High End Desktop PC

PoweruseR

Well-known member
  • Oct 12, 2011
    10,637
    8,427
    113
    In Your Heart
    Dear Sir la,

    Mata Ai walata ekka pc ekak oni used ekak thama balanne .. daraz eketh intel xeon apu pc thiyenawa. 80-134 athare

    Eth mata vram eka wadiyen oni . 2.50k wage max yata karanna puluwan. Kohenda ganna puluwan ram eka aduma 64gb oni.

    Gaming walata mage wenam pc ekak have rtx4060 + ryzen 7700x + 32gb 5600 ddr5

    * Mata 5060ti 16gb eka 1lak walata genna ganna puluwan brand new .eka dala ara machine ekama gannada? Rtx 5070ti 16gb gahannada eka 2 laks wedd ganna puluwan

    Note: lapa kota wahana sihinayata keliya a nisa eke dukara hitha hadenna meka karanne 🥹🥹🥹
     

    TNHM

    Well-known member
  • Jan 3, 2017
    5,328
    15,488
    113
    දැන් තියෙන PC එකට GPU එකක් ගහපන් 16gb වගේ හරි 24gb වගේ හරි.
    AI වලට කියලා වෙනම PC හදන්න එපා. උබ වැඩේ serious යනවා නම් ඉතින් cloud තමයි
     

    PoweruseR

    Well-known member
  • Oct 12, 2011
    10,637
    8,427
    113
    In Your Heart
    ලක්ශ දෙකේ එක අරං නූඩ්ස් හදලා මේකෙ දාපං
    Used ewa thiyenawa honda ewath.lankawe kukullu use karapuwa newe. Company wala use karapuwa

    දැන් තියෙන PC එකට GPU එකක් ගහපන් 16gb වගේ හරි 24gb වගේ හරි.
    AI වලට කියලා වෙනම PC හදන්න එපා. උබ වැඩේ serious යනවා නම් ඉතින් cloud තමයි
    Runpod thama eth salli aparade nisa baluwe.5070ti 16 ekak gahuwa nam hari wage
    ------ Post added on Aug 18, 2025 at 11:32 PM
     

    hapriyankara

    Well-known member
  • May 6, 2008
    5,279
    4,545
    113
    නොපෙනෙන මානයේ
    දැං NPU (AI accelerator) එක්කම එන CPU තියෙනව. ඒකක් බලල ගනිං

    AMD Ryzen 5 8600G (https://www.gamestreet.lk/product_view.php?pid=Mjg0MA==)

    MSI B650 Gaming Plus (Wi-Fi) DDR5 (https://www.gamestreet.lk/product_view.php?pid=MTk4NQ==)

    MSI Spatium M450 1TB PCIE 4.0 NVME M.2 (https://www.gamestreet.lk/product_view.php?pid=MjMwMA==)

    Corsair Vengeance LPX 32GB DDR5 5600MHz (https://www.gamestreet.lk/product_view.php?pid=MTkwOQ==)

    Gamdias Aura GC6 Elite ARGB Mid-Tower Case (https://www.gamestreet.lk/product_view.php?pid=Mjg2NQ==)

    GAMDIAS HELIOS M1-550B 550W 80 Plus Bronze (https://www.gamestreet.lk/product_view.php?pid=MjQ4Mg==)

     
    • Like
    Reactions: Asmodeus

    NEMISIS

    Well-known member
  • Nov 13, 2013
    11,348
    19,442
    113
    Colombo
    I will buy an AMD Strix Helo AI Max 395 for my home AI server. That comes with 128GB unified memory(96GB of which can be used as VRAM) and said to capable of running 120B+ LLM.
    cost is around $2000. Price is justifiable only if you are planning to run 24x7 AI work load.

    You can try to buy a MacMini with upgraded RAM, because of its unified memory, those are very capable AI Machines.
     

    PoweruseR

    Well-known member
  • Oct 12, 2011
    10,637
    8,427
    113
    In Your Heart
    දැං NPU (AI accelerator) එක්කම එන CPU තියෙනව. ඒකක් බලල ගනිං

    AMD Ryzen 5 8600G (https://www.gamestreet.lk/product_view.php?pid=Mjg0MA==)

    MSI B650 Gaming Plus (Wi-Fi) DDR5 (https://www.gamestreet.lk/product_view.php?pid=MTk4NQ==)

    MSI Spatium M450 1TB PCIE 4.0 NVME M.2 (https://www.gamestreet.lk/product_view.php?pid=MjMwMA==)

    Corsair Vengeance LPX 32GB DDR5 5600MHz (https://www.gamestreet.lk/product_view.php?pid=MTkwOQ==)

    Gamdias Aura GC6 Elite ARGB Mid-Tower Case (https://www.gamestreet.lk/product_view.php?pid=Mjg2NQ==)

    GAMDIAS HELIOS M1-550B 550W 80 Plus Bronze (https://www.gamestreet.lk/product_view.php?pid=MjQ4Mg==)

    mage thiyena machine eka 7700x , oya board eka thiyenne , 64 ddr5 5600 ,NZXT 80plus PSU.
     

