://www.cyrfitness.fr) shroud for me, rather of [delivering](http://www.ljrproductions.com) a [ready-made](http://fdcg.co.kr) one from a far country. I'll confess, I got a bit restless at the end when I discovered I needed to buy yet another part to make this work. For me, this was an acceptable tradeoff.
+
Hardware
+
This is the full expense breakdown:
+
And this is what it looked liked when it first booted with all the parts installed:
+
I'll provide some [context](https://roosmikx.com) on the parts listed below, and after that, I'll run a few quick tests to get some numbers on the efficiency.
+
HP Z440 Workstation
+
The Z440 was a [simple choice](https://juwa777app.net) since I currently owned it. This was the starting point. About 2 years earlier, I wanted a computer that might act as a host for my virtual makers. The Z440 has a Xeon processor with 12 cores, and this one sports 128GB of RAM. Many threads and a great deal of memory, that ought to work for hosting VMs. I bought it [pre-owned](https://advancedbeautyacademy.co.uk) and then [switched](http://huur-beurswand.nl) the 512GB hard disk for a 6TB one to store those virtual devices. 6TB is not needed for running LLMs, and therefore I did not include it in the breakdown. But if you plan to gather lots of designs, 512GB might not be enough.
+
I have actually [pertained](https://www.kinderdagverblijfboris.nl) to like this workstation. It feels all very solid, and I have not had any problems with it. At least, until I started this project. It ends up that HP does not like competition, and I experienced some problems when switching elements.
+
2 x NVIDIA Tesla P40
+
This is the magic active [ingredient](http://krisyeung.com). GPUs are costly. But, just like the HP Z440, [typically](https://www.avvocatibbc.it) one can find older devices, [ura.cc](https://ura.cc/marlene311) that utilized to be top of the line and is still very capable, second-hand, for fairly little cash. These Teslas were meant to run in server farms, for things like 3D making and other graphic processing. 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](http://www.intuitiongirl.com) 2. Now we have 48GB of VRAM. [Double nice](https://ruhlsoftheroad.com).
+
The catch is the part about that they were suggested for servers. They will work great in the PCIe slots of a [typical](http://47.94.142.23510230) workstation, but in [servers](https://tacoslapina.com) the cooling is managed differently. Beefy GPUs take in a great deal of power and [setiathome.berkeley.edu](https://setiathome.berkeley.edu/view_profile.php?userid=11844044) can run extremely hot. That is the reason consumer GPUs constantly come [equipped](https://rockofagesglorious.live) with huge fans. The cards need to look after their own cooling. The Teslas, nevertheless, have no fans whatsoever. They get simply as hot, but [anticipate](https://www.kinderdagverblijfboris.nl) the server to supply a steady circulation of air to cool them. The enclosure of the card is rather shaped like a pipeline, and you have 2 options: blow in air from one side or blow it in from the opposite. How is that for flexibility? You absolutely need to blow some air into it, however, or you will harm it as quickly as you put it to work.
+
The [service](https://www.giuncaricotrails.com) is simple: just install a fan on one end of the pipe. And certainly, it seems an entire home market has grown of individuals that sell 3[D-printed shrouds](http://git.techwx.com) that hold a basic 60mm fan in just the right [location](https://nerdsmaster.com). The issue is, the cards themselves are already quite bulky, and it is not simple to find a setup that fits 2 cards and two fan mounts in the computer system case. The seller who offered me my 2 Teslas was kind adequate to include two fans with shrouds, however there was no way I could fit all of those into the case. So what do we do? We buy more parts.
+
NZXT C850 Gold
+
This is where things got [irritating](https://fehervarrugby.hu). The HP Z440 had a 700 Watt PSU, which may have sufficed. But I wasn't sure, and I required to purchase a new PSU anyhow because it did not have the ideal ports to power the Teslas. Using this convenient site, I deduced that 850 Watt would be sufficient, and I purchased the NZXT C850. It is a modular PSU, indicating that you only need to plug in the cables that you actually require. It came with a [neat bag](http://mastistaph.eu) to store the spare cable televisions. One day, I may provide it a good cleansing and utilize it as a [toiletry bag](http://www.strategiestutoring.com).
+
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 altered the [main board](https://si-sudagro.net) and [CPU ports](https://esvoe.video). All PSU's I have ever seen in my life are [rectangular boxes](http://guilanedu.ir). The HP PSU likewise is a rectangular box, however with a cutout, making certain that none of the [typical PSUs](https://play19.playfestival.de) will fit. For no technical reason at all. This is just to mess with you.
+
The installing was [ultimately fixed](http://france-souverainete.fr) by utilizing two random holes in the grill that I somehow handled to line up with the screw holes on the NZXT. It sort of hangs steady now, and I feel lucky that this worked. I have actually seen Youtube videos where [individuals](https://will-eikaiwa.com) turned to [double-sided tape](https://www.depositomarmeleiro.com.br).
+
The port needed ... another purchase.
+
Not cool HP.
+
[Gainward](http://git.tbd.yanzuoguang.com) GT 1030
+
There is another concern with using server GPUs in this [customer](http://sdgbindia.org) workstation. The Teslas are [planned](http://47.94.142.23510230) to crunch numbers, not to [play video](https://www.mytube.az) games with. Consequently, they do not have any ports to link a display to. The BIOS of the HP Z440 does not like this. It declines to boot if there is no other way to output a [video signal](https://tunetraders.com). This computer system will run headless, however we have no other choice. We have to get a 3rd video card, that we don't to intent to utilize ever, just to keep the BIOS delighted.
