First Time Build for Research

[thesadnoodle]

Budget (including currency): ~$800 - $1,000 USD

Country: United States

Games, programs or workloads that it will be used for: Scientific Research, Machine Learning, light gaming (CS:GO, esports, etc.)

Other details (existing parts lists, whether any peripherals are needed, what you're upgrading from, when you're going to buy, what resolution and refresh rate you want to play at, etc): I'm upgrading from an 8-core intel macbook pro, so I'm going to need everything including a monitor. 

 

[lemij31400]

Personally id consider looking at Xeon Scalable 1st/2nd Gen and or EPYC Naples/Rome as they can be gotten for fairly cheap and allow for alot more expansion than consumer parts have to offer considering you will be using for ML and most likely wanting mGPU in the future.

 

Although have you considered using a cloud service for your research/ML model training so you only pay for what you use? as with such a tight budget its hard to get the best of both worlds with that kind of budget.

 

all being said if you want a PC for that and cant increase budget/need local machine could go with something like this to get you started:

https://pcpartpicker.com/list/yL6PQ6

[CPU: Intel Core i3-12100F 3.3 GHz Quad-Core Processor  ($104.99 @ Newegg) 
Motherboard: Gigabyte B660M DS3H DDR4 Micro ATX LGA1700 Motherboard  ($109.99 @ Amazon) 
Memory: Silicon Power GAMING 16 GB (2 x 8 GB) DDR4-3200 CL16 Memory  ($48.97 @ Amazon) 
Storage: TEAMGROUP MP33 1 TB M.2-2280 NVME Solid State Drive  ($69.98 @ Amazon) 
Video Card: MSI GeForce RTX 3060 Ti LHR 8 GB VENTUS 2X OCV1 Video Card  ($469.99 @ Amazon) 
Case: Phanteks Eclipse P300A Mesh ATX Mid Tower Case  ($69.99 @ Newegg) 
Power Supply: EVGA SuperNOVA G6 650 W 80+ Gold Certified Fully Modular ATX Power Supply  ($74.99 @ Newegg) 
Monitor: AOC G2490VX 23.8" 1920x1080 144 Hz Monitor  ($159.99 @ Amazon) 
Total: $1108.89]

 

its not the best by any means but with that kind of budget you dont really have any room to wiggle with for your needs

[thesadnoodle]

Awesome thank you! I know it is a tight budget, but I'm hoping to build something now and be able to upgrade it with the complexity of my projects in the future. 

[Ryuikko]

About what id reccommend https://pcpartpicker.com/list/yPnC4s

The 12400f is a solid cpu and is about the baseline for most cpu intensive stuff which irrc, ml is cpu based

Stock cooler just to save a bit of money

Good board for the price

Cheapest kit of 16gb ram

Decent 1tb storage 

6600xt is a solid performer, but if you need nvidia specific tools like cuda, I'd get the cheapest 3060 you can find instead

Good case with airflow

Good quality 750 watt psu for some future upgrading 

 

If this is out of your budget, a downgrade to a 6600 would be my first recommendation

 

[igormp]

On 7/24/2022 at 4:56 PM, Ryuikko said:

About what id reccommend https://pcpartpicker.com/list/yPnC4s

The 12400f is a solid cpu and is about the baseline for most cpu intensive stuff which irrc, ml is cpu based

Stock cooler just to save a bit of money

Good board for the price

Cheapest kit of 16gb ram

Decent 1tb storage 

6600xt is a solid performer, but if you need nvidia specific tools like cuda, I'd get the cheapest 3060 you can find instead

Good case with airflow

Good quality 750 watt psu for some future upgrading 

 

If this is out of your budget, a downgrade to a 6600 would be my first recommendation

 

Keep in mind that AMD is useless for anything ML, and ML uses a lot of GPU, not CPU.

 

OP is better off with a 3060 or 2060 Super.

[brob]

On 7/24/2022 at 2:54 PM, thesadnoodle said:

Awesome thank you! I know it is a tight budget, but I'm hoping to build something now and be able to upgrade it with the complexity of my projects in the future. 

 

Whatever you build now will need replacing rather than upgrading. Keep this in mind as spending more today to facilitate a possible upgrade is usually a waste.