Choose between the Ryzen 5 3600X and the Ryzen 7 3700X based on your needs.
Choose between the Ryzen 5 3600X and the Ryzen 7 3700X based on your needs.
having extra threads on a CPU can assist with deep learning, but it's less critical than a GPU with ample memory. A budget CPU works fine for basic tasks, though a more powerful CPU is better for reinforcement learning. For the cost of an 1100k, I’d recommend a system built around a Ryzen 5 3600x if gaming matters most.
Because of the high price, I chose 1650 Super right now. Upgrading to 20 Series could be possible later when prices stabilize. Gaming isn’t my main concern. If I go with 3700x, I’ll stick with a stock cooler now; choosing 3600x lets me save money and buy a CPU cooler later. I’m stuck between the two options.
The Stealth stock cooler works well and is a solid choice, particularly when you're not pushing the CPU to its limits and your case offers decent airflow. You don’t need to invest in a separate CPU cooler for this setup. The tasks you’re handling would gain from more cores, but six should suffice. I’d recommend the 3700X personally, though you have no reason to choose otherwise. Regarding Intel options, an 11600K could match the 3600X in efficiency, but the 3700X will outperform it.
It's not crucial. Most ML tasks depend on GPUs, and those two extra cores offer limited gain. It could help if you're building large projects; otherwise, upgrading to a faster SSD and more RAM would be more beneficial, especially when working with Node.js.
The 3600x stock cooler might not be sufficient; consider upgrading to a better CPU cooler for optimal performance.
If you're not aiming to boost performance or worry about overheating, the standard cooler will suffice. I own a gammax 400 since my 3700x wasn't included. The processing units aren't heavily used during training, and data handling stays on one core most of the time. Add more RAM if needed—64gb is sufficient for medium-sized datasets.