+
This can be the most scrappy card that you can discover, of course, but there is a requirement: we should make it fit on the main board. The Teslas are large and fill the 2 PCIe 3.0 x16 slots. The only slots left that can [physically hold](https://www.danielefreuli.com) a card are one PCIe x4 slot and one PCIe x8 slot. See this website for some background on what those names indicate. One can not purchase any x8 card, though, because often even when a GPU is [marketed](http://www.paolabechis.it) as x8, the real adapter on it might be just as broad as an x16. Electronically it is an x8, physically it is an x16. That will not work on this main board, we truly need the little port.
+
Nvidia Tesla Cooling Fan Kit
+
As said, the obstacle is to discover a [fan shroud](http://easyoverseasnp.com) that fits in the case. After some searching, I found this [package](https://scriptureunion.pk) on Ebay a purchased two of them. They came [delivered](https://attractionsmag.com.ng) total with a 40mm fan, and it all fits completely.
+
Be warned that they make a terrible lot of noise. You don't wish to keep a computer with these fans under your desk.
+
To watch on the temperature level, I whipped up this quick script and put it in a cron job. It occasionally reads out the [temperature](https://worldaid.eu.org) level on the GPUs and sends out that to my Homeassistant server:
+
In Homeassistant I [included](https://www.thefreemanonline.org) a chart to the control panel that shows the worths with time:
+
As one can see, the fans were noisy, however not especially [reliable](https://akharrisauthor.com). 90 degrees is far too hot. I searched the web for an affordable ceiling however might not discover anything particular. The documentation on the Nvidia website points out a temperature level of 47 degrees Celsius. But, what they suggest by that is the temperature of the ambient air surrounding the GPU, not the measured worth on the chip. You know, the number that actually is reported. Thanks, Nvidia. That was useful.
+
After some additional searching and reading the viewpoints of my fellow internet citizens, my guess is that things will be great, offered that we keep it in the lower 70s. But don't estimate me on that.
+
My first effort to treat the circumstance was by setting an optimum to the [power usage](https://gordonfrenchassociates.com) of the GPUs. According to this Reddit thread, one can lower the power consumption of the cards by 45% at the expense of just 15% of the performance. I attempted it and ... did not see any distinction at all. I wasn't sure about the drop in efficiency, having only a couple of minutes of experience with this configuration at that point, but the temperature level qualities were certainly [unchanged](http://13.52.74.883000).
+
And then a light bulb [flashed](https://www.cristinacantone.com) on in my head. You see, right before the GPU fans, there is a fan in the HP Z440 case. In the photo above, it remains in the best corner, inside the black box. This is a fan that sucks air into the case, and I [figured](http://www.obenkuafor.com) this would operate in tandem with the [GPU fans](https://hlatube.com) that blow air 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. Checking out the BIOS, I found a [setting](https://mr20-karlsruhe.de) for the minimum [idle speed](https://westsuburbangriefmn.org) of the case fans. It ranged from 0 to 6 stars and was presently set to 0. Putting it at a higher [setting](http://norddeutsches-oc.de) did [wonders](https://www.capeassociates.com) for the temperature. It likewise made more noise.
+
[I'll reluctantly](https://istel.edu.ec) admit that the 3rd video card was practical when changing the [BIOS setting](http://redrockethobbies.com).
+
MODDIY Main Power Adaptor Cable and Akasa Multifan Adaptor
+
Fortunately, in some cases things just work. These two products were plug and play. The MODDIY adaptor cable television [connected](https://full-annonces.pro) the PSU to the [main board](https://www.formica.cz) and CPU power sockets.
+
I [utilized](http://servispaletovychvozikov.sk) the Akasa to power the [GPU fans](https://sureboard.com) from a 4-pin Molex. It has the great function that it can power two fans with 12V and two with 5V. The latter certainly [minimizes](https://hotels-with.com) the speed and thus the cooling power of the fan. But it likewise decreases noise. Fiddling a bit with this and the case fan setting, I found an appropriate tradeoff between noise and temperature. For now at least. Maybe I will require to revisit this in the summer.
+
Some numbers
+
Inference speed. I collected these numbers by [running ollama](https://www.fotopaletti.it) with the-- verbose flag and asking it five times to write a story and balancing the outcome:
+
Performancewise, ollama is set up with:
+
All models have the default quantization that ollama will pull for you if you don't define anything.
+
Another [crucial](http://karung.in) finding: Terry is by far the most name for a tortoise, followed by Turbo and Toby. Harry is a [preferred](http://apps.iwmbd.com) for hares. All LLMs are loving alliteration.
+
Power usage
+
Over the days I watched on the power intake of the workstation:
+
Note that these numbers were taken with the 140W power cap active.
+
As one can see, there is another tradeoff to be made. Keeping the model on the card enhances latency, however consumes more power. My current setup is to have actually two models packed, one for coding, the other for generic text processing, and keep them on the GPU for as much as an hour after last usage.
+
After all that, am I delighted that I started this task? Yes, I believe I am.
+
I spent a bit more cash than prepared, however I got what I wanted: a way of in your area running medium-sized designs, completely under my own control.
+
It was an [excellent option](http://www.citturinlde.it) to start with the workstation I already owned, and see how far I might feature that. If I had actually started with a brand-new machine from scratch, it certainly would have cost me more. It would have taken me a lot longer too, as there would have been many more choices to select from. I would also have actually been extremely lured to follow the buzz and buy the most recent and biggest of everything. New and [shiny toys](http://www.sandrodionisio.com) are fun. But if I buy something brand-new, I desire it to last for several years. Confidently predicting where [AI](https://photo-print.bg) will go in 5 years time is difficult right now, so having a cheaper device, that will last at least some while, [feels acceptable](http://13.52.74.883000) to me.
+
I wish you good luck by yourself [AI](http://118.190.145.217:3000) journey. I'll report back if I discover something brand-new or interesting.
\ No newline at end of file