    MihiCherub

    Well-known member
  • Sep 14, 2009
    18,849
    1
    9,594
    113
    Gampaha
    I will buy an AMD Strix Helo AI Max 395 for my home AI server. That comes with 128GB unified memory(96GB of which can be used as VRAM) and said to capable of running 120B+ LLM.
    cost is around $2000. Price is justifiable only if you are planning to run 24x7 AI work load.

    You can try to buy a MacMini with upgraded RAM, because of its unified memory, those are very capable AI Machines.
    Comfy or anything Pytorch doesn't work. ROCm 6.4.1 supposedly supports Strix Halo. It kinda does. Somethings do work. But enough doesn't that Pytorch powered SD doesn't work.
     

    THE_RAIN

    Well-known member
  • Sep 2, 2016
    1,805
    997
    113
    Sri Lanka
    I will buy an AMD Strix Helo AI Max 395 for my home AI server. That comes with 128GB unified memory(96GB of which can be used as VRAM) and said to capable of running 120B+ LLM.
    cost is around $2000. Price is justifiable only if you are planning to run 24x7 AI work load.

    You can try to buy a MacMini with upgraded RAM, because of its unified memory, those are very capable AI Machines.
    Not into AI but what kinda work you guys do from those machines ?
    I mean what are the AI tasks that runs 24x7 on a local machine?
     

    NEMISIS

    Well-known member
  • Nov 13, 2013
    11,348
    19,442
    113
    Colombo
    Comfy or anything Pytorch doesn't work. ROCm 6.4.1 supposedly supports Strix Halo. It kinda does. Somethings do work. But enough doesn't that Pytorch powered SD doesn't work.
    Waiting for more reviews, But I've seen a lot of LLMs work with good enough performance. There won't be Nvidia level support but tool and library support will increase.
    Nvidia has GDX Sparx, but that will cost $3000, however I don't think we can buy this for MSRP, It will be a better option, but can't justify the cost.
    Not into AI but what kinda work you guys do from those machines ?
    I mean what are the AI tasks that runs 24x7 on a local machine?
    I have a fully smart home built using ESPHome devices + Home-assistant, I need an AI to fully automate and decision making to run the home.
    Things like predicting the solar output and optimizing the power usage. (As I am on fully off grid).

    One simple example, If there was less solar output in the last few days, make a prediction that today's output will also be lower, so in order to fully charge off-grid batteries, AI needs to turn off hot water, and save power.

    And I need this to work offline without any 3rd party API support, that will be my own HAL9000
     
    Last edited:

    infernocus

    Well-known member
  • Jul 8, 2025
    2,405
    4,787
    113
    Waiting for more reviews, But I've seen a lot of LLMs work with good enough performance. There won't be Nvidia level support but tool and library support will increase.
    Nvidia has GDX Sparx, but that will cost $3000, however I don't think we can buy this for MSRP, It will be a better option, but can't justify the cost.

    I have a fully smart home built using ESPHome devices + Home-assistant, I need an AI to fully automate and decision making to run the home.
    Things like predicting the solar output and optimizing the power usage. (As I am on fully off grid).

    One simple example, If there was less solar output in the last few days, make a prediction that today's output will also be lower, so in order to fully charge off-grid batteries, AI needs to turn off hot water, and save power.

    And I need this to work offline without any 3rd party API support, that will be my own HAL9000
    This is a stupid idea, all of these can be done by using ChatGPT or whatever other API for $20 per month, $2000/$20 = 100months (assuming your AI pc costs $2000). You can run this for 100Months = 8+ years. All these low end/home AI servers will be outdated in 3-4 years and will become a expensive e waste. But $20 API cost will be spread across 8 years and it is peanuts.
    There are cheap MIni PCs that are perfect to run rest of the system

    If I were you,
    I will use ChatGPT paid + couple of mini PCs to run agents and other automations. I don't see a huge advantage of running a sophisticated AI locally.