AMD Ryzen AI Max+ 395

“Both the AMD Ryzen 395/495 and the Spark have terribly slow bandwidth”

“memory bandwidth quickly becomes the limiting factor”

“None of them have something even close to the memory bandwidth or integrated GPU performance of the Ryzen AI Max processors.”

“The CPU of Ryzen is better than that of DGX Spark, especially for modern programs”

“I spent a while fiddling around with the right Linux kernel, kernel firmware, ROCm installs, etc.”

“These devices were great when they were cheaper than the DGX Spark.”

“AMD making lemonade is one of the best reasons to get a strix halo.”

“it's only useful for models under 64gb of vram due to the active parameter counts on larger MoEs”

“The issue is more to do with the CPU not being as good as say an AMD AI Max or an M4.”

“Desktops with the amd ai+ 395 are still fairly affordable for what they can do.”

“It's showing about 8-9 tk/sec which seems quite slow for something like a web search”

“would run the 4-bit quant of a trillion param model well enough for personal use”

“The AMD 395 chip supports up to 128GB unified RAM. So still not enough even at 1-bit quant unfortunately.”

“If you want more capable and are willing to spend a little more, go with the newer Ryzen AI Max+ 395 machines.”

“with enough spare memory to run another two small models on the side”

“if you arent moving to the AMD 395 or MACs then youre hitching aride on the expensive calory ride”

“it loads a big enough model and spits out tokens just fast enough that you can read what it's doing”

“Or get an AMD 395 laptop or mini PC for half the price of an equivalent mac device”

“Order of magnitude slower than dedicated GPU but at half the cost.”

“AMD made a CPU/GPU soc with unified memory specifically for ai inference.”

“chips like AMD AI MAX 395 with AMD not being beholden or useful to the AI firms”

“Great for that purpose. Not exactly a speed demon with LLMs”

“its got a local model that readily accomplishes medial tasks”

“grab a amd395 or mac studio and start having realiable workflows”

“Up until recently they didn't even support their cashcow Ryzen 395+ MAX properly.”

“they’re making many of the same compromises with Panther Lake and Ryzen AI Max.”

“first usable integrated graphics solution I've seen”

“I'd take the effortless local model capability over outright speed any day.”

“it really is amazing piece of hardware given what you can do with it”

“since it's not tied to a garbage OS nor reasonably affected by future dram costs”

“Trying to run LLM models somehow makes 128 GiB of memory feel incredibly tight.”

“I have a 13" gaming tablet with the 128 GB AMD Strix Halo chip (Ryzen AI Max+ 395, what a name).”

“Doubling down on the Ryzen Max product line might be a brilliant move.”

“For inference you want the one with 128GB soldered RAM.”

“it's still a laptop in a small desktop chassis! Depending on your use case it's great or just too underpowered”

“If you are not planning to batch, you can run it much cheaper with Ryzen AI Max SoC devices.”

“The AMD AI Max and Apple M chips are good for that audience.”

“the amd AI Max is the only one that's worth anything in my opinion”

“it does have decent performance at like 130b parameters”

“so you can DIY a Steam Machine that also runs LLMs (albeit a bit slow)”