Search results

Do not skimp out on a good ups. Most modern PSUs will not turn on or work well with ups that don't give a pure sine output. Most PSUs give an emulated digital sine wave, not a pure sine wave.

Yep, The limit on the max bandwidth, the pcie bus support, the throughput, all are an improvement in GDDR6 over GDDR5. Then again, Keep in mind the actual specs and their comparison, as how well and efficiently an architecture is used depends on the manufacturer.

Those telling intel is better for productivity have certainly never used it for AI and other machine learning workloads. Ryzen 7000 architecture is actually good for CPU offloading and CPU acceleration in tensorflow and PyTorch. It has to do with the cache size and architecture. For AI...

I do a lot of offline LLM finetuning and inferencing. Doing it offline always has limits. Don't ignore fast and large quantities of RAM, it makes CPU offloading bearable and also is cheaper per Gb than GPU. avoid AMD as Rocm hasn't gone that far yet. Stick to nvidia. If you want pure...

Have you considered AM5 ? It'd give you a path for Upgradation over AM4.

How about a Rack unit ? You can get them really cheap for second hand, and there's more room for expansion.

Depends on your budget and use. Avoid 4080 super over 4080. It costs more for realistically a small bump in performance. That applies to all super models over their base.

Ryzen 7000 series run really hot. You'd need a watercooler for sure. Don't skimp out on a good cooler, considering You'd thermal throttle fast and miss out on performance. Buy a good 360mm AIO. I have a 7950x and it runs really hot. I have a deepcool LE750, yet it can go upto 97° c. Use PBO in...